Ahmed Ekhmaj

UNIVERSITI PUTRA MALAYSIA
WATER AND SOLUTE DISTRIBUTION PATTERN IN SOIL UNDER

POINT SOURCE TRICKLE IRRIGATION
AHMED I.M. EKHMAJ.
FK 2005 11
Dedicated to
my Parents as they did carefor me when I was little
Abstract of thesis presented to the Senate of Universiti Putra Malaysia in fulfillment of

the requirements for the degree of Doctor of Philosophy
WATER AND SOLUTE DISTRIBUTION PATTERN IN SOIL UNDER POINT

SOURCE TRICKLE IRRIGATION
AHMED I. M. EKHMAJ
December 2005
Chairman: Professor Mohd Amin Mohd Soom, PhD, P. Eng
Faculty:
Engineering
A basic requirement in the design of a trickle irrigation system is to obtain more

information about the shape and size of the wetted soil zone. This will ensure precise
placement of water and nutrients in the active root zone to meet the requirements of
precision farming. A series of laboratory and field experiments were conducted to
determine water and solute distribution pattern in soil under point source trickle
irrigation. Three types of experiments were conducted. The focus of the first type of
experiment was to study the effect of water application rate and the amount of water on
water movement in the lateral and vertical direction. River sand and sandy loam soil
were used as the media in a plexiglass container. These experiments were conducted
under laboratory conditions where the application rates of 0.75 and 3.4 I/h were used for
river sand soil while 1 and 3.0 I/h were used for sandy loam soil. The second type of
experiments was conducted on river sand in the laboratory using a wooden box. The
purpose of this experiment was to study the effect of application rate and amount of
irrigation water on the surface wetted radius. Application rates of 3, 5.5 and 7 I/h were
used. The third type of experiment was conducted under field conditions. The
experiments were designed for field evaluation of water and solute movement from a
point source. Sandy and sandy loam soils were selected for these experiments and the
application rates varied from 1.5 to 6 Ilh. The results fiom the experiments revealed that
for all soil types, lateral movement of the wetting front and the surface wetted radius as
measured at the soil surface approached a limit with elapsed time. A linear relationship
was found between vertical wetting front advance and the square root of elapsed time.
The results obtained fiom both plexi glass and wooden box experiment showed that the
water application rates caused a notable effect on the surface wetted radius, where
increase in the application rates contributed to an increase in the surface wetted radius.
On the other hand the statistical analysis of the field experiment results showed
insignificant effect of the application rates on the surface wetted radius. Increase in the
discharge rate caused a decrease in the vertical advance of the wetting front for both
sandy and sandy loam soils under field conditions, and sandy loam soil in the plexiglass
experiments. The maximum volumetric moisture content after irrigation was found in
the region just below the irrigation source. The statistical analysis of moisture
distribution data under field conditions showed insignificant effect of water application
rate on the water content distribution within the boundary of 17.5 and 27.5 cm in radial
and vertical distance, respectively. The patterns of the chloride concentration
distribution were similar to those for moisture content distribution. The effect of inlet
chloride concentration on the distribution of chloride concentration was significant in
both soils. The greater the concentration at the inlet, the higher the chloride
concentration in the soil. For both types of soil, most of the treatments indicated
insignificant effect of application rate on the chloride distribution. Two simple models
based on the average change in volumetric water content (AO), total volume of water
PtWUS I'AKAAN SULTAN A E U L SAMAD
applied (V,), application rate (q,) and the saturated hydraulic conductivity (ks) were
developed to determine the surface wetted radius (r) and vertical advance of the wetting
front (z) produced from point source trickle irrigation, r=
z=
80-0.38
V,
0.36
q,
0.26
-0.03 -0 03
-V, ~ - q,~ ~ ks .
and
-0.1
ks0.19. These models were verified with the data from this study
and other published experiments under different conditions. The results obtained from
both types of data improved the capability of using these models for designing a trickle
irrigation system. In this study, Hydrus-2D model was used to simulate water and solute
distribution under point source trickle irrigation. Good agreements were found between
simulated and experimental results regarding location of the wetting front, water
distribution and solute concentration under different application rates.
Abstrak tesis yang dikemukakan kepada Senat Universiti Putra Malaysia sebagai memenuhi
keperluan untuk ijazah Doktor Falsafah
BENTUK PERTABURAN AIR DAN SOLUT DALAM TANAH DIBAWAH

PENGAIRAN CUCUR SUMBER TITIK
Oleh
AHMED I. M EKHMAJ
Disember 2005
Pengurusi: Professor Mohd Amin Mohd Soom, PhD, P.Eng

Fakulti:
Kejuruteraan
Keperluan asas dalam rekabentuk sistem pengairan cucur adalah untuk mendapat lebih banyak
maklumat tentang bentuk dan saiz kawasan tanah yang dibasahi. Ini memastikan ketepatan
pemberian air dan baja di ladang bagi memenuhi keperluan pertanian presis. Kajian di makmal
dan di ladang telah dilakukan bagi menentukan bentuk taburan air dan solut dalam tanah di
bawah pengairan cucur sumber titik. Tiga jenis kajian telah dilakukan. Tumpuan bagi kajian
pertama adalah untuk mengkaji kesan kadar pembubuhan air dan jumlah air bagi gerakan air
secara mendatar dan menegak. Pasir sungai dan tanah lom berpasir telah digunakan sebagai
media dalam bekas plexiglass. Kajian ini telah dilakukan di dalam makrnal yang mana kadar
pengairan 0.75 dan 3.4 l/j telah diberi bagi pasir sungai, sementara kadar 1.0 dan 3.0 I/j telah
diguna bagi tanah lom berpasir. Kajian kedua telah dilakukan bagi pasir sungai di makmal
dengan menggunakan bekas kotak kayu. Tujuan kajian ini adalah untuk mengkaji kesan kadar
pembubuhan dan jumlah air pengairan keatas jarak lingkungan permukaan yang dibasahi. Kadar
pembubuhan 3, 5.5 dan 7 I/j telah digunakan. Kajian yang ketiga telah dilakukan di ladang.
Kajian direka untuk penilaian di ladang bagi pergerakan air dari sumber titik. Jenis tanah
vii
dilakukan di ladang. Kajian direka untuk penilaian di ladang bagi pergerakan air dari
sumber titik. Jenis tanah berpasir dan lom berpasir telah dipilih bagi semua kajian ini
dan kadar pembubuhan 1.5 hingga ke 6 l/j telah digunakan. Keputusan dari semua kajian
tersebut di atas menunjukkan bahawa bagi semua jenis tanah, garisan basah dan jarak
lingkungan permukaan yang dibasahi bagi gerakan mendatar yang diukur di permukaan
tanah adalah terhad dengan masa yang berlalu. Satu hubungkait secara lelurns telah
diperolehi di antara garisan basah yang tegak dan masa ber lalu berkuasa seperdua.
Keputusan yang terdapat dari kajian plexiglass dan kotak kayu menunjukkan bahawa
kadar pembubuhan air menyebabkan kesan yang jelas ke atas jarak lingkungan
permukaan yang dibasahi yang mana peningkatan kadar pembubuhan melibatkan
peningkatan jarak lingkungan permukaan yang dibasahi. Dalam ha1 yang sama,
keputusan analisis statistik bagi kajian di ladang menunjukkan kesan yang penting bagi
kadar pembubuhan ke atas jarak lingkungan permukaan yang dibasahi. Peningkatan
kadar luahan menyebabkan kekurangan pergerakan menegak bagi kedua-dua jenis tanah
berpasir dan lom berpasir di ladang, dan jenis tanah lorn berpasir bagi kajian plexiglass.
Kelembapan isipadu yang maksimum selepas pengairan terdapat dibahagian pemancar
pengairan cucur. Analisis statistik bagi taburan lembapan di ladang menunjukkan kesan
yang penting bagi kadar pembubuhan air ke atas taburan lembapan dalam jar&
lingkungan sempadan 17.5 hingga 27.5 sm masing-masing bagi jarak lingkungan dan
jarak tegak.. Bentuk taburan klorida adalah sama dengan taburan lembapan. Kesan
klorida yang pekat di alur masuk ke atas taburan klorida yang pekat adalah penting bagi
kedua-dua jenis tanah. Kepekatan lebih di alur masuk menghasilkan klorida pekat yang
tinggi di dalam tanah. Bagi kedua-dua jenis tanah, kebanyakan rawatan menunjukkan
kesan yang penting bagi kadar pembubuhan ke atas taburan klorida dalam bentuk
vii
kepekatan tanpa dimensi. Dua jenis model mudah yang berdasarkan purata pertukaran
kandungan lembapan isipadu (AO), jurnlah air yang diberi (V,), kadar pembubuhan (q,)
dan keberkondukan hidraul tepu (k,) telah dibangunkan untuk menentukan jar&
lingkungan permukaan yang dibasahi (r) dan garis basah menegah (2) yang telah dihasil
dari pengairan cucur sumber titik, r = A O - ~ . ~ ~ V W O2 6 q ~ - 003ks-0.03d m z = A@'
O 'ksO19.
3 8 ~3 ~
6 q 0~ -
Model-model ini telah dibuat pengesahan dengan data dari kajian ini dan kajian
lain yang telah diterbitkan dalam keadaan yang berlainan. Keputusan yang dapat bagi
kedua-dua jenis data mendorong keupayaan kegunaan kedua-dua model bagi rekabentuk
sistem pengairan cucur. Dalam kajian ini, model Hydrus-2D telah digunakan untuk
simulasi taburan air dan solut di bawah pengairan cucur sumber titik. Persetujuan yang
baik telah diperolehi di antara keputusan simulasi dan kajian berkenaan lokasi garisan
basah, taburan air dan kepekatan dengan kadar pembubuhan yang pembubuhan
...
Vlll
ACKNOWLEDGEMENTS
First, all praise be to Almighty ALLAH SWT. The only creator, sustainer and efficient
assembler of the world, for giving me the strength, ability and patience to complete this
work.
I would like to acknowledge and thank my supervisor Prof. Mohd Amin Mohd Soom for
accepting me as one of his postgraduate students. I would also like to thank him for his
cheery nature, assistance, guidance and mentorship throughout the years. This
dissertation would have never been completed without his help. He has provided me
with all facilities needed to complete this work. Also I would like to thank my members
of supervisory committee, Prof. Salim Said at Faculty of Engineering, Civil Engineering
Department University Sarawak Malaysia and Dr. Abdul Aziz Zakaria at Faculty of
Engineering Biological and Agricultural Engineering, University Putra Malaysia, for
their assistance, constructive suggestions, and guidance for execution of the research
project.
Special thank goes to Drs. Jirka Simiinek and Rien van Genuchten from the U.S. Salinity
Laboratory in Riverside, for their assistance and advice regarding to usage of Hydrus 2d
model.
I am grateful to the peoples of Libyan Arab Jamahiriya and Al-Fateh University who
offer me the scholarship for pursuing the PhD degree at Universiti Putra Malaysia.
I would like to take this opportunity to thank my friends for their friendship. We all have
had memorable moments.
I am especially grateful to my mother, a person whose courage, fortitude and patience I

have always admired. Last but not least, I would like to gratefully express to my brothers
and sisters for their unwavering support, best wishes and encouragement through both
good and bad times.
I certify that an Examination Committee met on 2th December 2005 to conduct the
final examination of Ahmed I. M. Ekhmaj on his Doctor of Philosophy thesis
entitled "Water and Solute Distribution Pattern in Soil under Point Source Trickle
Irrigation" in accordance with Universiti Pertanian Malaysia (Higher Degree) Act
1980 and Universiti Pertanian Malaysia (Higher Degree) Regulations 1981. The
Committee recommends that the candidate be awarded the relevant degree.
Members of the Examination Committee are as follows:
Ir. Desa Ahmad, PhD

Professor
Faculty of Engineering
Universiti Putra Malaysia
(Chairman)
Asep Sapei, PhD
Lecturer
(Internal Examiner)
Ir. Lee Teang Shui, PhD
Associate Professor
(Internal Examiner)
Wan Sulaiman Wan Harun, PhD
Professor
Faculty of Resource and Science Technology
Universiti Malaysia Sarawak
(External Examiner)
School of Graduate Studies

Date:
19 JAM 2006
This thesis submitted to the Senate of Universiti Putra Malaysia and has been accepted
as fulfillment of the requirements for the degree of Doctor of Philosophy. The members
of the Supervisory Committee are as follows:
Mohd Amin Mohd Soom, PhD

Professor
University Putra Malaysia
(Chairman)
Salim Said, PhD
Professor
Universiti Malaysia Sarawak
(Member)
Abdul Aziz Zakaria, PhD
Senior Lecture
University Putra Malaysia
(Member)
AINI IDERIS, PhD

ProfessorDean
School of Graduate Studies
Date :
xii
0 7 FEB 2006
DECLARATION
I hereby declare that the thesis is based on my original work except for equations and
citations, which have been duly acknowledged. I also declare that it has not been
previously or currently submitted for any other degree at UPM or other institutions.
TABLE OF CONTENTS
Page
DEDICATION
ABSTRACT
ABSTRAK
ACKNOWLEDGEMENTS
APPROVAL
DECLARATION
LIST OF TABLES
LIST OF FIGURES
LIST OF ABBREVIATION
v1
ix
xi
...
Xlll
xvi
xix
xxvii
CHAPTER
INTRODUCTION
1.1 Trickle Irrigation
1.2 Statement of the Problem
1.3 Objectives of the Study
1.4 Scope of the Study
1.5 Organization of the Thesis
LITERATURE REVIEW
2.1 Introduction
2.1.1 Water Flow Mechanism
2.1.2 Saturated Water Flow
2.1.3 Unsaturated Water Flow
2.1.4 Continuity Equation
2.1.4 Richard's Equation
2.2 Water Movement under Trickle Irrigation
2.2.1 Field Studies
2.2.2 Laboratory Studies
2.2.3 Mathematical Model Studies
2.3 Solute Transport in Soil
2.3.1 Convective Transport
2.3.2 Diffusion Transport
2.3.3 Dispersive Transport
2.3.4 Convection-Dispersion Equation
2.4 Previous Studies on Solute Movement under Trickle
Irrigation
2.5 Hydrus-2D Model
2.5.1 The Governing Equations
MATERIALS AND METHODS
3.1 Design of Study
3.2 Design of Study Laboratory Experiments
xiv
3.2.2 Wooden Box Experiments

3.3 Field Experiments
3.3.1 Description of Field Experiments
3.3.2 Soil Water Determination
3.3.3 Chloride Analysis
3.4 Semi Empirical Model
3.4.1 Theoretical Considerations
3.4.2 Model Optimization
3.5 Hydrus-2D Numerical Model
3.5.2 Problem Definition
3.5.3 Boundary condition
3.5.4 Inputs for Simulation Model
3.6 Statistical Criteria
3.7 Determination of the Soil Hydraulic Properties
3.7.1 Soil Texture
3.7.2 Bulk Density
3.7.3 Porosity
3.7.4 Soil Water Retention Curve
3.7.5 Saturated Hydraulic Conductivity
3.7.6 Soil Water Diffusivity
RESULTS AND DISCUSSIONS

4.1 Evaluation of Water and Solute Distribution under Point
Source Trickle Irrigation
4.1.1 Laboratory Experiments
4.1.2 Field Experiments
4.2 Predictive Model for wetting Pattern Dimensions
4.2.1 Wetted Surface Radius Equation
4.2.2 Vertical Advances of Wetting Front Equation
4.2.3 Sensitivity Analysis
4.2.4 Model's Verification
4.3 Model Evaluation Hydrus-2D
4.3.1 Wetting Front Movement
4.3.2 Moisture Content Distribution
4.4.3 Solute Distribution
SUMMARY, CONCLUSIONS AND
RECOMMENDATIONS
5.1 Summary
5.2 Conclusions
5.3 Recommendations for Further Work
REFERENCES
APPENDICES
BIODATA OF THE AUTHOR
LIST OF TABLES
Page
Table
2.1
Summary differences between saturated and unsaturated flow 2.6

(adapted from Rattan and Shukia, 2004)
Properties of the -soils used in moisture movement studies 2.12
(adapted from Goldberg et al. (1976)
Numerical values of input variables used in the predicted models.
3.25
Parameters for estimation of surface wetted radius (Wooden Box 4.17

experiments)
The regression results of the surface wetted radius as a function 4.26
of cumulative water infiltration for sandy soil (field experiments)
The regression results of the vertical advances of wetting front as 4.28
a function of the elapsed time (Sandy soil-field experiment)
The regression results of the surface wetted radius as a function 4.38
of the total water applied for sandy loam soil.
The regression results of the vertical advances of wetting front as 4.40
a function of the elapsed time (Sandy loam soil-field experiment)
Nonlinear model summary for surface wetted radius
Coefficients and diagnostic tests for nonlinear regression model 4.69
for surface wetted radius
Nonlinear model summary for vertical advances of wetting front
4.70
Coefficients and diagnostic tests for nonlinear regression model 4.70

for vertical advance of wetting front
Values of the statistical parameters used in comparison for 4.93
surface wetted radius for sandy soil (field experiment)
vertical advance of wetting front for sandy soil (field
experiments)
xvi

surface wetted radius for sandy loam soil
vertical advance of wetting front for sandy loam soil
surface wetted radius and vertical advance of wetting front for
sandy soil
Summary of observed wetted radius for Cecil sandy loam soil 4.1 16
(adapted from Risse et a]., 1989)
Comparison of the observed and predicted wetted radius for 4.1 17
Cecil sandy loam soil
Physical properties of three soils studied by Yitayew et a]., 4.118
(1998)
surface wetted radius for various soil types
Soil hydraulic parameters as fitted with Hydrus-2D
surface wetted radius for sandy soil
vertical advance of wetting front for sandy soil
surface wetted radius for sandy loam soil
vertical advance of wetting front for sandy loam soil
moisture distribution for sandy soil
moisture distribution for sandy loam soil
Values of the statistical parameters used in comparison for solute 4.164
distribution for sandy soil
xvii
Values of the statistical parameters used in comparison for solute 4.1 67

distribution for sandy loam soil
xviii
LIST OF FIGURES
Page
Figure
2.1
General elevation for unsaturated conductivity (K (6)) versus 2.7

suction h for three different soil textured (adopted from James et
a]., 1999)
Constant head and flow rate point source
Front view of the Plexiglass soil containe
Side view of plexiglass soil container
Inside view of the Wooden box
Measuring of surface wetted radius
General view of the experimental setup for sandy soil
General view of the experimental setup for sandy loam soil
Top view of horizontal wetting pattern in sandy loam soil
Side view of vertical wetting pattern in sandy loam
Wetting front movements in the vertical direction in sandy soil
Wetting front movements in the vertical direction in sandy loam 3.15
soil
Schematic diagram showing the physical layout of the trickle 3.26
irrigation system implemented in Hydrus2D, including the
boundary conditions.
Set up of the soil water diffusivity experiment
Wetting front advance at different times for river sand soil under 4.3
trickle discharge rates of 0.75 I/h and 3.40 l/h
The distance of the wetting front from the center of the trickle 4.5
source as a function of elapsed time for river sand
Wetting front advances at different times for sandy loam soil 4.5
xix
under trickle discharge rates of 1.0 and 3.0 l/h

source as a function of elapsed time for sandy loam soil
source as a function of elapsed time (measured by Angelakis et
a1., 1992)
Vertical advance of wetting front as a function of cumulative 4.9
water infiltration as measured by Angelakis et al., 1992
Vertical advance of wetting front versus square root of elapsed 4.1 1
time for river sand and sandy loam soil
Horizontal advance of wetting front versus square root of elapsed 4.12
time for river sand and sandy loam soil
Volumetric moisture distributions in soil at different discharge 4.14
rates for river sand and sandy loam soil
Surface wetted radius as a function of elapsed time for river sand 4.15
under different application rates
Surface wetted radius as a function of the cumulative water 4.16
applied for river sand under different application rates.
Volumetric moisture distribution in river sand at different 4.18
discharge rates:
Final saturated radius as related to emitter application rate for 4.20
sandy soil
The relationship of flux densities, q, versus the reciprocal of 4.21
steady-state radii, llr, for sandy soil
Wetting front advances for sandy soil under trickle discharge 4.22
rates of 2.0, 4.5, 4.8 and 6.0 I/h
Surface wetted radius as a function of elapsed time for sandy soil 4.24
under 1.5,2.1,2.6, 3.5,4.0 and 4.5 I/h application rates
Surface wetted radii as a function of elapsed time for sandy soil 4.25
under 6.01lh application rate.
Surface wetted radii as a function of cumulative water applied for 4.25
sandy soil under different application rates
Vertical advances of wetting front versus elapsed time for sandy 4.27
soil under different application rates
Vertical advance of wetting front versus square root of elapsed 4.29
time at different time for sandy soil under application rate of 2.0,
4.5,4.8 and 6.0 I/h
Vertical advances of wetting front versus cumulative water 4.30
applied for sandy soil under different application rates
source as a function of cumulative water applied for sandy soil
Final saturated radius as related to emitter discharge rate for 4.33
sandy loam soil
The relationship of flux densities, q, versus the reciprocal of 4.34
steady-state radii, llr,
Wetting front advances for sandy loam soil under trickle 4.35
discharge rates of 2.3,2.6,3.O and 4.0 l/h
Surface wetted radius as a function of elapsed time for sandy 4.36
loam soil under application rates of 2.0, 2.5,3.5and 4.0 I/h
Surface wetted radius as a function cumulative water applied for 4.38
sandy loam soil under different application rates
Vertical advances of wetting front versus elapsed time for sandy 4.39
loam soil under different application rates
Vertical advance of wetting front versus square root of elapsed 4.41
time at different time for sandy loam soil
Vertical advances of wetting as a hnction of cumulative water 4.42
applied for sandy loam soil under different application rates
4.3 1
source as a function of cumulative water applied for sandy loam
Surface wetted radius in sandy and sandy loam soil as a function 4.45
of cumulative water applied under different application rates
xxi
Vertical advance of wetting front in sandy and sandy loam soil as 4.46
a function of cumulative water applied under different
application rates under trickle discharge rate of 2.6 I/h
Wetting front position as a function of discharge rate q and 4.47
cumulative infiltration water in liters (the numbers labeling the
lines, Bresler; 1978)
Volumetric moisture distribution in sandy soil at different 4.49
discharge rates of 1.5,2.0,2.5 and 2.7 Ih
Volumetric moisture distribution in sandy soil at different 4.50
discharge rates of 3.5, 4.5,4.8 and 6.0 Ilh
Volumetric moisture distribution in sandy loam soil at different 4.55
discharge rates of 2.3,2.6, 3.0 and 4.0 Ik
Chloride distribution (meqll bulk soil) in sandy soil at different 4.59
discharge rates of 1.5, 2.0,2.5 and 2.7 Ilh
Chloride distribution (meqll bulk soil) in sandy soil at different 4.60
discharge rates of 3.5,4.5,4.8 and 6.0 Ik
Chloride distribution (meqll bulk soil) in sandy loam soil at 4.66
discharge rates of 2.3,2.6, 3.0 and 4.0 Ih
Sensitivity analysis of the surface wetted radius to the change in 4.73
saturated hydraulic conductivity
emitter discharge rate.
average change of soil moisture content
cumulative infiltrated water
Sensitivity analysis of the vertical advance of wetting front to the 4.78
change in saturated hydraulic conductivity.
change in emitter discharge rate
change in average change of soil moisture content
xxii

change in cumulative infiltration water
Correlation of observed and predicted surface wetted radius for 4.84
various application rates for river sand
Observed and predicted surface wetted radius for river sand soil 4.86
under 3.0, 5.5 and 7 I/h application rates
sandy soil (all application rates)
Observed and predicted surface wetted radius for sandy soil 4.89
under 1.5,2.10 and 2.6 I/h application rates
under 3.5,4.0 and 4.5 l/h application rates
under 6.0 Ilh application rate
Correlation of observed and predicted vertical advance of wetting 4.94
front for sandy soil (all application rates)
Observed and predicted vertical advance of wetting front for 4.95
sandy soil under various application rate
sandy loam (all application rates)
Observed and predicted surface wetted radius for sandy loam soil 4.99
under 1.5,2.0, 2.5, 3.5 and 4.0 Ilh application rates.
Correlation of observed and predicted vertical advance of wetting 4.102
front for sandy loam (all application rates)
Observed and predicted vertical advance of wetting front for 4.103
sandy loam soil under various application rate
Observed and predicted surface wetted radius for fine sandy loam 4.107
soil
Observed and predicted vertical advance of wetting front for fine 4.108
sandy loam soil
Observed and predicted surface wetted radius for sandy soil 4.1 12
xxiii
under application rate of 0.5, 0.7, 1.O, 1.4 and 2 l h .

Observed and predicted vertical advance of wetting front for 4.1 13
sandy soil under application rate of 0.5, 0.7, 1.O, 1.4 and 2 1/h
Observed and predicted surface wetted radius for loamy sand soil 4.12 1
under application rates of 2.0, 3.0 and 4.0 l/h
Observed and predicted surface wetted radius for loamy soil 4.122
under application rates of 2.0, 3.0 and 4.0 I/h
Observed and predicted surface wetted radius for silty clay soil 4.123
under application rates of 2.0, 3.0 and 4.0 Ih
4.68
Observed and predicted surface wetted radius for coarser sandy 4.126
soil under application rates of 3.0 I/h
Soil water retention curve fitted by van Genuchten (1980) for 4.129
sandy and sandy loam soil
Soil water diffusivity as a function of soil water content as fitted 4.1 30
by van Genuchten (1980) for sandy and sandy loam soil
Observed and computed surface wetted radius for sandy soil 4.134
under application rates of 1.5,2.0,2.5 and 2.7 Ih
Observed and computed surface wetted radius for sandy soil 4.135
under application rates of 3.5,4.5,4.8 and 6.0 l/h
Observed and computed vertical advance of wetting front for 4.136
sandy soil under application rates of 2.0 and 4.5 I/h
sandy soil under application rates of 4.8 and 6.0 l/h
Observed and computed surface wetted radius for sandy loam 4.140
soil under application rates of 2.3,2.6, 3.0 and 4.0 I/h
sandy loam soil under application rates of 2.3and 2.6 lh
sandy loam soil under application rates of 3.0 and 4.0 l/h
Observed and predicted moisture distribution under application 4.147
rate of 1.5 1/h for sandy soil, elapsed time 320 min
xxiv
rate of 2.0 I/h for sandy soil, elapsed time 360min

rate of 2.5 l/h for sandy soil, elapsed time 290min
Observed and predicted moisture distribution under application 4.15 1
rate of 4.5 Ik for sandy soil, elapsed time 420min
rate of 6.0 llh, or sandy soil elapsed time 340 min
rate of 2.3 I/h for sandy loam soil, elapsed time 270 min
Observed and predicted moisture distribution under application
rate of 2.6 I/h for sandy loam soil, elapsed time 400min
rate of 3.0 Ik for sandy loam soil, elapsed time 350min
rate of 4.0 I/h for sandy loam soil, elapsed time 360min
Observed and predicted chloride distribution under application 4.168
rate of 1.5 I/h for sandy soil
rate of 2.5 Ik for sandy soil
Observed and predicted chloride distribution under application 4.17 1
xxv
rate of 4.5 l k for sandy soil

rate of 4.8 l/h for sandy soil
rate of 6.0 lk for sandy soil
rate of 2.3 Ik for sandy loam soil
rate of 2.6 l k for sandy loam soil
rate of 3.0 I/h for sandy loam soil
rate of 4.0 I/h for sandy loam soil
xxvi
LIST OF ABBREVIATION
CDE= Convection-Dispersion Equation

CRC= Chloride Relative Concentration
MBE= Mean Bias Error
RETC= Retention Curve

RMSE= Root Mean Square Error
SPSS= Statistical Package Social Science Software
SWR= Surface Wetted Radius
U= Theil's inequality coefficient
USDA= United State Department of Agriculture
xxvii
CHAPTER 1
INTRODUCTION
Trickle Irrigation
The relation between population growth and the increased demand for food and
available water resources gives a warning of persistent disorder. The signs of water
shortage are quite clear not only in arid and semi-arid areas but also in part of areas
known as rich in water resources. The work to rationalize water consumption, therefore,
is a priority to reach a permanent water security. The saving of water will enable us to
create a new water resource, otherwise lost to wastage. The rationing of irrigation water
is of paramount importance, especially in view of the fact that agriculture consumption
takes hold the larger water amount. To increase the efficiency levels of irrigation water
consumption several steps should be adopted including the introduction of modern
technology, the development of administrative procedures by irrigation water
management staff and by the farmers themselves and the improvement of the structure
of water distribution and user institutions.
As we move into the 21-st century, emphasis is on alternative water allocations and their
impacts on the environment and economy, the utilization of water resources to meet
agricultural and environmental needs, and the management of water resources to ensure
the integrity, productivity, diversity, and vitality of aquatic ecosystems and their
watersheds. Trickle irrigation is one of the technologies that offer many unique
agronomic, water conservation, and economic advantages needed to address the
challenges for irrigated agriculture in the future. Trickle irrigation has become one the
most popular systems of irrigation, especially in areas where water is expensive. Trickle
irrigation often provides economical and efficient means of crop irrigation by supplying
water and chemicals directly into the soil root zone. Interest in trickle irrigation is
prompted by a variety of factors, including increasing cost of water and limited supplies,
potential for improved water and chemical application uniformity, potential for
improved yields and quality, and great capacity for preventing deep percolation losses of
water and chemicals. Basically, trickle irrigation involves the slow application of water
and chemicals at particular points where the water infiltrates the soil and spreads through
it without appreciable flow of free water on the soil surface.
Under trickle irrigation, reducing the portion of the soil surface that is wetted can save
water and chemicals. This wetted portion of the soil is maintained in a continuously
moist state and the soil volume is never allowed to be depleted or to approach the socalled wilting point. This creates a uniquely favorable soil moisture regime and gives
trickle irrigation a distinct advantage over surface and sprinkler irrigation, especially for
sandy soil of low moisture storage capacity and in arid climates of high evaporation.
In trickle irrigation, the soil serves less as a reservoir for water than for conventional
irrigation because the water that is withdrawn from the root zone is continually
replenished.Thus the soil type and the application rate of water both influence the
moisture and chemical movement patterns. The pattern of infiltration is such that water
flows radially away from the point source under influence of capillary suction forces, as
well as downward under the influence of gravity.
One of the most important considerations in the design of a trickle system is the
volume of wetted soil required to meet the peak water demands of the plant. This
requires knowledge of the soil water and chemical distribution from a point or line
source. In a majority of the trickle systems, emitters apply water and chemicals to a
point on the soil surface or slightly below the soil surface, From this point, the water
will be distributed by means of the water potential gradients to an ultimate wetted
width and depth. This water movement can be characterized as being three-dimensional
and mostly unsaturated.
The final distribution of the water in the soil is primarily a function of the soil texture,
the application rate, and the total volume of water applied. This final water and
chemical distributions determine the volume of soil wetted, the amount of water
applied per emitter and the adequate amount and rate of chemicals. Knowing this it is
then possible to determine the number of emitters required per plant in order to wet a
prescribed portion of the plant's root zone. This is one of the most important design
parameters of trickle irrigation systems.
1.2
Statement of The Problem
Trickle irrigation is a popular practice in many parts of the world. It allows high water
application efficiency continuous, high level of readily available soil water content, and
direct application of fertilizers. A clear advantage of trickle irrigation is that soluble
fertilizers can be applied directly to the plant root zone with both high uniformity and
application efficiency. The expected on shape of water and solute distribution within
the wetted volume from a point source is important to discover before an efficient
design, operation and management of a trickle irrigation system can be performed. It is
within this volume that a plant obtains water to meet its nutrient requirements. To
obtain maximum crop production per unit of water applied, water must be located in a
soil volume where the plant roots have easy access to water and nutrients.
Although much research have been conducted to determine the moisture and chemicals
distribution within the wetted volume under varying rates and amount from trickle
irrigation, the results often seem contradictory and insufficient. Moreover, very little
attention has been paid to estimate the soil water and chemical distributions during
trickle irrigation under field conditions. This will be discussed more thoroughly in the
chapter two.
Objectives of the Study
With the aim of a better understanding of the geometry and distribution patterns of water
and chemicals in soil profile which are of vital importance in the optimum design,
operation and management of trickle irrigation system, this study was conducted with
the following specific objectives:
1. To determine the shape of the wetted soil volume and the moisture distribution
pattern within the wetted volume.
2. To determine chloride distribution in the soil profile as related to the salt content
of the water and trickle irrigation application rates.
3. To develop a simplified model for predicting the geometry of the wetted soil
volume.
To validate and verify the hydrus -2D model with gravimetric field data for water and
chloride movement through the soil profile.
Scope of the study
This study is focused on estimating water and solute distribution pattern under point
source trickle irrigation. There were three main components:
laboratory and field
evaluation of the water and solute movements from a point source trickle irrigation using
gravimetric method, developing a new method to determine the major components of
wetting pattern produced from a point surface trickle irrigation i.e., surface wetted radius
and vertical advance of wetting front and finally evaluating HYDRUS-2D model for
predicting water and solute movement under point source trickle irrigation.
The signficance of this research can be summarized as follow:

A precise knowledge of wetted and chemigated soil volume would ensure
economically sound and environmentally favorable chemigation practices.
Close agreement between the model simulated values and experimental data
would verify the possibility of predicting the movement and distribution of added
water and chemicals.
Organization of the Thesis
To achieve the objectives mentioned above, the study consists primarily of laboratory
and field experiments as well as semi-emprical and computer modeling. The thesis is
arranged in five chapters. After the introduction, Chapter 2 contains the literature
review, where the mechanism of water and solute movement in soil is presented. The
field, laboratory and mathematical studies on water and solute distribution under trickle
irrigation are discussed in detail in this part of thesis. Chapter 3 describes the
experimental designs and methodologies. Chapter 4 presents the results of the
experimental evaluation of moisture and solute movement under point source trickle
irrigation. A comprehensive discussion to these results is presented. Chapter 4 also
contains the description of predictive model for wetting pattern dimensions. The
theoretical considerations, identification of candidate predictors, model optimization,
model verification and testing are conducted and discussed in this chapter. Evaluation of
Hydrus-2D model for water and solute distribution under point surface trickle irrigation
are described in Chapter 4 as well. The conclusions and recommendations of this study
are presented in chapter 5. Finally, the references and appendices are presented at the
end of this thesis.
CHAPTER 2
LITERATURE REVIEW
2.1
Introduction
The knowledge of the shape of water and chemical distribution within the wetted soil
volume is the most important consideration in the design, operation and management of
a trickle irrigation system. This requires understanding of water and solute movement
and leads to knowledge of how water and solute behave in soil and what factors
influence their behavior. This chapter covers topics on water flow mechanism, saturated
and unsaturated flow, in general, with special emphasis on water movement from a point
source. The second section focuses on solute movement in soil including convective
transport, diffusion transport, dispersive transport and the convection-dispersion
equation, in general, with special emphasis on solute movement from a point source.
2.1.1
Water Flow Mechanism
Soil water flow is caused by a driving force resulting from a potential gradient, that flow
takes place in the direction of decreasing potential, and that the rate of flow is
proportional to the potential gradient and is affected by the geometric properties of the
pore channels through which flow takes place. These principles apply in unsaturated, as
well as saturated soils (Taylor and Ashcraft 1972).
The moving force in a saturated soil is the gradient of a positive pressure potential. On
the other hand, water in an unsaturated soil is subject to a sub atmospheric pressure
potential. The gradient of this potential likewise constitutes a moving force. Matric
suction is due to the physical affinity of water to the soil-particle surfaces and capillary
pores. Water tends to be drawn from a zone where they are thinner and from a zone
where the capillary menisci are less curved to where they are more highly curved. In
other words, water flows spontaneously from where matric suction is lower to where it
is higher.
When suction is uniform all along a horizontal column, water soil is in a state of
equilibrium and there is no moving force. Not so when a suction gradient exists. In that
case, water will flow in the pores which are water filled at the existing suction and will
creep along the hydration films over the particle surfaces, in a tendency to equilibrate
the potential.
The moving force is greatest at the wetting front zone, where water invades and
advances into originally dry soil. In this zone, the suction gradient can amount to many
bars per centimeter of soil. Such a gradient constitutes a moving force thousands of
times greater than the gravitational force. Such strong forces are sometimes required for
water movement to take place in the face of extremely low hydraulic conductivity,
which a relatively dry soil often exhibits. The most important difference between
unsaturated and saturated flow is in the hydraulic conductivity. When the soil is
saturated, all of the pores are water filled and conducting, so that continuity and hence
conductivity are maximal. When the soil de-saturates, some of the pores become air
filled and the conductive portion of the soil's cross-sectional area decrease
correspondingly. Furthermore, as suction develops, the first pores to empty are the
largest ones, which are the most conductive, thus leaving water to flow only in the
smaller pores. The empty pores must be circumvented, so that, with de-saturation,
tortuosity increases. In coarse-texture soils, water sometimes remains almost entirely in
capillary wedges at the contact points of the particles, thus forming separate and
discontinuous pockets of water. In aggregated soils, too, the large inter-aggregate
spaces, which confer high conductivity at saturation, become barriers to liquid flow
from one aggregate to its neighbors.
For these reasons, the transition from saturation to unsaturation generally entails a steep
drop in hydraulic conductivity, which may decrease by several orders of magnitude
(sometimes down to 1/100,000 of its value at saturation) as suction increases from 0 to 1
bar. At still higher suctions, or lower wetness values, the conductivity may be so low
that very steep suction gradients, or very long time, are required for any appreciable
flow to occur. At saturation, the most conductivity soils are those in which large and
continuous pores constitute most of the overall pore volume, while the least conductive
are the soils in which the pore volume consists of numerous microspores. Thus a
saturated sandy soil conducts water more rapidly than a clayey soil. The very opposite
may be true when the soils are unsaturated. In soil with large pores, these pores quickly
empty and become nonconductive as suction develops, thus steeply decreasing the
initially high conductivity. In soil with small pores, many of the pores retain and
conduct water even at appreciable suction, so that the hydraulic conductivity does not
decrease as steeply and may actually be greater than that of a soil with large pores
subjected to the same suction.
2.1.2 Saturated Water Flow
Soil is considered saturated when all pores are filled with water. Saturated water flow is
rather rapid, and occurs mainly by draining the gravitational water occupying the pore
spaces between the soil particles. Gravitational force is the main driving force for
saturated water flow. Darcy's law, the fundamental equation describing water movement
in soil, relates the flow rate to these two factors, (Rattan, 2004). Mathematically, the
general statement of Darcy's law for vertical, saturated flow is:
where: the flow rate V/At is what soil physicists call the flux density, i.e., the quantity of
water V moving past an area A, perpendicular to the direction of flow, in a time t. The
vertical saturated hydraulic conductivity Ks is the reciprocal, or inverse; of the resistance
of the soil matrix to water flow. The term dH/dZ is the hydraulic gradient, the driving
force causing water to move in soil, the net result of all forces acting on the soil water.
A distinction is made between saturated and unsaturated hydraulic conductivity.

Saturated flow occurs when the soil water pressure is positive; that is, when the
soil matric potential is zero (satiated wet condition). In most soils this situation
takes place when about 95 percent of the total pore space is filled with water.
The remaining 5 percent is filled with entrapped air. If the soil remains saturated
for a long time the percent of the total pore space filled with water may approach
100. Saturated hydraulic conductivity cannot be used to describe water

movement under unsaturated conditions.
2.1.3 Unsaturated Water Flow
The fundamental driving forces in both saturated and unsaturated flow are the potential
gradient and hydraulic conductivity. As stream of water is passed through the
unsaturated soil matrix, the incoming water replaces the air present in the soil pores; it
increases the total water volume of water inside the soil, thus increasing the moisture
content (0) of soil. This agrees with the fundamental of continuity equation, which states
that the difference in the inflow and outflow rate is equal to the change of water storage
in soil. The gradient causing flow in unsaturated soils is of negative pressure potential.
The flow paths in unsaturated flow are more tortuous as several pores are filled with air.
The hydraulic conductivity in an unsaturated zone can vary by as much as four to five
orders of magnitude. Some of the differences in saturated and unsaturated flow are
summarized in Table 2.1.
Water has an affinity for soil particle surface and capillary pores, which result in matric
suction. When matric suction (h) is uniform throughout the soil matrix, the soil system
remains in equilibrium or at a steady state conditions. When matric suction (h) is not
uniform within the soil profile, water moves from a site of higher (h) till the system
attains equilibrium.
Table 2.1 Summary differences between saturated and unsaturated flow (adopted from
Rattan and Shukia, 2004)
Parameter
Saturated flow
Unsaturated flow
Water content
Constant
Variable over space and time
Air content
Zero (close to zero)
Variable over space and time
Potential gradient
Positive and constant
Negative and variable
Hydraulic conductivity
Maximum, constant
Low and variable
Vapor flow
None
Possible provided temperature

gradient also exist
Water flow
Steady
Steady as well as unsteady
Flow paths
Continuous
Tortuous
Continuity equation
Inflow=outflow
Inflow =outflow+ source or

sink, of water
Flow descriptions
Darcy's law
Darcy-Buckingham
Richards equation
Flow parameter
Ks
K(0)
equation
Edgar Buckingham (1907) suggested a modification of Darcy's law in order to describe

the flow through soil. This modification was based on two assumptions:
1. The driving force for water flow is the sum of the matric potential and gravitational
potentials.
2. The unsaturated hydraulic conductivity of soil is a function of water content

(K=K (8)) or matric potential (K=K (h)), where: (0) is soil water content and h is soil
matric potential (James et al., 1999).
For unsaturated flow the Buckingham-Darcy flux law may be expressed as:
Where: V H is the hydraulic head gradient which may include a suction (h) and
gravitational (z) component (H=h + z), and K (h) hydraulic conductivity as a function of
suction head. Unsaturated hydraulic conductivity is a nonlinear function of water
content or matric potential.
A general relation for unsaturated hydraulic conductivity versus matric potential for
three different textures of soil (Clay, Loam and Sand) classes is shown in Figure 2.1.
Loam
Sand
Figure 2.1: General elevation for unsaturated conductivity (K (6))versus suction h for
three different soil textured (adopted from James et al., 1999)
The sandy soils have higher conductivity than clayey soils at saturation conditions
because they contain large pore spaces, which are filled and conduct water. Conversely,
sandy soil at unsaturated conditions produce a dramatic decrease in hydraulic
conductivity, while clayey soils at the same matric potential have higher K values
because the latter soils retain more water.
2.1.4 Continuity Equation
The mass conservation law, expresses in the equation of continuity, states that if the rate
of inflow into a volume element is greater than the rate of outflow, the volume element
must be storing the excess and increasing its water content (and, conversely, if outflow
exceeds inflow, water content must decrease with time).The equation of continuity in
multidimensional systems according to (Rattan and Shukla, 2004) can be written as
~ ) t is the time (T)

where 8 is the soil water content ( L ~ L -and
In some situations, positive and negative accumulations of a component are due to not
only outside the system. For the flow of water in soil the proper continuity equation
(Koorevaar et al., 1983) is then
where S is an extraction term.

2.1.5 Richard's Equation
The flux density equation and the continuity equation must both be satisfied in solving a
particuIar water flow problem. Substitution of equation (2.2) into the continuity equation
(2.4) yield the combined flow equation that can be written as
This equation, along with various alternative formulations, is known as the Richards
equation. Note that the extraction term is also considered a term dependent on the actual
soil flow equation
where K,(h) and K,(h) are the unsaturated hydraulic conductivity at horizontal and
vertical directions, respectively.
In equation (2.6) time and space dimension are independent variables, whereas soil
water pressure and volume water content are the dependent variables. The equation can
be simplified introducing soil differential capacity defined as
C(L-') is called soil water capacity and its use provides the option of expressing the
Richards equation in either an h-based form , a 8- based form, or a mixed- form.
Applying equation (2.7) into equation (2.6) the general flow equation in twodimensional form and h- based can be expressed as
ah
a
ah
C(h) - = -[Kx(h) -1
at
ax
ax
a [Kz(h)(-ah
+az
az
- 111 - S(h)
The 0- based equation can be written as
where D(0) -K(O)/C(B) is the unsaturated diffusivity (L~T-'),D,(e) and D,(B)are the
horizontal and vertical unsaturated diffusivity, respectively.
The most commonly used form of Richards equation is the h- based form. Its advantages
are that it can be applied to both saturated and unsaturated conditions, as well as to
layered soils, where h is generally continuous but t3 may not be (Hanks and Ashcroft,
1980).
2.2
Water Movement under Trickle Irrigation
Water moves downwards vertically and sideways horizontally from the source for bare
soils (Hillel, 1982). This downwards and sideway movement of water forms a wetted
zone where the main root zone of the plant develops (Benami and Offen, 1983). The
wetted zone is of practical importance as it represents the amount of soil water and
chemicals stored in the root zone. Its depth and width should at least equal to if not
slightly bigger than rooting depth and width of the plant root system. Therefore, the
volume and the pattern of the wetted soil becomes an objective rather than an end result
of the design process of an irrigation system (Zur, 1996). This wetted volume will be a
function of the soil water distribution which is determined by soil physical properties,
the application rate, and the total volume of water applied. The wetted area is largest for
heavy soil and smallest for light soil. Similarly, the shape of the wetted area for heavy
soils has more spread horizontally than vertically and vice versa for light soils (Zur,
1996) Many field and laboratory studies related to water distribution patterns in soil
have been conducted world wide. These studies have provided a good understanding of
the effects of trickle irrigation on water distribution patterns in soil. However, there is a
need for continuing research on the effects of application rates and volumes on the soil
water distribution in a variety of soils and conditions.
2.2.1 Field Studies
Roth (1 974) developed a simple procedure, which reasonably predicted the wetted soil
volume, lateral movement and vertical movement of water from a point source. The
underlying assumptions were that the soil moisture in wetted profile is at field capacity
and trickle irrigation was defined to exclude large flow rates which would cause
excessive ponding and surface runoff or small flow rates which would not increase the
soil moisture so it can approach field capacity. Moisture content in excess of field
capacity would be lost to deep percolation because of gravity. This procedure was
verified with field tests on a fine sand soil and in laboratory on a Gadsden clay soil. He
also found that the moisture movement in soil from a trickle source is more a function of
volume applied than the application rate. Higher flow rates can cause greater moisture
content in the soil during the application but the values decrease and approach water
contents from lower rates if given similar redistribution periods. He expected that the
procedure for predicting wetted soil volume, lateral movement and vertical movement
can be used by both designers and managers of trickle irrigation system.
Horizontal water movement in two soils was studied by Goldberg et al., (1976) using
five different flow rates (1.0, 2.0, 4.2, 9.2 and 18.4 liters per hour) and five total
volumes of water (12, 24, 48, 96 and 192 liters). The two soil types were Kfar Hayarok
and Hatzerim and their properties are shown in Table 2.2. The point source with the
greatest discharge (1 8.4 llh) caused surface runoff. This rate is not suitable for practical
use. The ponded diameter of the water under the point source increased 3-fold as the
flow rate increased from 1.0 to 9.2 l/h for each soil type with equal application volume.
However, these studies showed that the maximum lateral water movement increased less
than lo%, indicating that lateral movement is more a function of the total volume of
water applied than of the flow rate.
Table 2.2: Properties of the soils used in moisture movement studies (adapted from
Goldberg et al., 1976)
Property
Kfar Hayarok
Hatzeerim
Bulk density (gm/cm3)

Volumetric moisture at saturation (%)
Hydraulic conductivity (cmlmin)
Clay content (%)
Lafolie et al., (1989) analyzed the water flow under trickle irrigation on stratified and
anisotropic soils and compared the data from a field experiment using loamy clay soil
from 24 field plots irrigated at different rates and frequencies. A previously proposed
numerical model for predicting soil profile ,water content distributions during trickle
irrigation on stratified and anisotropic soils was used. Results for a typical set of
experiments involving four hours of irrigation and twenty hours of redistribution have
been presented in this treatise. Many discrepancies were noted when the observed water
contents were simulated with the model and a majority of them could be attributed to the
inadequate characterization of unsaturated soil hydraulic conductivity. Moreover other
problems like formation of thin soil crust on soil can be explained with the help of
heterogeneous behavior of fine textured soil. The inclusion of soil crusting and soil
anisotropy resulted in improved predictions of soil water content. The conclusion drawn
from this paper can be summarized as a need for accurate description of soil hydraulic
properties for reliable prediction of soil water content during trickle irrigation.
Jobling (1974) presented field data from several different soil types. In general, the total
size of the wetted area is about the same regardless of the point source discharge,
provided the same total amount of water is applied. It was also noted that the wetted soil
volume could be slightiy smaller with higher water contents for the faster flow
compared to the slower flow rates of equal water volume applied. The data indicated
that the vertical and horizontal movement was approximately proportional to the volume
of water applied, raised to the 0.33 power for a heavy soil. For coarse soils, the
horizontal movement was roughly proportional to the volume of water applied raised to
the power of 0.2 and vertical movement is approximated by the water volume applied
raised to the power of 0.6. To get greater lateral movement on coarse soils, he
recommends that higher flow rates can be used to increase the pond area under a point
source for a short time which in turn increases the lateral movement
Muluneh and Akbar et al., (1995) conducted field plot study to evaluate water and solute
movement from a point source. The purpose of the study was to understand chemical
and water distribution in the vadose zone. The study revealed a clear relationship
between the application rate and the shape of the wetted soil zone. For the same
application volume, an increase in rate resulted in an increase in the wetted horizontal
area and a decrease in the wetted soil depth. The study also found solute concentration
increased with high input concentration, applied volume, and the application rate. Larger
wetted and chemigated soil volume was found for higher volumes of applied water.
Al-Qinna et al., (2002) studied the influence of emitter discharge on the wetting pattern
and distribution in arid and semi-arid soil suffering from surface crust formation. Four
application rates were tested with a constant volume of water applied through a
controlled Palm-type emitter using eosin red indicator for the delineation of the wetting
front advance. The shape of wetting patterns on crusted soil varied considerably from
that of a typical bulb shaped to that of a "V-shaped" cross-section depending upon the
application rate. A regression analysis indicated that increasing the application rate
increases horizontal advance by increasing the increments but reduces the vertical
advance of the wetting front by decreasing the increments owing to the negative effect
of the soil surface crust.
Risse and Cheness (1989) conducted a series of field experiments to determine the
differences in the size and shape of the wetted area at two different water application
rates (3.78 and 7.57 l/h) with a total volume of 37.8 liters. The soil at the site was
classified as Cecil sandy loam with a relatively deep A horizon. Their results showed
that the average wetted radius for the 3.78 l/h treatment was 37.1 cm while the wetted
radius for 7.57 l/h treatment was 35.3 cm. The difference between these average wetted
radii was statistically insignificant at 5% level. Although no tests could be run to
determine if there was any difference in the wetted depths of the two treatments, it did
appear that the 3.78 I/h treatment wet the soil to a slight greater depth than the 7.57 l/h
treatment. This shows that a slower rate tends to have a greater vertical movement.
Yitayew et al., (1998) investigated the saturated water entry zone for loamy sand, loam
and silty clay soils under point source trickle irrigation with application rates of 2.0, 3.0
and 4.0 l/h. This field study indicated that increase in the size of saturated radius was
rapid at the start and slowed down as time increased. The saturated water radius
increased with the application rate for specific soil. They also found that the saturated
radius was smaller for the light soil (loamy sand with high saturated hydraulic
conductivity) than for medium and heavy soil (with lower saturated hydraulic
conductivities). The results showed that the spread of wetting front which is a distinct
boundary that exists between the wetted zone and the dry zone increased with increasing
of the flow rate.
Goldberg et al., (1971) examined the effect of trickle irrigation on distribution of roots,
water, and minerals in three-dimensional soil profile. They reported that it is possible to
somewhat increase the lateral spread of the wetting front by increasing the emitter
discharge rate or the amount of water applied. They also concluded that if the soil's
infiltration capacity is high and the lateral movement of water is low, it is essential to
use a closer emitter spacing.
2.2.2 Laboratory Studies
Limited data have been colleted on the water movement and distribution in soil profile
from a point source under field conditions. The amount of time and effort required to
obtain these data has been the greatest deterrent. To use these data to verify the
mathematical models would also require the knowledge of the unsaturated hydraulic
conductivity relationship. Thus, most studies have been conducted in laboratory
conditions where a uniform soil is packed into an observation box with one side of
plexiglass. The water movement is observed with time and moisture content determined
by sampling along a radius at various times.
Soil water movement was studied by Haq (1973) in the laboratory using a plexiglass
box. The soil type he used was Hawaiian clay of volcanic origin with a bulk density of
0.75 g/cm3, field capacity of 50 % and saturation of 63 % by volume. His studies
involved applying water at rates of 1.5, 4.0, 5.0, 10 or 16 cclmin for various duration of
up to 90 minutes. It appears from his results that the water movement were the same for
equal volumes of water applied at different flow rates. The moisture content in the
wetted profile had a slight gradient from the source to the wetted front and
approximately equal to field capacity. If the initial moisture content was increased, the
water would move faster through the soil and attain the same moisture level as a similar
test where the initial moisture content was less.
Water movement was observed by Hachum et a1 (1 979) in the laboratory for the two soil
types. The two soil types were a loamy sand and silt loam with bulk densities of 1.62
and 1.4 grnlcc, respectively. Each soil was placed in a narrow box to simulate a line
source. They observed that for similar total of water applied, as the discharge rate
increased, the depth of wetted zone decreased while the lateral spread increased. They
suggested that this was a direct result of the increased in the width of saturated soil strip
at the surface in the vicinity of the trickle source. This was particularly more noticeable
with the silt loam soil compared to the loamy sand soil. After turning water off, the
horizontal movement almost stopped while the vertical advance of the wetting front due
to gravity continued at the same rate for loamy sand soil. The horizontal and vertical
advance continued equally when the application was terminated for the silt loam soil.
The two dimensional shape of the wetted fiont was approximated as a semi-ellipse.
Bar-yosef and Sheikholslami (1976) studied moisture movement in the laboratory using
a sandy soil and two discharge rates of 0.25 and 2.5 llh. One liter of water was applied
for each flow rate and the moisture movement and water content were measured 48
hours after the irrigation was terminated. It was found that the vertical movement
increased from 13 to 3 1 cm and the horizontal movement decreased from 19.5 to 15.5
cm as the flow rate increased from 0.25 to 2.5 l/h. This indicated that horizontal
movement did not change appreciably with increasing flow rate. It appeared that
redistribution after the application slowed the horizontal moisture movement and
resulted in lower water content values in the higher flow rate profile compared to the
lower flow rate profile.
Mostaghimi et a1 (198 la, 198lb, and 1983) did a series of studies in a Drummer silty
clay loam. Their first study found that the wetting pattern changed as a function of
application rate. They showed that an increase in the application rate resulted in an
increase in the vertical component of water movement and a decrease in the horizontal
component. In their second study they concluded that a pulsed application resulted in
less vertical water movement than did a continuous application. Their final study
compared soil water distribution under bare soil to those under cropped conditions. They
observed a nonuniform distribution under the cropped condition which resulted in a
smaller total volume of wetted soil.
Li et a1 (2004) studied the wetting patterns from a surface point source for sandy soil. A
15- wedge-shaped plexiglass container was used to represent one twenty fourth of the
complete cylinder. The variables affecting water and solute distributions, including
application rate and applied volume were investigated and their effects were presented
on a basis of a completely cylindrical system. In the experiments, the apparent
cylindrical application rate was varied from 0.5 to 2.0 l/h, the apparent cylindrical
applied volume from 5.9 to 7.0 liters. Monitoring of the wetting front movement in a
sandy soil revealed that the increase in the surface wetted radius and in the vertical plane
with water volume applied can be represented by a power function with power values of
about 0.3 and 0.45, respectively. Similar values reported for a clay soil (Bar-Yosef and
Sheikholslami's, 1976) which suggested that these relations could be applied to a wide
range of soil textures, from sand to clay. The effect of water application rate on water
distribution pattern in the soil was analyzed. For a given volume applied, increasing the
application rate allows more water to distribute in the horizontal direction, while
decreasing the rate allows more water to distribute in the vertical direction.
Armstrong and Wilson (1983) conducted a study on the water distribution in Lakeland
sand. Their results showed that the size and shape of the wetted zone beneath an emitter
was more a function of the total amount of water applied than the rate of application for
the application rate of 3.9 to 16.1 l/h and application volumes of 3 1 to 160 liters.
Youngs et al., (1999) studied the surface ponding of coarse-textured soils under
irrigation with a line of surface emitters. Experiments were conducted in a glass-fronted
tank. Two soil materials were used, a coarse sand and a fine sand. These initially air-dry
materials were packed to a uniform packing density into the tank. A peristaltic pump
was used to feed water at a constant rate from a line of ten identical capillary tubes on to
the surface of the sand in the centre of the tank. The extent of surface ponding and the
progress of the wetting front advance pattern were observed through the glass front of
the tank for a range of steady flow rates. Flow rates used in case of the coarse sand
ranged between 0.85 and 8.83 m3/mld, and for the fine sand between 0.168 and 0.885
m3/mld. The results showed that the ponded area on the soil surface fairly quickly
approached a constant width for all application rates although the wetting front
continued to advance laterally as well as vertically.
2.2.3 Mathematical Model Studies
Mathematical and numerical models have been developed to describe saturated and
unsaturated flow through soil under laboratory conditions. Often these models are based
on assumptions, which are unlikely to be met in the field, or they require soil property
inputs, which are difficult to obtain.
Bresler et al., (1971) compared the theory of transient infiltration from a trickle
irrigation source with the experimental results. As a part of experimental verifications,
loamy sand was tested under conditions similar to the conditions of two dimensional
flow models. Another comparison was made under conditions similar to those of
cylindrical flow model. The functional relationship between water diffusivity and water
content was estimated using computer technique that complements the conventional
diffusivity determination methods. This study indicates that despite discrepancies
between the experimental and theoretical results, especially at low application rates, the
coherence is sufficient for practical applications of the theory. The study also gives the
probable reason for discrepancies in some cases such as inadequate assumptions, lack of
precision in estimation of soil water parameters, inaccurate boundary condition
definition and hysteresis during the infiltration process. The hystesesis affect may form
due to the conditions at the soil surface. These soil conditions are more complicated than
those dealt with in theory especially in the initial stages and it becomes hard to ascertain
the factor causing the discrepancy between the observed and theoretical results.
However, in general, hysteresis becomes prominent when trickle irrigation discharge is
small. The results show an increase in the horizontal wetted area and a decrease in depth
of wetted soil with an increase in trickle discharge rate. This can be possibly accounted
by the effect of trickle discharge rate on the size of the water entry saturated zone. This
zone becomes larger with increase in trickle rate.
Levin and Van Rooyer (1979) compared laboratory and field studies to the
mathemathical model developed by Brandt et al., (1971). They used sandy soil
containing 98% sand and a bulk density of 1.5 g/cm3. Their tests used flow rates of 2
and 4 l/h continuously and 2 l/h pulsed in 30 minute increments. The volume of water
applied to all tests was 12 liters and measurements of soil water movement and water
content were taken 12 hours after each test was initiated. Total volume of soil wetted in
each test was determined and average amount of water found in the soil profile was 11.9
liters. The amount of water in the wetted soil volume was also predicted using the
mathematical model. The average amount of water found in this case was 11.5 liters
compared to the actual amount of 12 liters applied. Their results indicated that greater
lateral and vertical movement resulted from higher flow emitters compared to lower
flow emitters 12 hours after the start of each test. They also suggested that greater
amounts of water could be utilized by the plant if low or pulsed flow rates were used
compared to high discharge rates. Further they recommended that lower flow rate
emitters or high flow rate emitters that are pulsed would result in more water remaining
in the upper soil profile to be utilized by the plant.
Schwartzman and Zur (1986) 'developed a semi-empirical approach to predict the

wetting patterns under a point or line source. They assumed that the extent of wetting
depended on the emitter discharge rate q, saturated hydraulic conductivity k,, and total
amount of water in the wetted volume V, to the vertical depth of wetting z,and lateral
diameter of the wetted volume d (measured at its widest point). Based on results
presented by Bresler et al., (1971) for surface dripper (point source) they estimated the
following relationships:
Consistent units must be used in these approximations such as d and z [m], q[m3/s].
k,[m/s] and v [m3]
These relations were then used in equations which could predict the depth and width of
the wetted soil volume at various flow rates and durations. They demonstrate that this
method could accurately predict the extent of wetting and went on to propose a
procedure for selecting the optimal emitter spacing based on both physical and cost
considerations. The worth of this procedure to engineer would depend on the availability
of the empirical relations needed. While the time and expense involved in performing
actual field experiments would probably be too great for the process to be practical,
simulations using a proven computer program could be used if the input data is easily
obtainable.
Ben-Asher et al., (1986) investigated the infiltration from a point trickle source in the
presence of water extraction using an approximate hemispherical model.
They derived analytic expressions for the position of the wetting front. For infiltration
from a point source without water extraction they found that
where: R(t) is an effective radius of a wet hemisphere at time (t), I, is discharge rate, t is
the time of infiltration, and A0
0,,- 8, ,where: 8,, is average water content, and 8, is
initial water content. This approximation appears to work best for fine-textured soils,
dry initial conditions, and relatively short times. For longer time (and for wetter
conditions) the approximation becomes invalid as gravity effects distort the shape of the
wetted volume from hemispherical to semi- ellipsoidal.
Dasberg and Or (1999) indicated that the elapsed time at which the effect of gravity on
the flow process become dominant (equal capillarity) and begins to distort the
hemispherical approximation is related to the soil properties (and to the change in water
content within the wetted soil volume). An estimate for this "gravity time" tgravfor flow
from a small pond is given by (Philip, 1986):
It is assumed that a typical change in volumetric water content (for relatively dry initial
conditions) is about Ae= 0.125 cm ~ m ('which
~ represents the difference between water
content at field capacity and wilting point for many soils). The factor of a [Eq. (2.14)]
may be increased for wetter initial conditions. This estimate of gravity time may be used
to predict whether the hemispherical approximation is reasonable for the soil type and
irrigation time t, ,They estimate that for soil types and irrigation times with t gr,/ti > 10,
the hemispherical geometry is like to provide reasonable estimates for the dimensions of
the wetted soil volume (a large ratio is indicative of strong capillarity effect, resulting in
spherical soil wetting). A similar approximation for spherical wetting from buried
emitters (volume of wetted soil V= 4m3/3) may be used with the same criterion for
gravity time (t pav).
Gupta et al., (1995) conducted experiments to evaluate the effects of application rate on
wetted regime developed near the emitter. The following empirical relationship was
established to describe the wetted area as a function of q and V:
1
(8.257q - 1 . 2 2 4 ) ~ ; + 54.46
where: D is wetted diameter of soil surface (cm).The experiment data indicated that the
wetted surface area increased with increases in the trickle application rate and amount of
the applied water. They concluded that the trickle application rate and the volume of
water applied control the size of lateral spread of water on the soil surface, and the radii
of wetted regime per unit application rate decrease with increase in the application rate
indicating that the water spread more on soil on soil surface at the lower application rate
compared to the higher rates.
Healy and Warrick (1988) developed numerical solutions in a dimensionless form of

Richards'equation for flow from a point source (on the soil surface) to estimate the
empirical coefficient of "generalized" equations for the shape of the wetted soil volume.
The resulting equations and coefficients are applicable for a wide range of soil types and
source discharge rates. The general form of equations is:
Where: ri*is the dimensionless distance from the source (ri*=avgri where ri the physical
distance and avga length scaling factor associated with the water retention model of
van Genuchten (1980), i is an index of the direction of r: taken vertically (i=l; r= 0),
diagonally (i=2; r=z), and along the soil surface (i=3; z=0), z is dimensionless time
(F [av, k, t]/Ae), the coefficients Ai, Bi and Ci are coefficients dependent on the
parameter n (a van Genuchten retention parameter), and dimensionless emitter discharge
q*=( a2vgq/ks).The dimensionless wetted soil volume is also given in an equation
similar to Eq.(16). Healy and Warrick presented an extensive table for the coefficients
that cover most practical situations (soil type and emitter discharge) and may be useful
for design purpose (when soil hydraulic properties are known).
Angelakis et a1 (1993) did a comprehensive work in the area of prediction of soil water
distribution, under a circular trickle source, using numerical techniques such as timedependent linearized infiltration theory, finite element solution to transient water flow
equation and analytical methods like effective radius of a wet hemisphere model, Healy
and Warrick (1 988) generalized solution. The wetting front and soil-water distribution in
two- dimensional soil profiles under different frequencies and discharge rates of water
were measured for Yo10 clay loam and Yo10 sand, which have different textural
characteristics. Both the horizontal and vertical wetting fronts are important for efficient
design, operation and management of a trickle irrigation system, so the experiments
were performed to get a two-dimensional soil water distribution. The results showed that
for the homogenous soil, predictions of the shape of the advancing fronts, done by
numerical and linearized model were reasonably precise. Predictions with the linearized
model for clay loam or sandy profiles were adequate for the horizontal component,
however there was an over or underestimation in the vertical component of the wetting
front. The possible reason for discrepancy can be the inadequacy of the assumption that
k (0) was a unique linear function. Predictions with numerical simulation were better
than with linearized model, for both the soils, especially in the vertical direction.
However, the numerical model overestimated the horizontal component in sandy soil.
The effective hemispherical model, in which the radius of the wet hemisphere expands
proportional to tIJ3, under-predicted the vertical wetting front and over-predicted the
horizontal wetting front, for both soils. The effect of gravity in distorting the spherical
symmetry was invisible, especially under low application rates and long times. On the
other hand, the Healy and Warrick (1988) generalized model showed good agreement in
vertical and diagonal positions, especially under high application rates with clay soil.
However significant overproductions were observed in the reason for errors between
calculated and measured results can be attributed to the assumption that the source is in
the point condition rather than as a disc source. The soil water tensions computed by
using the linearized solution were higher than those observed, but the finite model
showed the opposite estimation. Measured soil water contents were also slightly greater
than those predicted by these models. The results were generally better for clay loam
than for sandy soil as there exists a better approximate linear function k (0) and a
varying size of the surface source used between two application rates linearized and
numerical models.
Risse and Chesness (1989) simplified design procedure to determine the wetted radius
for a trickle emitter. This design was based on Dasberg and Bresler's (1985) method for
predicting the wetted radius from a point source, by using soil texture, emitter flow rate
and soil water depletion potential as the only inputs required. They used Bresler (1978)
equation to predict two dimensional water flows:
-2
Pu = - + (
an-
+- 1 OOOQ )0.5
(a2n2>
As
where:
p, the ultimate radius of the saturated zone of water ponding, cm
a= a constant expressing the relative importance of gravity and capillarity to

water of the soil, cm-'
Q= the emitter discharge rate, L h ,
Ks= the saturated hydraulic conductivity, cmlh.
Generally the results from the prediction equation and measured wetted radius value
showed the over all-average difference between the predicted and observed values using
individual inputs were within - 19% and 1 1% of the measured values.
Using experimental data, the continuity and the cumulative infiltration equations, a new
expression for predicting the wetted soil depth (Eq 2.1 3) was described by Hammami et
al. (2002).
Where: Zfo
the maximum wetting front computed from the surface, along the
symmetry axis.
Rf (t)
= the
wetting front radius measured from the point source, on the soil surface, cm
Kf = the soil hydraulic conductivity in the wetting volume, cmfh

Oi = the initial soil moisture content (decimal) and Of the soil moisture content behind the
wetting front ( decimal)
t = infiltration time, h
To verify Equation 2.18, a series of infiltration experiments on sandy and silty clay soils
were carried out in the laboratory using a semi-cylindrical container, 1.20 m high and
1S O m in diameter. A glass plate, constituting the straight vertical face was installed to
observe the wetting front advance. The discharge flow rates used were 1, 2 or 4 l/h with
the infiltration time 12 hours for all cases. The wetting front radius Rf (ij was measured
from the point source, on the soil surface and the "maximum" wetting front depth zf(0
was measured (through the transparent side) from the surface along the symmetry axis.
The results of zf
(0 predicted
values as a function of Rf (ij were comparable with
measured values, even though the relative deviation between predicted and measured
was in the range between -10% to + 10% for sandy soil. In silty clay soil, the relative
deviation was in the range between -5 % to +5%.
Clothier et al., (1982) developed a comparatively simple analytical solution of the flow
equation considering the pressure head gradient across saturated zone at the soil surface.
The model may be expressed in the following form:
Where:
is the width of the strip wetted by a trickier, and 6 is the average surface
pressure head across the saturated zone and varies from 1.0 to 2.5 for clay-loam soils to
sandy soils.
Howell et al., (1983) have given a simple solution by approximating the wetted volume
of soil as a hemisphere. They assumed that the soil volume was wetted from initial
uniform volumetric water content (Oi) to final water content (Of). The solution is of the
following form:
where: q is the volumetric rate of application (m3/h), t is application time (h), r, is

wetted radius (m). The final water content will correspond to a hydraulic conductivity of
1 mlday and can be very different for different soils depending on their physical and
chemical characteristics (Jensen, 1983).
By using dimensional analysis technique Zazueta et al. (1995) developed a simple

equation to estimate the diameter of the wetted-bulb resulting from infiltration into a
dry, deep, uniform soil.
where: q is the discharge rate of emitter, 1 h-'; rd is the root depth, cm; Id is the basic
infiltration rate 1 h-'; D is the diameter of the wetting bulb. The equation compared well
with other equations determined from analytical and regression analysis.
Philippe et al. (1996) reviewed four existing analytical theories [Wooding (1968), Raats
(1971), Lockington et al. (1984) and Philip (1985)l that provided the flee-water pond
radius, r,, under trickle source. They have shown that these different theories do indeed
give different r, values and they indicated that it was due to the different hypotheses
used to describe the geometry of the saturated region. Their results have shown that
Ratts (1971) solution appeared to give the best estimates of the free-water pond radius
across all the discharge rates in a field study. They also noted that the free-water pond
radius varies quite slowly with the dripper discharge rate. The study suggested that more
efficient means to reduce ponding, and increase the volume of soil wetted by drip
irrigation, is to increase the number of emitter and then run each of them at a lower rate.
Multidimensional infiltration and subsequent redistribution of water in the soil can be

described by the Richard's equation that combines ~ a r c ~law
' s with conservation of
mass as:
where: 8 is the volumetric water content; h is the matric head; H is the total head,
k(h) is the unsaturated hydraulic conductivity; and V is the del operator (gradient or
divergence). For flow from point source, analytical solution for Eq. 2.22 can be
obtained by quasi-linearization (Philip, 1971) which considers the exponential model
for unsaturated hydraulic conductivity proposed by Gardner, (1 958):
in which k, is the saturated hydraulic conductivity (LT-') and a is a constant

characteristic of the soil
(L-I).
The quasi-linearization also uses the matric-flux
potential transformation defined by Philip (1 971) as:
2.24
where: 0 is the rnatric flux potential (L~T"),h,=-o3 is the limit of the integration.
For steady state flow from point source, i.e., dOldt=O, the quasi-linearization yields a
second - order differential equation in term of 0 :
2.25
A solution for Eq. 2.25 considering a buried point source of strength q at r= 0 z= 0 was
derived by Philip (1968) for a cylindrical coordinate system as :
where BB is the dimensionless matric flux potential defined as rliB=8naQBlaq.

Raats (1971) derived as analytical solution to Eq. 2.25 for a point source on the soil
surface:
a,(r, z )
2QB
age"
E
+-
4n
a
- -(
2
(r 2
+ z2)"
')
where: Ei is the exponential integral:
El
(e-x I
x)&
(Abramowitz and Stegun,
1964).
The steady state distribution of the matric potential h and soil water content 6, (when the
h- 6 relationships for the soil are known) for surface or buried emitters can be calculated
For transient flow from point source, Warrick (1974) used similar transformations
(Eqs. 2.24 and 2.25) coupled with the additional assumption that dWd0=k (or d0/d0=
alk) where k is constant, to obtain the following time dependent linearized from Eq,
2.25.
according to Ben Asher et al. (1978), k can be determined by: (i) from known K(0)
functions; (ii) using the equality k= a D(0) with D (soil water diffusivity) evaluated at
some average 0; (iii) from flow experiments by matching predicted and observed h and
0 using k as a fitting parameter. The dimensionless variables, R= ar12, Z= a~-12,T=
aktl4, p= ( R ~ + z ~ ) "and
~ , 0 = aq@l8n, are introduced to solve Eq. (2.29) analytically,
subjected to the initial condition: 0(r, z, O), and to the boundary condition (1301dz) + a 0
for z=0, r#O
The solution for a buried point source is:
4B( R , Z , T )
e2
-{ePerfc(2P
2JT
f i )+ e - p e r j i c ( L f i
-
2JT
And the solution for a surface point source is
4,w,
2 , T ) = 2(4,
e22
q ,-22'
z
[ 4 12=z+dZ1
~
1
The integration in Eq. 2.31 can be solved by Gauss-Leguerre quadrature, 15- point
formula, with Z'=Z+x/2 (Sen et al., 1992):
where the weight value mi and sampling point xi maybe obtained from Carnahan et
al. (1969). In case of cyclic inputs (e.g. irrigation cycles), or temporal variations in
source strength, the value of 0 may be obtained by superposition in time ( Warrick,
1974):
With q-1 = 0, To=O, and P T , .

Finally, matric potential values may be obtained from the analytical solutions by Eq.
(2.28) and the corresponding 0(r, z, and t) can be obtained using any soil water
retention model (van Genuchten, 1980; Russo, 1988).
The Richards water flow equation (2.22) is difficult to solve due to the highly
nonlinear dependency of k on the unknown water content 0. Thus except for a few
special situations, Richards' equation can be solved using numerical methods such
as finite differences ( Brande et al., 1971; Lafolie et al., 1989) or finite elements
( Taghavi et al., 1984). One of the most comprehensive solutions to the problem of
water content from an emitter placed on the soil surface was developed by Brandt et
al., (1971). They assumed the soil to be stable, isotropic, homogenous, porous
medium. The initial water content (i.e. initial conditions) was assumed to be uniform
and low, relatively dry soil). The emitter is placed on the soil surface at the center of
a circle of radius r equal to half the distance between emitters and has a certain
known volumetric flow rate or discharge q. Based on observations that in the
vicinity of the emitter a radial area of ponded water develops, and that the water
content at the soil surface beneath this ponded area is at saturation (0= O,), it was
assumed that water infiltrates into the soil only through this ponded- saturated area.
The comparisons of the computed soil-water content data and the data obtained from
laboratory and field studies verified the validity of the model (Bresler et al., 1971
and Bresler and Russo, 1975).The results showed good agreement between
calculated and experimental and field data. In all these comparisons the surface
evaporation was taken to be negligible. However, even a high evaporation rate (10
mrnlday) would not have an appreciable effect on water content distribution
(Bresler, 1975) since the rate of evaporation is normally low compared to the soil
saturated hydraulic conductivity (10- 100 mmih). Consequently, only when
evaporation is extremely high, or when the hydraulic conductivity of the soil is very
low (<1.0 mmlh), does one need to consider the effect of evaporation during the
infiltration phase of trickle irrigation.
Peter et al., 2003 calculated the dimensions of wetted soil from hydraulic properties of
29 soils covering a wide range of textures and soil hydraulic properties to assess the
impact of soil texture and/or type on soil wetting patterns. The soils they used came
from two groups that differed in the extent to which hydraulic properties depended on
soil texture. Vertical and radial distances to the wetting front from both surface and
buried emitters were calculated (using Philip, 1984) for conditions commonly associated
with daily irrigation applications in a widely spaced row crop (sugarcane) and
horticultural crops. In the first group of soils, which had least expression of field
structure, the wetted volume became more spherical (i.e. the wetted radius increased
relative to the depth of wetting below the emitter) with increasing clay content, as is
commonly accepted. However, in the second group of soils in which field structure was
preserved, there was no such relationship between wetted dimensions and texture. For
example, five soils with the same texture had as great a variation in wetting pattern, as
did all 11 soils in the first group, indicating the considerable impact of field structure on
wetting patterns. The implications of the results for system design and management
were illustrated by comparing current recommendations for trickle irrigation systems in
coastal northeastern Australia with the calculated wetted dimensions. The results
suggested that (1) emitter spacing recommended for sugarcane are generally too large to
allow complete wetting between emitters, and (2) the depth of wetting may be greater
than the active root zone for both sugarcane and small crops in many soils, resulting in
losses of water and chemicals below the root zone. They also concluded that texture is
an unreliable predictor of wetting and there is no basis for adopting different dripper
spacing in soiIs of different textures in the absence of site-specific information on soil
wetting. Such information is crucial for the design of efficient trickle irrigation systems.
Chu, (1994) used a three-dimensional Green-Ampt analysis to describe the wetting

pattern of a surface emitter with a constant discharge by matching the emitter discharge
with the average infiltration capacity rate. He derived algebraic solutions of the wetted
radius and the maximum wetting- pattern depth. The time relationship between wetting
pattern of water dripping from a surface emitter and soil- profile characteristics is as
follows
ro2 )
( k , Hro / M)t
where R, in mm is the wetted radius on the soil surface; r, in mm is the source radius
(saturated radius around the source); H in mm is the green -Ampt potential parameter;
M in mm3 mm-3 is the Green -Ampt water content parameter (difference between soil
saturated water content and initial soil water content); and t in second is the time of
application.
2.3
Solute Transport in Soil
The transport of solutes (water- soluble chemicals) through soils is coupled with the
flow of water. The convective streams of flowing soil water carry salts and other
constituents. Self-diffusion of constituents in the liquid phase is another mechanism for
solute mixing and transport. Solutes may interact with the soil matrix (adsorption and
desorption), they may precipitate whenever their solubility is exceeded, and they react
among themselves. The primary mechanisms controlling the movement of salts in the
soil with the irrigation water are convection, diffusion, and mechanical dispersion
(Dasberg and Or, 1999). A brief review of these mechanisms follows.
2.3.1 Convective Transport

This is the passive transport of dissolved solutes with representing flowing water. In this
case, water and solutes move at the same average rate:
where c is the volume-averaged solute concentration, J, is the solute flux , and Jwis the
water flux (Darcy velocity). To estimate solute travel or arrival times, the mean apparent
velocity or pore water velocity (v) is used:
v=-
J w
2.36
Thus solute flux may also be characterized as J,= v 0 c. The water flux Jw represented
the flow velocity averaged over an entire cross-sectional area. However, because
convection occurs in the liquid phase only, v (>Jw) is used to represent the average
interstitial flow velocity in the liquid- filled pores.
2.3.2 Diffusion Transport
This is a spontaneous process resulting from random thermal motion, collisions, and
deflections of dissolved molecules. A net transfer of molecules of a solute species
usually occurs from regions with higher to lower concentrations as the result of
diffusion as described by Flick's first law. For free or bulk solution, the one
dimensional mass flux (Jd) due to molecular diffusion is given by:
where Do is the diffusion coefficient in bulk water. The diffusion coefficient in porous
media is lower than for bulk water. Because air and soil particles form barriers to liquid
diffusion, the apparent soil-liquid difhsivity Ds (L~IT)is a function of the available path
for diffusion determined by tortuosity z(B), resulting from the geometry of the medium
(i.e., texture and structure) and the volumetric water content. An example of the
relationship between Do (bulk water) and D, (soil) are given by Jury et al., (1991):
where n is porosity. For an unsaturated porous medium the flux of diffusing solutes can
be expressed as:
2.3.3 Dispersive Transport
Differences in flow velocities at the pore scale (due to different pore sizes, shapes, and
connectivity) cause the solute to be transported locally at different rates and thus lead to
mixing (or dispersion) of an incoming solution within an antecedent solution. The
process is macroscopically similar to mixing by diffusion (thermal motion); however, it
is entirely dependent on water flow (i.e., not driven by concentration gradients). The
solute flux due to mechanical (or hydrodynamic) dispersion (Jh) is described by an
equation similar to Fick's Law for diffusion:
Where Dh is the hydrodynamic dispersion coefficient (L~IT). This coefficient is

dependent on the interstitial pore flow velocity v (LIT), and on the dispersivity h (L) of
the soil ( a function of pore sizes and shapes) according to:
where a is an empirical factor usually assumed to equal 1. (i.e., a linear dependency of

Dh on v). The value of h may range from 1 cm in small columns to few meters in field
experiments. In most cases the relative effect of hydrodynamic dispersion can exceed
that of diffusion. Because of the macroscopic similarity between diffusion and
hydrodynamic dispersion it is common to combine their coefficients (assuming that they
are additive) into a diffusiondispersion coefficient (D,):
2.3.4 Convection-Dispersion Equation

The total flux of dissolved solutes in soil (j,) is the result of combined transport by the
three mechanisms discussed above, and may be described by the convection-dispersion
model :
where Js is the total mass of solute transported across a unit cross sectional area of soil
per unit time, Jw is the water flux (Darcian flux), De is the combined diffusiondispersion coefficient, and dc/& is the spatial solute gradient (partial derivative indicates
that the gradient may aIso vary with time). D, is dominated by the dispersion process
under most flow conditions. Combining Eq (2.43) with the continuity equation
(Conservation of solute mass) yields:
Where Bc is the mass of solutes in solution. The continuity equation may be written as:
Assuming steady-stae water flow in a homogenous soil profile (Jwand 6 are constant in
time and space), reduce Eq. 2.39 to the familiar form of the convection-dispersion
equation (CDE) for inert and non-adsorbing solutes:
Where D=D$6, and v=J,/ 6. The CDE may be expanded to describe two- and three
dimensional solute transports:
These forms of the CDE become more complicated when modi=ling reactive and
adsorbing solutes. In general, additional terms to account for various forms of solute
adsorption and transformations may be incorporated into the same basic form of Eqs
2.46 and 3.47.
2.4
Previous Studies on Solute Movement under Trickle Irrigation
The modeling of solute transport in soils is linked with water flow. To date, very limited
literature on solute transport from a point source is available. This was one of the main
reasons to select the topic of field evaluation of water and solute movement from a point
source.
Goldberg et al. (1971) studied salt and water distribution in the soil of a vineyard. They
found salt distribution in the soil followed the general pattern of water flux, with distinct
lateral and downward components. The authors found there were three salinity zones
with trickle irrigation: a high salt concentration zone near the soil surface which
increased toward the emitters, a deep level of accumulation, and a leached onion shaped
zone between the upper and the lower zone situated beneath the row.
Hills et al. (1991) presented results of Las Cruse trench experiment. The experimental
results indicate that water and bromide move fairly uniform during infiltration, while the
bromide move ahead of the tritium due to anion exclusion during redistribution.
Comparison between measurements and predictions made with a two-dimensional
model show qualitative agreement for two of the three water content measurement
planes. Model predictions of tritium and bromide transport were not as satisfactory.
Measurements of both tritium and bromide show localized areas of high relative
concentrations and a large downward movement of bromide relative to tritium during
redistribution. While the simple deterministic model does show larger downward
movement for bromide than for tritium during redistribution, it does not predict the high
concentrations of solute observed during infiltration, nor can it predict the
heterogeneous behavior observed for tritium during infiltration and for bromide during
redistribution.
A leaching study conducted by Bresler et al., (1971) demonstrates the effect of trickle
discharge and soil hydraulic properties on the shape of the actual salt concentration
distribution. In two soils (loam and sand) and trickle discharges (94 to 20 liters per
hour), the solution concentration of the saturated water entry zone is identical to that of
the infiltration water. It is clear that shape of the overall wetted zone is soiI and
discharge dependent. Therefore, the salt distribution pattern should also be affected by
these properties. In addition the leached part of the soil in the verticaI component of the
wetted zone is deeper and narrow than the radial component as the soil becomes coarser
and as the trickle discharge becomes slower. The salt quantities from the leached part of
the soil are accumulated and reach a maximum at a certain distance from the source. The
location of this maximum salt accumulation zone is largely dependent on both soil type
and the discharge rate of the point source. The size of this maximum salt content,
however, is affected mainly by the soil properties and not by the discharge rate.
Solute distribution around a trickle source for three dimensional steady state flow
was studied in terms of an approximate analytical model (Dasberg and Or, 19999.
Ion and water uptake by the root system were taken into account. Water flow and ion
concentration profiles were assumed to have hemispherical symmetry. The
distributions were nearly constant near the source with a sudden increase or decrease
closer to the wetting front depending on the ion balance (Zemel et al., 1987).
The model of solute flow was based on piston displacement travel and steady three
dimensional water flows from a surface point source. Experimental evidence was
presented to justify the omission of the solute diffusiondispersion term in the flow
equation. Laboratory experiments were described to demonstrate the rapid onset of

steady state water flow, especially near the emitter. Results from bromide tracer
laboratory experiments using Manawatu fine sand with constant infiltration into a
small surface cavity confirmed the isochrones predicted from theory (Clothier, 1984).
(Philip, 1984) considered the problem of predicting travel times of marked solute
particles emanating from continuous (i.e., steady) buried and surface point sources.
Defining a dimensionless time, T as:
T
qt
-
16
where: t is the actual travel time (T), q is the emitter discharge (LIT), a
parameter and
2.48
is is a soil
i (L~L-') is the average volumetric water content in the wetted soil
volume, an approximation for solute travel time from a buried emitter in the upward and
downward directions is :
Upward (Z< 0),
Where: Z=az/2. Solute travel times from a surface emitter in terms of dimensionless
radial (R= arl2) and vertical (Z) directions are:
Radial (Z=O),
Downward (R=O)
The SWMS-2D computer program for simulating water and solute movement in two
dimensional variable saturated media was developed by Simunek et al., (1992). The
program numerically solved Richards' equation for saturated-unsaturated water flow
and advection- dispersion equation for the solute transport using finite elements. A sink
term has been incorporated to account for water uptake by plant roots. The transport
equation includes provisions for linear equilibrium adsorption, zero order production,
and first order degradation. Water and solute movement in unsaturated, partially
saturated or fully saturated porous media in regions delineated by irregular boundaries
have also been handled. The Chain-2D code also simulates two- dimensional, variablysaturated water and solute movement. The governing partial differential equations were
solved numerically using Glarkin-type linear finite element schemes (Simunek and van
Genuchten, 1993)
Li et al. (2003) studied nitrate distribution from a point source trickle irrigation in loamy
soil. In the experiments, the apparent cylindrical application rate was varied from 0.6 to
7.8 I/h, the apparent cylindrical applied volume from 6 to 15 1, and the input
concentration from 100 to 700 mg/l. The results showed that the concentration in the
soil around the source was uniform for a given input concentration. For any input
concentration, the accumulation of nitrate at the boundary of the wetted volume was
observed. The nitrate concentration in the soil was primarily affected by the input
concentration. There was an increased nitrate concentration with a higher input
concentration. The effects of application rate and applied volume on nitrate distribution
were not significant up to a radius of 15 cm, beyond which either a higher application
rate or a smaller applied volume resulted in a higher nitrate concentration in the soil.
2.5
Hydrus-2D Model
The Hydrus -2D model simulates water flow, heat transport and the movement of
solutes involved is sequential first-order decay reaction in two-dimensional variably
saturated media. The program numerically solves the Richards's equation for saturated unsaturated water flow and the advection-dispersion equation for solute transport, using
a Galerkin-type finite element scheme. The flow equation incorporates a sink term to
account for water uptake by plant too. Flow and transport can occur in the vertical plane,
the horizontal plane, or in the three-dimensional region exhibiting radial symmetry
about the vertical axis. A detailed description of H y d m s 2 D is given by USDA.
2.5.1 The Governing Equations

2.5.1.1 Soil Water Movement
The differential equation governing the nonhysteretic flow of water in the elemental
field can be written in indicia1 notation as (Neuman, 1973):
1.h-based
a8
a
-=-(K~(o)K;
dt ax,
ah(e) K' ( 0 ) ~ :-)
ax,
the hydraulic properties near saturation. In our simulation the original expression of Van
Genuchten was used:
K ( h )=
K ( h ) = K ,s~,'[l (I -
s1Irn
lml2
I = the pore connectivity parameter [-I

13, = the saturated water content [ L ~ L - ~ ] .
8, =the residual water content [L3L"], may be defined as the water content at which both
dWdh and K goes to zero when it becomes large.
cx- an empirical factor [L-'I, whose inverse is referred to as the air entry value or
bubbling pressure.
n and m = the empirical factors affecting the shape of retention curve [dimensionless].
Where
8 = the volumetric water content [L3/L3 ]

h = the pressure head [L]
S = a sink term [T-'1

xi (i=1,2) = spatial coordinates [L],
t = time [TI
K'(@
K:
=the relative hydraulic conductivity a function of 8,
= the
hydraulic conductivity tensor at saturation [LT-'1

dh
D, = K r ( 0 ) ~(-);
80
= the
soil water difhsivity tensor.
In this study, h may be eliminated to have a ''@-based equation or 8 may be eliminated

to have an "h-based " equation.
The model permits the use of three different analytical models for the hydraulic
properties (Brooks and Corey, 1966, Van Genuchten 1980, and Vogel and Cislerva,
1988). The unsaturated soil hydraulic conductivity properties in the Hydrus- 2D model
are described by a set of closed form equations resembling those of Van Genuchten,
(1980) who used the statistical pore-size distribution model of Mualem (1976) to obtain
a predictive equation for the unsaturated hydraulic conductivity function in terms of soil
water retention parameters. Hydrus2D uses a modified version of the original Van
Genuchten equation with nine unknown parameters for added flexibility in describing
n and m = the empirical factors affecting the shape of retention curve [dimensionless].
S,= the effective degree of saturation, also called the reduced water content
(05Se51).
Estimating of Van Genuchten parameters from the experimental data requires:

(1) Sufficient data points of I9 (h) pair (at least 5 to 8) and data points of h (8) or
D (19) pair.
(2) A program for performing nonlinear regression.
Recent versions of many computer spreadsheets provide relatively simple and effective
mechanisms for performing nonlinear regression. However, RETention Curve, RETC,
(Van Genuchten et al., 1991) for describing soil water hydraulic properties can be used
to evaluate the parameters used in Hydrus- 2D model (Simfinek et al., 1999)
2.5.1.2 Solute Movement
In the governing transport equation of a single non-reactive ion with ignoring chemical
interactions and biological processes, the partial differential equation two-dimensional
chemical transport during transient water flow in a variably saturated rigid porous
medium is taken as:
where C is the solution concentration [ML-3], qi is the i-th component of volumetric

flux[LT-'I, S is the sink term in the water flow equation (I), C, is the concentration of
the sink term [ML"], and DcCis the components of the dispersion tensor [L2T-'1.These
components are given by (Bear, 1972)
q2
B D , = D,, >+
14
2
OD,
-+
= DL 4z
14
4 +
D, A
BzD,
14
2
D T %14
EkD,
Where
DLand DT=the longitudinal and transversal dispersivities [L]
1'
= the
absolute value of the volumetric flux density [LT']
Dw=the molecular diffusion coefficient of the solute in free water [L2T-'1

z =tortuosity factor [dimensionless]
CHAPTER3
MATERIALS AND METHODS
Design of Experiment
Laboratory and field experiments were conducted to determine water and solute
distribution pattern in soil under surface point source trickle irrigation. There were three
types of experiments in this study. The focused on the effect of water application rate
and amount of water on the lateral and vertical movement of water in sandy and sandy
loam soil by using plexiglass container. This type of experiments was conducted under
laboratory conditions. The application rates used for these experiments were 0.75 and
3.40 liters per hour for sandy soil and 1.0 and 3.0 liters per hour for sandy loam soil. The
second type of experiments was conducted in the laboratory using a wooden box. The
purpose of these experiments was to study the effect of application rate and amount of
application rates on the surface wetted radius. Three application rates were used for
these experiments 3, 5.5 and 7 liters per hour. The third type of experiments was
conducted under field conditions. These were designed for field evaluation of water and
solute movement from a point source. Two types of soil were selected for these
experiments, sandy and sandy loam. The application rates were selected to be within the
normal range for trickle irrigation from 1.5 to 6 liters per hour for sandy soil and from
1.5 to 4.0 liters per hour for sandy loam soil. The treatments of solute concentration
(chloride) were selected to be 13, 20, 35 and 54 meqll for sandy soil. In the case of
sandy loam soil, the solute concentrations were 16,25,30 and 47 meqll.
3.2
Design of Laboratory Experiments
Plexiglass and wooden box experiments were conducted under laboratory conditions at
the Soil and Water Laboratory (Universiti Putra Malaysia).The plexiglass experiments
were conducted to represent the flow of water in two dimensions while the wooden box
experiments conducted to represent the flow of water in soil in phase of three
dimensions.
3.2.1 Plexiglass Experiments
These experiments were conducted to determine the horizontal and vertical advances of
wetting front for two different types of soil under various water application rates. The
position of the wetting front was marked on the transparent wall of the plexiglass
container at fixed time intervals. The lines were then copied on transparent drawing
paper. The depth and width of the soil volume at various times during infiltration was
then determined from the results. Further descriptions of the experimental setup are
presented in the following sections.
3.2.1.1 The Water Application Apparatus
The apparatus used in this investigation was designed to provide uniform water
application from a point source at various rates. Essentially, the apparatus consisted of a
modified Marriott tube (Yitayew and Watson, 1986) was used as a reservoir joined to
emitter by means of hose. The reservoir, Figure 3.1, was composed of a circular
cylinder, and inner transparent tube. The cylinder has an inner diameter of 32 cm and a
length of 150 cm with a plixiglass side to monitor the level of water inside the reservoir.
The inner tube with both ends open has a 2 cm diameter and 4 m long. The reservoir was
sealed from both ends whereas the top has a hole to accommodate the inner cylinder.
The reservoir contains an outlet valve 25 cm above its bottom. A hose with 10 mrn
diameter was used to convey the water from the reservoir to the emitter. The reservoir
was at an elevation of 2 m above the ground surface. By changing the elevation between
the inner tube and the emitter point, the flow rates could be varied.
Inner tube
I
Reservoir
----.----------
Valve
-----------
Constant head
Figure 3.1 : Constant head and flow rate point source (adapted from Yitayew et a1 1998)
Application rate is one of the critical aspects in the design and management of trickle
irrigation system. It was therefore important to investigate the effect of application rate
on the horizontal and vertical distribution from a point source. On line, turbo-key,
button, and non pressure compensating emitter (Appindex A) was chosen to achieve the
requirement of a constant flow rate, and the same emitter was used for all treatments.
Prerequisite for the investigation was an accurately calibrated constant flow point
source. The material used during calibration was a stop watch, a graduated cylinder, and
two beakers.
3.2.1.2 The Soil Container
The soil container was built using 4 mm plexiglass in both faces of the container, with
inside dimensions of 120 cm in height, 120 cm in length, and 5 cm in width supported
by a steel frame as shown in Figures 3.2 and 3.3. In this manner it was easy to follow the
advance of wetted front during water application; one face of the container was
constructed with a removable plexiglass sheet.
3.2.1.3 Soil Profile Preparation
The experiments were conducted on two soil types, namely, a river sand and a sandy
loam soil. The air dried soil, which passed through a 2.0 mm sieve, was packed into the
soil container. A big funnel attached to a flexible tube 100 cm long with a 2.0 cm inner
diameter was used in order to prevent fractionation of the soil particles. The funnel and
flexible tube rested on the top side of the container and was filled by taking small
quantities from the bulk soil. The filled funnel and the tube were slowly moved
horizontally and rotated simultaneously. The soil was compacted as uniformly as
possible by progressively filling sections of the container with equal weights of the soil.
In this manner, an average bulk density of 1.56 jgrn/crn3 for the river sand and 1.31
gm/cm3 for the sandy loam soil was achieved.
Figure 3.2: Front view of the Plexiglass soil container (120 cm x 120 cm x 5 cm)
Figure 3.3: Side view of the plexiglass soil container
3.2.1.4 Description of Soils
The river sand soil was taken from a river bank near Universiti Putra Malaysia. The
texture of soil is medium sandy soil (98 % sand, 1% silt and 1% clay), with 90 % of the
sand having a mean diameter range of 0.25-0.5 mm. The sandy loam soil was taken from
a farm located within Universiti Putra Malaysia. The soil texture was sandy loam (15 %
clay, 24% silt and 61% sand), with 52% of the sand having a mean diameter range of
0.1-0.25 mm.
3.2.1.5 The Experiment
After preparing the soil profile, the water was continuously applied to the midpoint of
the soil surface. For sandy soil the selected application rates were 0.750 and 3.40 l/h for
120 and 70 minutes, respectively. In the case of sandy loam, the water was applied at
application rates of 1.0 and 3.0 l/h for 270 and 250 minutes, respectively. The
application rates were chosen to be within a wider range whereas the first application
rate i.e., 0.750 and 1.0 l/h represent a low application rate while 3.40 and 3.0 represent a
high application rate. On the other hand, the selected time was based on the type of soil
smaller in case of sandy soil and larger in sandy loam soil. After the water was turned
on, the advance of the wetted front was recorded periodically by marking the
displacement on the transparent plexiglass surface. When the water was turned off, the
soil profile was immediately placed in a horizontal position, the screws holding the sides
of the apparatus were removed, and the detachable plexiglass was removed to permit
sampling. To facilitate sampling, a steel grid with 6.0 cm spacing was used to make a
trace on the soil surface. Within each grid location, the section of soil was cut with a
spatula and placed in a weight can. The water content of each segment was determined
gravimetrically then converted to the volume basis by multiplying the average bulk
density of the soil profile. The sampling procedure required approximately 30 minutes.
It was assumed that the water redistribution during sampling was negligible.
3.2.2 Wooden Box Experiments
The apparatus used in this investigation was designed to provide a uniform water
application from a point source at various rates. Essentially, the apparatus consisted of a
modified Marriott tube as was used by Yitayew and Watson (1986) described earlier.
The point source was calibrated very carefully for the application of required volumes at
desired rates of application. The outlet was located on the soil surface at the center of the
soil container, Three-application rates, i.e., 3.0, 5.5, 7.0 l/h were chosen. The
experiments were conducted on river sand (98 % sand, 1% silt and 1% clay) with 90 %
of the sand having a mean diameter range of 0.25-0.5 mm. The air-dried soil, which
passed through a 2.0 mm mesh, was packed into a plywood soil container 1 m2 in area
and 1 m deep (Figure 3.4).
The soil was compacted as uniform as possible by
progressively filling sections of the container with equal weights of soil. In this manner
an average bulk density of 1.56 gm/cm3 was achieved. After preparing the soil profile,
water was continuously applied to the midpoint of the soil surface at 3.0, 5.5, 7.0 l/h for
380, 300 and 180 minutes, respectively. During each experiment, the advance of the
wetted front on the soil surface was measured fiom different axes, and the average was
taken to infer equivalent surface wetted radius, (Figure 3.5). After termination of
irrigation, soil samples were taken fiom one half of the wetted area with the aid of a 30
mm diameter sampling tube. Sampling for moisture determination was executed at a
distance of 60 mm from the trickle source and with depth increments of 60 mm along
the radial line. The sampling procedure required approximately 30 minutes. It was
assumed that the water redistribution during sampling was negligible. After the final
sampling, a vertical transaction was made through the point where the source was
located to determine the wetting-front position.
Figure 3.4: Inside view of the Wooden box
Figure 3.5: Measuring the surface wetted radius
3.3
Field Experiments
It was decided that field experiments should be conducted under conditions similar to
those in the actual field application of trickle irrigation. Field experiments were
conducted on sand and sandy loam bares soil found at two locations situated 20 km
south of Tripoli-Libya along the way to Tripoli International Airport. This area was
selected because of the potential of trickle irrigation on these types of soils. The sandy
soil was classified as loose, very friable, and highly drained and has very fine roots,
while the sandy loam was classified as massive, moderately hard, and fine and medium
irregularly shaped of lime concentrations with common fine roots.
3.3.1 Description of Field Experiments
The apparatus of trickle irrigation system used in this investigation was designed to
provide uniform water application from a point source at various rates. The apparatus
consisted of 120 liter plastic drum as water supply reservoir to provide water at a
constant rate. The reservoir serves as a modified Marriott tube (Yitayew and Waston,
1986) and joined to (5 to 6) adjustable emitters by means of hose 75 mm ID. Much
efforts have been made to ensure that each emitter gave a constant application rate. On
line, button, openable, adjustable and non pressure compensating emitter (Appendix A)
was chosen to achieve the requirement of a constant flow rate and the same type of
emitter was used for all treatments. Prior to conducting each experiment, a Plexiglass
sheet (4 mm thick), 100 wide and 100 cm long was installed alongside the wall of
trenches dug to identify and to follow the advance of the wetted front during water
application. Efforts were required to level the soil surface where emitters were located.
After preparing the test site, the reservoir was taken to the test site and fixed to a height
of 1.7 m above the ground, and then it was filled by saline water obtained by mixing tap
water with several volumes of 1N C1' (as NaCl). Chloride was chosen because it has
desirable characteristics such as a low concentration in soil, high solubility,
micronutrient, conservative, simply, accurately and economically measured. Four
different concentrations of chloride were used 13, 20, 35, and 54 meqll chloride for
sandy soil experiments and 16, 25, 30 and 47 meqll for sandy loam soil experiments.
The wide range of chloride concentrations would be helpful for comparisons purposes.
The Emitters were calibrated for the application of the required volume at the desired
rate. The outlets were located on the soil surface where each dripper irrigated a distinct
area without any overlap with others; meanwhile the last dripper was located 10 cm
away of the edge of the trench (Figure 3.8).
Various application rates were chosen between 1.5 and 6 liters per hour for the sandy
soil and 1 to 4 l/h for the sandy loam soil. These wide ranges of application rate would
be helpful for comparison purposes. During the application, the advance of the wetted
front on the soil surface and the perpendicular face of the trench was recorded
periodically, (Figures 3.9 through 3.1 1). At the end of irrigation a vertical soil profile
perpendicular to the plane of the emitter source was exposed. This was conducted for the
sites where the plexiglass was installed. Due to accessibility of sandy soil, other sites
were also exposed. This was not possible for the case of the sandy loam soil sites. A
coordinate system was established on the profile with the origin at the soil surface
directly where the emitter source was located. Due to the symmetrical shape of wetting
pattern, only one-half of the wetted profile was sampled for water and solute
determination. The sampling layout for soil water and solute measurements was
horizontal and vertical intervals of 5 centimeters, starting from the application and
moving outward to the edge of the wetted surface and the same 6 cm interval for vertical
data was fixed up to the wetting front.
.-
- -
Figure 3.6: General view of the experimental setup for sandy soil
Figure 3.7: General view of the experimental setup for sandy loam soil
--
Figure 3.8: Top view of horizontal view of wetting pattern in sandy loam soil
Figure 3.9: Side view of vertical wetting pattern in sandy loam
Figure 3.10 Wetting front movements in the vertical direction in sandy soil
3.3.2 Soil Water Determination
The gravimetric with oven drying method involves weighing a moist sample, oven
drying it at 105 O C , reweighing and calculating the mass of water lost as a percentage of
the mass of the dried soil. Using the bulk density of soil, water content on mass basis
was converted to water content on volume basis.
3.3.3 Chloride Analysis
The soil samples taken from field experiments and used for moisture content
determinations were extracted with deionized water in a 1:l soil/water ratio. The
chloride content of the extract was made. The chloride content of the extract was
determined by titration with 0.005 N AgN03 (Black, 1965). The following steps
describe the chloride analysis in the soil:
Soil Extraction
The soil extraction was prepared by adding 100 ml of distilled water to 100 gm of 105OC
oven dry soil (1 :1 extract volume ratio), shacked well for 30 min, the soil was allowed to
settle briefly, and then filtered through extractant- washed filter paper.
Reagents
A. Potassium Chromate Solution (K2Cr04), 5% in water

Dissolve 5 g potassium chromate in 50 mL distilled water.
Add dropwise 1 N silver nitrate (AgNO,) until a slight permanent red precipitate is
formed.
Filter, and bring to 100-mL volume with distilled water.
B. Silver Nitrate Solution (AgN03), 0.01 N

Dry about 3 g silver nitrate in an oven at 105OC for 2 hours, cool in a desiccator, and
store in a tightly stoppered bottle.
Dissolve 1.696 g dried silver nitrate in distilled water, and brings to 1-L volume with
distilled water.
Procedure
The procedure is described as follow:
I . Pipette 5 mL soil saturation extract into a 100-mL Erlenmeyer flask.

2. Add 4 drops potassium chromate solution.
3. Titrate against silver nitrate solution until a permanent reddish-brown color appears.
4. Always run two blanks containing all reagents but no soil, and treat them in exactly
the same way as for the samples. Subtract the blank titration reading from the readings
for all samples.
Calculation
Where: V = Volume of 0.005 N AgNO, titrated for the sample (mL).

B = Blank titration volume (mL)
R = Ratio between total volume of the extract and extract volume used for titration.
N = Normality of AgNO, solution.
Wt = Weight of air-dry soil (g).
The concentration of chloride is expressed in term of soil solution (as meqll). Using the
information of the chloride concentration based on soil solution and the volumetric
moisture content of the soil, the chloride concentration in term of bulk soil can be
calculated using the following relation:
Conc (meqll bulk soil) = (0v) Conc ( meqll soil solution)
3.4
Semi Empirical Model
For many practical situations, detailed information on matric potential or water content
distribution within the wetted volume is not necessary, and predictions of the boundaries
and shape of the wetted soil volume is suffice. Moreover, the routine use of numerical
models or analytical transient flow for design purpose is still considered cumbersome
and impractical due to the lack of information on soil hydraulic properties. In trickle
irrigation the distribution of water in the soil is primarily a function of the soil texture,
the application rate, and the total volume of water applied. Based on these variables,
semi empirical method was developed to determine the main component of the wetted
pattern of soil, i.e. the surface wetted radius and the depth of wetted volume. Saturated
hydraulic conductivity, average moisture content in the wetted volume, emitter
application rate and total volume of water applied were selected to represent the inputs
of the predicted equations.
3.4.1 Theoretical Considerations
As evident from the previous studies, three factors are responsible for determining the
soil-wetting pattern under trickle irrigation; soil properties, emitter discharge rate and
the total amount of water applied. The derivation of a new relationship for wetting
pattern is presented in the following sections.
3.4.1.1 Soil Properties
The effect of soil properties on the wetting pattern emerged from trickle irrigation is
strongly related to its porosity (n). On the other hand, the concept of soil porosity
reflects its hydraulic conductance (k). The gravity force would act on macro-pores which
could result in greater vertical movement. In micro-pores water movement is more
commonly a factor of capillary action, which can move water laterally. Furthermore,
raising the value of porosity may indicate to raise the value of soil water retentivity. This
is additional reason explaining the slower movement of soil water in heavy textured soil.
3.4.1.2 Emitter Discharge Rate
The effect of discharge rate on wetting pattern is more interacted with the soil properties.
If water applied at a fast rate, then some ponding could occur and the lager capillaries
would fill, causing greater moisture content in the wetted volume the gravity force
would act on the lager pores which could result in greater vertical movement. This
would probably be more noticeable in a sandy soil because of the larger capillaries
compared to a clayey soil. A higher flow rate on a heavy soil could have a different
effect. Since the capillaries are smaller for a clayey soil, the water movement by gravity
would be slower compared to sand soil. This could result in a larger ponded area which
would increase the lateral water movement.(Roth 1983).
3.4.1.3 Total Amount of Water Applied
The effect of total amount of water applied on the geometry of wetting pattern of soil is
to be in direct proportion. This would seem logical, as the excess water applied would
have to wet a greater volume of soil. Thus, to state the geometry of wetting pattern as a
fimction of the total amount of water applied rather than elapsed time would be more
adequate from a physical perspective. Many studies found that the total amount of water
applied has significant effect on wetting pattern distribution, Schwartzman and Zur
(1%6), Eq. 2.10 and Eq. 2.1 1 and Gupta et al., (1995), Eq. 2.15.
So in any soil, the functional relationship between all variables can be defined as
follows:
where
r = wetted radius
k = soil hydraulic conductivity
n =soil porosity
q, =application rate
V,
= volume
of water applied
z = the depth of wetted zone.
It appears that a simple procedure based on previous variables could be developed to

predict the wetting pattern geometry. The accuracy of results depends on the following
approximations:
i.
A single surface point source irrigated a bare soil with a constant discharge rate
(4w).
..
11.
...
The soil is homogeneous and isotropic.
111.
There is not a water table present in the vicinity of root zone.
iv.
The evaporation losses are negligible.
v.
The effect of soil properties is represented just by its porosity and saturated
hydraulic conductivity.
vi.
The value of saturated moisture content porosity equals the value porosity of. It
could be obtained using an equation given by Hillel, (1982) which states:
and
Where
n = porosity of the soil
0,= Moisture content at 0 bars

pb=bulk density of the soil (measured)
p,=particle density of the soil (assumed 2.67 gm/cm3).
According to previous approximations, the relationships of 3.1 and 3.2 become:
Ben-Asher et al. (1986) investigated the infiltration from a point trickle source in the
presence of water extraction using an approximate hemispherical model. They suggested
that the position of the wetting front is a function of the half value of saturation moisture
content. For infiltration from a point source without water extraction they found that:
I
R(t)a (
The new variable A@ is called the average change of soil moisture content.
This leads to:
Shwartzman and Zur (1986) proposed simple relationships of the following form
between the wetted diameter and vertical distance to wetting front and emitter discharge
rate, soil hydraulic conductivity, and the total volume of water in the soil:
ks
w = K , (V, ) 0 2 2 (-)
4,
-0.17
where
W = wetted width or diameter of water pattern, m

Z = vertical distance to wetting front, m
K1 = 0.03 1 (empirical coefficient).

K2 = 29.2 (empirical coefficient);
V,
= volume
of water applied, L;
K, = saturated hydraulic conductivity of the soil, mls;

q = point-source emitter discharge, L h ;
Despite Equations 3.12 and 3.13 offering a simple and useful means for predicting
wetting pattern, including the expected distortion in wetted volume, which is not
predicted by the hemispherical approximation, they need to be validated against
experimental values.
According to the approaches introduced by Shwartzman and Zur (1986) and Ben Asher
et al., (1986), the nonlinear expressions describing wetting pattern may take the general
forms as:
Where
r = the surface wetted radius (L)
z = the vertical advances of wetting front (L)
A 6 = the average change in volumetric water content within the wetted zone ( L ~ / L ~ )
V,= the total volume of water applied (L3),

q,
= the application rate
(L3/~)
Ks = the saturated hydraulic conductivity (L/T)

a, p, 7, h, p, o,6 and S= the best fit coefficients.
3.4.2 Model Optimization
Once the model structure and order have been identified, the coefficients that
characterize this structure model need to be estimated in some manner. To determine the
coefficients of Equations 3.13 and 3.14, four available published experimental data were
adopted, Taghavi et al., (1984), Anglelakis et al, (1993), Hammami et al, (2002), and Li
et al, (2003). The choice of these experiments was essentially based on their convenient
data. The procedures of these experiments are available in their original papers.
However, Table 3.1 displays the input variables used in Equations 3.14 and 3.15.
3.5
Hydrus-2D Numerical Model
Water and solute movements in two directions are complicated compared with one
direction movement behavior. Numerical models are useful solutions to predict water
and solute movement behavior in active root zone of the crop. Simunek et a1 (1999)
developed a general purpose two dimensional numerical model known as Hydrus-2D.
Table 3 1 Numerical values of input variables used in the predicted models.

Input Parameters
-
Taghavi et a1
Anglelakis et a1
Hammami et a1
Li. et a1
Clay loam
a. clay loam
Silt
loam
--
Texture of soil
b. sand.
Emitter application 2.1 and 3.3
rate (lph)
a. 2.1 and 7.80
1.O
b. 9.0 and 12.3.
2.0
4.0
Saturated moisture 0.53

content (vol)
Saturated hydraulic 0.85
conductivity ( c m h )
a.0.85
5.8
1.85
b.5.8
The Hydrus-2D program numerically solves the Richards' equation for saturated and
unsaturated water flow and convection-dispersion equation for solute transport. The
program is based on finite- element scheme and incorporates a graphical user interface
for data entry (soil properties, atmospheric conditions, emitter discharge, and boundary
conditions). The model will be used to compare the agreement between the
measurements and calculated values of water and solute movement under point source
trickle irrigation.
3.5.1
Problem Definition
Simulation of the movement of water and a dissolved solute from a point source into the
soil profile, which contains only one layer inherently, prescribes itself to an
axisymmetric situation. This is one of the reasons that Hydrus-2D code was used since it
has the capability of simulating axisymmetric flows.
The following boundaries were considered for this analysis. Because of symmetry, a
vertical cross section (ABCDE, Fig. 3-12), along the radius of the soil sample, was
selected for simulation by the computer model. The boundary conditions for the cross
section were shown in Figure 3.12 are as follows:
AB= constant flux boundary, represents the saturated water entry radius.
BC, CD, DE and EA = no flow boundary.
no flow
r
L
3
0
F'
0
C
Z 1r
no flow
Figure 3.12: Schematic diagram showing the physical layout of the trickle irrigation
system implemented in HydrusW2D,including the boundary conditions. The upper solid
line indicates the saturated water entry radius.
The boundary EA was a no flow boundary due to symmetry. The boundaries BC, CD
and DE were selected as no flow because in the actual experiment no water reached
these boundaries. No flow boundary conditions is a special case of Neumann type
boundary conditions where Darcy's law relates flow to pressure head gradient where,
A constant boundary condition was selected for the leftmost corner nodes of the domain
to represent the point of application of water.
3.5.2
Boundary condition
The initial and boundary conditions can be mathematically formulated as follows:

3.5.2.1 Water Movement
In order to investigate the influence of trickle discharge, soil hydraulic properties and a
constant flux boundary condition at the saturated water entry radius were used.
The initial condition that must be satisfied within the flow domain to be uniform
throughout the soil profile is
and the boundary conditions are
dh
i3r
for
-= 0
for
-=
ah
dr
ah
dh
-=-I
a~
for
for
During water application, the point source boundary will have a constant saturated water
entry radius. The water application rate becomes:
A = xrsi
where Q is the flow rate from dripper [L~/T],A is the area of surface saturated water
entry [L*] and r, is saturated water entry radius [L]
3.5.2.2 Solute Transport
The initial solute concentration (c,,) within the flow domain is selected to be uniform
throughout the soil profile;
~ ( rZ,) =tin t=O
3.25
Solute is applied with irrigation water and a third-type boundary condition is used to
prescribe the concentration flux along the boundary of the emitter.
3.5.3 Inputs for Simulation Model
Hydrus-2D model requires the inputs of the soil hydraulic and solute parameters. The
inputs of soil hydraulic parameters are residual water content (Or), saturated water
content (O,), empirical factors of a, n, and m in Eq.2.55 saturated hydraulic conductivity
(k, ) and the pore connectivity parameter (I ). In addition, the longitudinal (DL) and
transversal (DT)dispersivities are required to simulate the solute movement.
3.6
Statistical Criteria
The data obtained from this study was evaluated statistically. Both F and t-test were used
for this purpose. The regression analysis technique was used to study the relationship
between the dependent variable (e.g. the wetted radius advance and the vertical advance
of the wetting front) and the independent variable (e.g. the elapsed time, volume of
water applied).
The comparisons between the predictions as estimated by the semi-empirical model and
Hydrus-2D model with the observed data were evaluated statistically. The indicators
used to evaluate the agreement between model (simulated) data and experimental data
were: Mean bias error (MBE), root mean square error (RMSE) and Theil's inequality
coefficient (U). These indicators allow for the detection of the differences between the
model estimates and experimental data. These indicators can be calculated as follows:
(PI -
J g Z T + Z
r=l
Where P, = the predicted data from the model, A,
= the
experimental data and n
= the
number of records, (Naylor 1970 and Hossein et al., 2004). The MBE, RMSE statistics
have as lower limit the value of zero, which is the optimum value for them as it is for U
The linear regression equation with intercept equals zero was also used to determine the
goodness of the predicted model. The general form of this equation is expressed as:
Where y represent the predicted data, a the slope of the regression line and x represents
the measured data. For perfect fit the value of the slope equals one.
A correlation analysis was performed between the measured and predicted data. The
value of coefficient of determination ( R ~ )given in the correlation analysis would
indicate the accuracy. of the predictions under experimental conditions. In addition, the
results of comparison between the measured and predicted data were plotted against a
1:l line. In case of perfect prediction all paired data would fall on the line across the
graph signifying a slope of one (Observed= Predicted).
3.7
Determination of the Soil Hydraulic Properties
Like the hydraulic information (emitter discharge rate and total water applied), the soil
hydraulic properties also govern the wetting pattern in soil. Therefore the determination
of soil hydraulic properties can help to understand and envision how the wetting pattern
would be and explain the differences of the wetting pattern within various textured soil.
In addition, these properties are important in order to be able to run HYDRUS-2D. The
selected soil hydrauIic properties and their determinations are described below.
3.7.1 Soil Texture
Soil physical properties such as texture and pore-size distribution are major determinants
in water movement in soil. For various purposes a detailed description of porous matrix
is required: Granular materials such as soils are often described by their particle -size
distribution. The hydrometer method and sieve analysis were used to determine the
particle-size distribution (soil texture). Hydrometer is the commonly used method to
determine the particle size distribution (percent of clay< 0.002 mm, silt 0.25-0.002 rnm,
and sand 2-0.25 mm). The particle size analysis was performed to determine the particle
size distribution of the sand particles.
3.7.2 Bulk Density
The bulk density is a basic soil property that is defined as mass of dry soil per unit
volume of soil
MIL^). In case of laboratory experiments the bulk densities of river sand
and sandy loam soil were determined based on the total mass of soil filled in the
volumes of the soil container. The average bulk density obtained was 1.56 and 1.31
gmlcm3 for river sand and sandy loam soil, respectively. For field experiments, thirty six
samples of soil were taken from various depths using a cylinder with dimensions of 5cm
length and 5 cm diameter. The average bulk density was found to be 1.57 and 1.38
gm/cm3 for sandy and sandy loam soil, respectively.
3.7.3 Porosity
The porosity, n is given as the ratio of voids to the total initial volume of soil, which
includes both voids and solids and expressed as:
where V, is the volume of the voids and V is the total volume of the soil.
The total porosity, n is a measure of the volume percentage pore space and is usually
derived from measurements of soil bulk density (pb) and soil particle density (p,). In this
study the total porosity is calculated based on particle and bulk densities approach
applying the following equation:
In this method the bulk density pb is determined as explained in the previous section.
The particle density p, of soil is assumed to be 2.67gm/cm3, For soils used in laboratory
experiments the porosity was calculated and values of 41% and 50 % were obtained for
river sand soil and sandy loam soil, respectively. The porosity was also calculated for
soils used in field experiments and values of 41% and 48 % were obtained for sandy and
sandy loam soil, respectively.
3.7.4 Soil Water Retention Curve
The relationship between matric suction and volumetric water content (the soil water
retention curve) was determined under laboratory conditions. Sand box method was used
to measure the soil moisture content at values of head ranged from 0 to 100 cm. For the
range larger than 100 cm, the traditional pressure plate membrane apparatus was used.
3.7.5 Saturated Hydraulic Conductivity
There are several methods that can be used to estimate the saturated hydraulic
conductivity (k,) under laboratory and field conditions. In this study, hydraulic
conductivity was estimated in laboratory using the concept of constant head
pennearneter. The test was carried out for ten samples for each type of soil whereas the
samples of laboratory experiments were determined based on disturbed soil columns and
undisturbed samples for field experiments soil. Based on this approach, the outflow
volume and time was recorded at fixed head. The hydraulic conductivity was measured
by using Darcy's equation:
where, k, is the saturated hydraulic conductivity (LIT)
v = the average flow velocity (LIT) at cross-sectional area A ( L ~ )and

, calculated as:
Q= the total application rate (L~/T)
i = the hydraulic gradient (dimensionless) calculated as:
where, Ah is the loss of hydraulic head (L) along the distance 1 (L).
The average value of saturated hydraulic conductivity of the samples was found to be 30
crnlh and 1.48 cm/h, respectively for river sand soil and sandy loam soil in the case of
laboratory experiments. For field experiments, the average saturated hydraulic
conductivity of the soil samples was found to be 20.1 and 5.8 cm/ h for sandy and sandy
loam soil, respectively.
3.7.6 Soil Water Diffusivity
The diffusivity fimction within the interval of saturation and arid dried soil was
measured by employing the method of Bruce and Klute (1956). The apparatus used for
making these soil water diffusivity measurements consists of a horizontal plexiglass tube
with a 3 cm inner diameter and a length of 60 cm, sectioned into 2cm rings,
(Figure 3.13). Soil samples were packed into the tube at desired bulk density. Water at 2
cm suction was allowed to enter the tube at one end. At the conclusion of the experiment
the column was sampled to determine its water content at different distance from the
wetted end. The water content 8 and its distance x from the wetted end was plotted.
From the graph the D (8) was then calculated as
where
CHAPTER 4
RESULTS AND DISCUSSION
This chapter discusses the results of water and solute distribution under point source
trickle irrigation. There are three sections in this chapter corresponding to the objectives
of the study. These are:
1. Evaluation of water and solute distribution under point source trickle irrigation
2. Predictive model for wetting pattern dimensions
3. Model evaluation Hydrus -2D
4.1
Evaluation of Water and Solute Distribution under Point Source Trickle

Irrigation
The results of the laboratory and field experiments are presented to develop an
understanding for a water and solute distribution pattern applied through a point source.
The results of the laboratory and field experiments are further shown in the appendices
Band C.
4.1.1
Laboratory Experiments
As described earlier, the laboratory experiments were categorized into two types
according to the wetted zone, i.e. two dimensional flow represented by the plexiglass
box experiments, and three dimensional flow represented by the wooden box
experiments.
4.1.1.1 Plexiglass experiments

Wetting Front Movement
Figure 4.1 shows the wetting front advance at different times for river sand profiles with
the application rates of 0.75 and 3.40 l/h. The horizontal water movement, after water
application, was less than the vertical advance. The elongation in the vertical direction,
probably due to the preferential flow channels in the soil and to the effect of the gravity
as additional force which represented by the second term in the right hand side of
Richards equation (Eq.6) i.e., q,= K (h)(d(htz)/z). The geometry of the two dimensional
water movements in the soil was characterized by the lateral advance of wetting front
(H) along the x- axis and the vertical advance of wetting front (Z) along the z- axis. The
point source (emitter) was located at the point with coordinate having the origin, of (0,
0). The wetting fronts were observed for river sand and sandy loam soil visually through
the plexiglass sheet. Figure 4.2 shows the wetting fronts versus the elapsed time for river
sand soil with application rates 0.75 and 3.4 l/h. The cumulative volume of water applied
were 1.5 and 4 liters for 0.75 and 3.4 llh application rates, respectively. As it can be
observed, the horizontal advances of wetting front rate decreased with increasing of time
while the vertical advances of wetting front progresses steadily through the irrigation.
The reason the vertical water advance progresses steadily throughout irrigation, while
the horizontal advance rate decreases with time during irrigation is that at the beginning
of the irrigation, for any light textured soil (like the river sand), capillary forces are the
dominating force which pull the water and caused it to move in all directions. But, after
a certain period of time , when the water content of the upper part of the wetted soil
approaches saturation, water begins to move easily and quickly downward under effect
of the gravity forces which then become more significant.
Water application rates caused a notable affect on both horizontal and vertical advances
of wetting front. To be able to compare the effect of trickle discharges rate and to place
the two different rates on an equal basis, the results are compared for an identical
amount of cumulative infiltration and not for an equal time allowed for the irrigation. It
can be seen the horizontal advances of wetting fronts is 11 cm and 17 cm for 0.750 and
3.4 l/h application rate, respectively when the cumulative application water wasl.5 liters.
On the other hand, at the same cumulative of water application, 1.5 liters the vertical
advance of wetting fronts is 62cm and 74 cm for application rates 0.750 and 3.4 l h ,
respectively.
Both horizontal and vertical advances of wetting front have a good relationship with
elapsed time. The following forms can describe such relationships,
H =atb
where Hand Z are the horizontal and vertical advance of wetting front ,respectively, t is
the elapsed time and a and b are empirical constants.
The regression analysis of the horizontal and the vertical advances of wetting front with
elapsed time indicated that the a-values were increased as the emitter discharge rate
increased. In the case of horizontal advance of wetting fronts, the a-values were found to
be 3.96 and 4.86 with respect to the emitter discharge rates of 0.750 and 3.40 llh,
respectively. Regarding to the vertical advance of wetting fronts, the a- values were 3.12
and 6.98 under the application rates of 0.750 and 3.40 l/h respectively. This result
concurs with Hachum, et al., (1976), Li et al., (2003) and Elmaloglou et al., (2003)
findings. They indicated that a- values increased as the water application rate increased.
The spreading of the wetting front at different time intervals for sandy loam soil with
water application rates of 1 and 3 l/h is shown in Figure 4.3. The vertical advancements
of wetting front were less than those for river sand and the horizontal advances were
more expanded.
- - . - . . . - .
100
- - - . . - - . .
q=0.750Ilh
-----
80
H= 3.96 t022 p=0 94

z=3 12 t o " F?=OAO
/
/
-5
/
/
/
60.
(D
>
m
I
/
/
/
40
/
/
I
I
20
/
/
./
)-.-"
yh
---.-
w--
J-
--..------ '
Elapsed time (min)
Elapsed time (min)
Figure 4.2: The distance of the wetting front from the center of the trickle source as a
function of elapsed time for river sand
Distance, x (cm)
Dis
-20
-10
10
20
30
40
Figure 4.3: Wetting front advances at different times for sandy loam soil under trickle
discharge rates of 1.O and 3.0 l/h
The spreading of the wetting front for sandy loam soil with the application rates of 1 and
3 l/h were plotted versus the elapsed time (Figure 4.4). The cumulative volume of water
applied were 4.5 and 12.5 liters for 1 and 3 l/h application rates, respectively. The
horizontal advances of the wetting front immediately below the point source moved very
fast compared to the vertical advances at the beginning of irrigation. However, after a
period of time, the vertical advance of wetting front surpassed the horizontal advances
by considerable distance.
-6
8s
([I
>
u
([I
-0
Elapsed time ( min )
Elapsed time (min)
function of elapsed time for sandy loam soil
This is probably affected by the change in the size of the water entry saturated zone as a
function of time (Bresler, 1978). At the beginning of irrigation the size of water entry
saturated zone was very small and identical to the area formed at the soil -water drop
interface. Thus, a portion of the water discharging out of the emitter was not able to
penetrate the soil through this area and because of conservation principle, a factor which
increased the horizontal advance of water. The rate of growth of the saturated water
entry decreased with time and then reached a steady value when the vertical flux of
water equal the saturated hydraulic conductivity of the soil. At that moment, water
begins to move easily and quickly downward under the effect of the gravity forces
which then become more significant. On the other hand, the leveling of the soil surface
may induce a compaction which results in decreasing the saturated hydraulic
conductivity at the soil surface. Many studies are in agreement with these finding where
the horizontaI advance of the wetting pattern at the soil surface moved very fast
compared to the vertical advances at the beginning of irrigation (Bresler, 1971, Hachum
et al., 1976, Angelakis et al., 1993 and Li et al., 2003). For example, Angelakis et al.,
(1993) found that the horizontal advances of the wetting front immediately below the
point source moved very fast compared to the vertical advances at the beginning of
irrigation (Figure 4.5).
1000
2000
3000
Elapsed time (min)
4000
1000
3000
Elapsed time (min)
hnction of elapsed time (measured by Angelakis et al., 1992)
The effect of the discharge rate on the wetting front can be seen very well when
comparing between both treatments. To consider an example of the effect of discharge
rate on the horizontal advances of wetting front, the cumulative water applied of 0.5, 1.5
and 2.30 liters were selected. The horizontal advances of wetting front were 13 cm, 16
cm and 21 cm, respectively for 1 l/h application rate.
In case of 3.40 l/h application rate, the horizontal advances of wetting front were 13,21
and 23 cm, respectively. These results can be explained by the fact that : as the water
application source increases, the width of saturated soil zone near the trickIe source
increases; this in turn enhances the horizontal advances of the wetting fronts. Unlike the
river sand soil, the increasing of the vertical advances of wetting front in sandy loam is
higher at lower discharge rate than under the higher application rate. This was a
consequence of a larger saturated water entry. When the water had spread over a larger
surface the movement of water was not as deep for a given quantity of water. On the
other hand, the deeper wetted depth in smaller discharge rate treatments may be caused
by the fact that these treatments had a longer time to infiltrate into the soil profile
(Bresler, 1978). Another cause of low vertical advance of wetting front is the effect of
entrapped air block (Warrick, 2002), where the mobile air is displaced from the finer
into larger pores as the fine pores fill by capillarity during water infiltration. Thus,
entrapped air blocks the larger pores which have a disproportionate effect in reducing
flow. Roth (1983) indicated that the increased flow rate would require a greater
unsaturated hydraulic conductivity value to conduct more water. Thus larger pores
would be required to transmit this water.
For application rate of I Ih, the vertical advances of wetting front were 16, 28 and 34
cm, under water application of 0.5, 1.5 and 3.4 liters, respectively. On the other hand,
under application rate of 3 l k the vertical advances of wetting front were 10, 20 and 24
cm for 0.5, 1.5 and 3.4 liters, respectively. These are in accordance with previous
findings where higher water application rates result in a decrease in the vertical
component of the wetted soil volume. (Goldberg et al., 1971, Ragab et al., 1974, Bresler
1978, Bucks et al., 1985, ~chwartzman1986, Ah Koon et a1 1990, Angelakis et al., 1992
Hamamai et al., 1999, and Li et al., 2004,.). For instance, Angelakis et al., (1992) noted
that the higher water application rates result in a decrease in the vertical component of
the wetted soil volume (Figure 4.6).
30
60
90
120
Cumulative water infiltration (Liter)
Figure 4.6 Vertical advance of wetting front as a function of cumulative water

infiltration as measured by Angelakis et al., 1992
The relationships between the horizontal and vertical advance of wetting front with the
elapsed time were fitted as described in Equations 4.1 and 4.2. Unlike the sandy soil, the
a- values were found to be lower as the application rate increased, 6.16 and 5.3 for
application rates at 1 and 3.0 I/h, respectively. On the other hand, these values were not
showing any variance regarding the vertical advance of wetting front
- elapsed time
relationships. This contradicts with the findings of Hachum, et al., (1976) and Li et al.,
(2003) who found that an increase in the discharge rate resulted in a decrease in the
vertical component of the wetted soil depth. However, the rates of discharge and the soil
they studied were much different, which could explain such discrepancies.
For one-dimensional vertical flow into homogenous soils having a constant initial
moisture content, it has been shown by Philip (1957) that the advance of the wetting
front, z can be expressed as a power series in t'I2, or
Where h(8) are unique functions of soil moisture content 8, and t is the time. Phillip
showed that the power series in Eq. (4.3) converges rapidly. Therefore, only the first few
terms are needed to provide as accuracy required in practice. Figure 4.7 shows the
vertical advances of wetting front i.e., z at H=O, versus the square root of elapsed time
for river sand and sandy loam soil. As shown from this Figure, the relationship between
z and t'" is almost linear for different water application rates. This result indicates that
for the two-dimensional flow of water in soil, the vertical infiltration Eq. (4.3) using the
first term of equation, or
where A is a parameter which depends on the soil properties and water application rate,
having dimension of l/tIn.
q= 3.40 Ilh
- - -Z= 6.91 t
R~= 0.86
0
River sand
River sand
1 0
/
"
12
Square root of elapsed time (min
-2
-
15
12
15
Square root of elapsed time (min 05)
5)
100
q=l Ilh
90
8 0
---2=2.80to5 F?=0.99
70:
Sandy loam
0)
c
.-
60.
50
/*
a,
40
'
'
B
-m
.B
r
30.
10:
20
/
X)
ff
0
0
'
0
'
P
s'
5
10
15
20
Square root of elapsed time (min O.5)
5)
Figure 4.7: Vertical advance of wetting front versus square root of elapsed time for
river sand and sandy loam soil
With respect to river sand soil, the A-values were found to be 6.16 and 6.9l~mlmin"~
under emitter discharge rates of 0.75 and 3.4 l/h, respectively. In case of sandy loam
soil, these values were 2.8 and 4.0 cm/minl" for 1 and 3.0 l/h application rate,
respectively. In contrast to the vertical advances of wetting front, the linear relationship
does not hold between the horizontal advances of wetting fronts and the square root of
elapsed time (Figure 4.8). This has been discussed by Bruce and Klute, (1956) Grander
and Mayhugh (1958), Hachum et al., (1976) and Levin et al., (1974) and Muluneh et al,
(1995) who indicated that the lateraI water advance will approach a limit at large values
of water application time.
q= 0.750 Ilh
- - - Fitted 1.32 f
F?= 0.31
20
/
/
0)
.-K
w-
River sand
/'6
0
a,
>
10
m
m
/
0
12
15
-5 45
'
q=
- - - H=2.49 to F? = 0.70
I
6
q= 3 40 Ilh
//
0 1
I
I
I
/
/
'0
6
.-N
River sand
'9
15
n
"
0
12
15
Square root of elapsed time (min '-=)
5,
I lih
=: 40. - - - ~ = 1 . 7 7 t O ~F?=0.56
K
:
35.
.-5 30.
0
/
/
/
25'
/
/
9 20:
/d
73
.-
/ 'O
m
3 15
/o/
2
8
L
OO,~'
10'
0
/ / /
Sandy loam
/'
"
0
'-5)
10
15
20
0.5)
Figure 4.8: Horizontal advance of wetting front versus square root of elapsed time for
river sand and sandy loam soil
Moisture Distribution
The results of soil water distribution are illustrated in Figure 4.9 for river sand and sandy
loam soil, respectively. Similar water content on both sides of the central axis was
assumed. Figure4.9 shows the contours for the water content (i.e. the lines along where
the water content is equal). Contours were drawn using Kriging option (Gridding
Method) of the Graphic Software SURFER Version 7, (1999). The water content
increment between each contour is 0.05 cm3 cm
-3
. The irregular shapes of wetting
pattern may have been caused by the variability of the bulk density as a result of packing
the soil in the container. In highly permeable river sand soil, the wetting pattern is
elliptical with the wetted depth larger than the wetted radius. The elongated distribution
of moisture content in river sand in the vertical direction probably reflects the effect of
gravity forces. Preferential flow channels may also induce such elongated Gardenas et
al. (2005) attributed the vertical elongation of wetting front in sandy soil irrigated by
drip line to the preferential flow channels in the soil. The moisture content rapidly
decreased in horizontal direction while in vertical direction this decrease was more
slowly. The maximum moisture contents were found below the point source and in the
radii of 15 cm. In case of sandy loam soil, the moisture distribution was more
hemispherical rather than elliptical shape. Sandy loam soil has a smaller pore sizes
resulting in more lateral spread of applied water occurred to the capillary forces, which
was apparently quite different from the distribution for river sand. The maximum
moisture content was stored within radii of 30 and 40 cm and depths of 40 and 50 cm for
1.0 and 3.0 llh, respectively. However, the Figures indicate that the water content
steadily decreases with depth along any vertical line in the soil profile.
Distance (cm)
Distance (cm)
10
20
30
40
50
1
t
P P
t
C
I
90 I
Sand river q = 0.75 1h

Elapsed time = 120 min
Distance (cm)
Sand river q = 3.40 llh

D istance (cm )
Sandy loam q = 1 .D l/h

Elapsed t i m e = 270 min
Figure 4.9: Volumetric moisture distributions in soil at different discharge rates for river
sand and sandy loam soil
4.1.1.2 Wooden Box Experiments

Wetting Front Movement
The wetting fronts were observed for river sand soil under application rates of 3, 5.5 and
7 liters per hour for 380, 300 and 180 minutes, respectively. No saturated zone was
noted below the emitters as a consequence to the higher saturated hydraulic conductivity
of this soil. Figure 4.10 shows the progress of the wetting front, as a surface wetted
radius, for river sand soil with 3, 5.5, and 7 l/h emitter discharge rates. It was observed
that the wetting front immediately below the point source moved very fast at the
beginning, and then it slowed down as duration of irrigation increased. The wetting front
in the later hours was not moving as fast in all three cases.
Elapsed time (min)
Figure 4.10: Surface wetted radius as a function of elapsed time for river sand under
different application rates
The relation between cumulative infiltration water and surface wetted radius is shown in
Figure 4.1 1. It should be noted that the location of the surface wetted radius is shown
(Fig. 4.1 1) as a function of the cumulative amount of infiltration water and not as a
function of the cumulative amount of infiltration water and not as a function of elapsed
time. This is done in order to place the different application rates on an equal basis. This
procedure was suggested also by many studies Bresler et al., (1971); Li et al., (2003) and
Li et al., (2004).There appears to be a strong relationship between cumulative water
volume applied and the wetting fi-ont movement. This suggests that for a certain volume
of water applied a corresponding surface wetted radius can be predicted.
A least squares regression analysis was performed to the cumulative water volume
applied and the wetting front movement. For each group of flow rates, the regression
equation would be of the form:
Where a and b are constants determined from the regression analysis, r is the radius of
wetting front at the surface and V is the cumulative water volume applied.
q= 3.0 llh - - - Fttted

q= 5.5 1/h
Fltted
q= 7 0 Ilh - Fltted
Cumulative water applied (liter)
Figure 4.1 1: Surface wetted radius as a function of the cumulative water applied for river
sand under different application rates.
The results of this analysis are shown in Table 4.1. The determination coefficient, R2, for
this analysis was 0.94, which indicates a high correlation between the surface wetted
radius and the cumulative water volume applied. Examining Table 4.1 which yielded
from the fitting data with Eq. (4.5) reveals that the surface wetted radius is proportional
to the volume applied with power value approaching to 0.3. This result is in agreement
with Bar-Yosef and Sheikholslami's (1976) and Li et al. (2003) findings. They reported
that surface wetted radius is proportional to (qt) 'I3:
The depth of the downward soil water movement helps to estimate the depth for the
leaching of salts including applied fertilizers. The maximum wetted depth under the
point source when irrigation was terminated was 84 cm for the 7 llh application rate,
while the lower one was 72 cm for 3 l h .
Table 4.1: Parameters for estimation of surface wetted radius (Wooden Box
experiments)
Hence the maximum wetted depths are higher than the maximum surface wetted radius
for all application rates. This was expected because of the additional gravity forces on
the water movement through light soil when compared to the capillary forces.
Moisture Distribution
Spatial soil-moisture distribution was monitored at the termination of water application.

The soil samples were only taken on one side of the wetted profile in 6-centimeter
increments. Figure 4.12 shows the gravimetric moisture distribution for 3.0, 5.5 and 7.0
l/h application rates. Again, the duration was 380, 300 and 180 minutes for, 3.0, 5.5 and
7.0 l/h application rates respectively. Moisture decreased with depth under the point of
application and also in the horizontal direction. It can be seen that more moisture is
stored within 20 cm around the point source. This increase is correlated better to the
cumulative water applied than to the discharge rate. Muluneh et al. (1995) and Li et al.
(2004) reported similar results where the maximum moisture content noted in the
distance close to the point source.
90
Figure 4.12: Volumetric moisture distribution in river sand at different discharge rates:
(A) 3.0, (B) 5.5 and (C) 7.0 l/h
4.1.2 Field experiments
4.1.2.1 Wetting Front Movement
Both surface wetted radius and the vertical advances of wetting front were observed
visually through conducting the experiments in sand and sandy loam soil under field
conditions.
A total of 19 trials were conducted for sandy soil using average flow rates of 1.5, 2.1,
2.6, 3.5, 4, 4.5 and 6.0 l h . The average flow rates were all within 10% of the treatment
flow rates, where the duration of each test was between 280 and 460 minutes .Close
observations indicated that the actual water source did not behave as an ideal point
source, but water was spread over a finite circular water saturated area. The border of the
saturated area was determined visually as the point where the free-water was indicated
by where the glistening surface zone ended on the soil surface. It was also noted that the
saturated zone varied according to the discharge rate. Figure 4.1 3 displays the final
saturated radius versus emitter discharge rate. In all the trickle irrigation experiments
depicted in Figure 4.13, the steady radius values were achieved after about 25 minutes. It
is clear from the Figure that the saturated radius increased as the discharge rate increased
where the maximum value of 7 cm was under 6 l h application rate. However, the rate of
increase and limiting extent of the central saturated radius is a function of soil type and
application rate (Bresler et al., 1977 Clothier et al; 1983., Shani et al., 1987; Rev01 et al.,
1991; Yitayew et al., 1998; and Salem et al., 2002).
Emitter discharge rate (Ilh)
Figure 4.13: Final saturated radius as related to emitter application rate for sandy soil
It can be stated here that the information of the saturated radius as a function of emitter
application rate can be used to estimate the saturated hydraulic conductivity. Regression
analysis was performed using the following Wooding's equation to find ks.
where Q is discharge rate of the point source (L?
T-I),
k, is the saturated hydraulic
conductivity (L T-I), r, is steady-state saturated radius (L), a is a constant characteristic

of the soil related to soil sorptive properties and describes the rate of reduction in
conductivity with matric head (L").
Figure 4.14 depicts the relationship of flux densities, q, versus the reciprocal of steady
state radii llr,, produced by the point source for sandy soil. Values of k, and a are shown
in Figure 4.14. The determination coefficient of the linear regression describing the
relation between q and llr, is relatively good.
60
50
y=11246x+1546 ~ ~ ' 0 . 7 5
Ks= 15 46 cmlh
alpha= 0.175Cm
,/
.-
_,I,,--
,/'
.*-
-C
40-
.,-_I-
.-
0,:'-
tn
c
,-'
'
a,
X
3
30-
LL
20
10 . . . . ' - . . . ' . . . . ' . .

0.10
0.15
0.20
0.25
" " " '
0.30
0.35
Figure 4.14: The relationship of flux densities, q, versus the reciprocal of steady-state
radii, llr, for sandy soil.
There is good agreement between k, values as produced by the point source method and
that produced by constant head method (difference of 26%). The discrepancy maybe
attributed to the impact of surface topography where no guarantee that point source
infiltration area would be exactly circular.
The boundaries of the wetting fronts were reasonably well defined and were surrounded
by drier soil. Figure 4.15 shows the spread of the wetting front with time for the sandy
soil profiles under application rates of 2.0,4.5,4.8 and 6.0 l/h. The surface wetted radius
below the point source moved very fast at the beginning of irrigation, then it slowed
down as duration of irrigation is increased. The increasing hydraulic gradient at the

beginning of water application allows the lateral movement to spread very fast; by
continuing applying the water, the soil becomes wet which induced decrease of
hydraulic gradient, slowing down the advance.
Distance.x (cm)
Distance, x (cm)
Distance, x (cm)
Distance, x (cm)
Figure 4.1 5: Wetting front advances for sandy soil under trickle discharge rates of 2.0,
4.5,4.8 and 6.0 l/h
Figures 4.16 and 4.17 display the surface wetted radius (SWR) for each group of
discharge rate in centimeters versus elapsed time in minutes. As can be noted, these
relationships took the form of nonlinear relationship where the rate of surface wetted
radius advance decreased with the elapsed. However, the surface wetted radius reached
steady values at the latter of elapsed time. Regression analysis of the eIapsed time and
the surface wetted radius were performed. These analysis yielded power equations as
described by Eq. 4.1. The results of the regression analysis were plotted as dashed line
and the regression equations are shown inside each Figure for each trial. For each group
of discharge rate fiom 1.5 to 6 l/h, the regression equation gave high determination
coefficient ranging from 0.97 to 0.99.
To be able to compare the effect of trickle discharges rate and to place the different rates
on an equal basis, the result are compared for an identical amount of cumulative
infiltration and not for an equal time allowed for the irrigation (Figure 4.18).The surface
wetted radius is increased by increasing the total water applied. However, for most
application rate groups, the surface wetted radius approaches finite values "around 30
cm". The statistical analysis for given volumes of water applied 5 , 10 and 12 liters with
application rate groups (2.1, 2.6, 3.5, 4.0, 4.5, and 6.0) is given in Appendix A (Table
A.l). It can be indicated that the surface wetted radius is not affected by the discharge
rate, where F
calculated
<F
critical
at 0.05 significance level. The cumulative volume of
applied water would appear to be more influential on the size and shape of the wetted
soil volume in soils with the same texture. The statistical analysis showed the volume of
applied water has a high significance on the surface wetted radius advance i.e., F calculated
> F critical-
Sandy Soil q= 1.5 Ilh
50
100 150 200 250 300 350 400 450
50
100 150 200 250 300 350 400 450
Elapsed time (min)
Elapsed time (min)
Elapsed time (min)
Elapsed (min)
35,-
q= 4.0 Ilh
"
0
50
100 150 200 250 300 350 400 450
Elapsed time (min)
50
100 150 200 250 300 350 400 450
Elapsed time (min)
Figure 4.16: Surface wetted radius as a function of elapsed time for sandy soil under
1.5,2.1,2.6, 3.5,4.0 and 4.5 l/h application rates.
Elapsed time (min)
Figure 4.17: Surface wetted radii as a function of elapsed time for sandy soil under
6.01lh application rate.
35
-Eo
-2
3 0 ~
a
6
:
"
I . .
.- 25
OD"
a?-.
o.
-0
. .
A~~ -
do
20
10
15
20
q=1.5 11h
q=2.1 Vh
q=2.6 11h
q=3.5 11h
q=4 0 Ilh
q=4 5 11h
q=6.0 11h
25
30
35
Figure 4.18: Surface wetted radii as a function of cumulative water applied for sandy
soil under different application rates
This result is obviously in agreement with the finding of Armstrong and Wilson (1983),
which showed that the size and the shape of the wetted zone beneath an emitter is more a
function of the cumulative amount of water applied than the rate of application for
application rates of 3.9 to 161 l/h and application volumes from 3 1 to 160 liters.
The same form of regression equation, Eq. 4.5, was obtained by using regression
analysis of the water applied and the surface wetted radius advances for each group of
application rate and all flow rates combined. Table 4.2 lists the best fitted parameters of
Eq 4.5. The high values of determination coefficient (~*>0.96)for each discharge rate
group indicates that the fitted equation is appropriate to describe the relationship
between surface wetted radius and cumulative volume of water applied. The slope of the
equation decreased as discharge rate decreased. Conversely, the exponent values were
found to be increased by increasing the application rate.
Table 4.2: The regression results of the surface wetted radius as a function of cumulative
water infiltration for sandy soil (field experiments), SWR=a vb
Flow rate (lh)
R~
6.0
10.34
0.32
0.98
All
12.15
0.29
0.96
The vertical advances of wetting front were traced on the plexiglass sheet at various time
of irrigation. Figure 4.19 shows the vertical advances of wetting front versus the elapsed
time for 2.0, 4.5, 4.8 and 6 l/h application rates. For all application rates the vertical
advances of wetting front are in non-linear relationship with the elapsed time. Unlike the
surface wetted radius, the advance of vertical wetting front continues progressing
steadily throughout the irrigation. For surface wetted radius advance, the potential for
water movement decreases and reaches a finite value. This does not hold in the vertical
direction because of the gravity effect. The gravity forces keep the potential gradient
from approaching zero, thus water movement continues downward after lateral
movement reaches a finite value. It was found that Eq. 4.2 represented the best formula
that can describe the relationship between the vertical advances of wetting front and the
elapsed time.
Elapsed time (min)
Figure 4.19: Vertical advances of wetting front versus elapsed time for sandy soil under
different application rates
Table 4.3 shows the regression results of vertical advances of wetting front as a function
of the elapsed time. The exponent value of Eq. 4.2 was found to be 0.56. This value
represents the exponent value for all application rates combined. Li et al. (2003) and
Elmaloglou and Nalmamos (2003), in their study about water distribution from point
source they found that the exponent ranged around the value of 0.50. As discussed
previously, it was found that the vertical advance of wetting front is a function of the
square root of elapsed time (Eq. 4.4)
Figure 4.20 shows the vertical advances of wetting front as a function of the square root
of elapsed time under different application rates. This Figure indicates that there is an
excellent linear relationship between the vertical advances of wetting front. Hachum et
al. (1976) in their study of two dimensional water flow found that a is a linear h c t i o n
of the water application rate and square root of elapsed time. a -values were found to be
2.91, 3.32, 3.24 and 3.85 cm/minO.'with respect to the trickle discharge rates of 2, 4.5,
4.8 and 6 11, respectively. This means, in soil with similar structure, the vertical water
advance can be calculated by Eq. 4.4 with estimated a- values based on the information
of water movement under different application rates.
Table 4.3: The regression results of the vertical advances of wetting front as a hnction
of the elapsed time (Sandy soil-field experiment). z=a tb
2.0
4.5
4.8
6.0
All
The effect of cumulative water applied on the vertical advances of wetting front was
observed. Figure 4.21 shows that the vertical advance of the wetting front is increased
by increasing the cumulative water applied for all application rates 2, 4.5, 4.8 and 6 l/h.
For 10 liters of cumulative applied water, the lower application rate induces greater
depth of the wetting front. However, this was also noted under larger amount of
cumulative applied water.
-,
V)
8
g
>
30
-0
q= 2 11h
mm
20
gr
I0
q= 4.5 Ilh
"- - - Fitted
R2 = 0 97
z= 2.93 t
---
Fitted z = 3.32 t
R2= 0.97
I
01
5
10
15
20
25
30
5
10
15
20
25
30
0
0
0
Square root of elapsed time (min)
Square root of elapsed time (mino5)

80
70
1
/
/D
60
m
5
5
50
40
/
/
a,
0
30
30
/
/
73
10
n
"
20
L-
20
0
.-
V)
mm
50 .
/n
C, 40
g>
/ID
q= 6.0 Ilh
q= 4.8 Ilh
Fitted z= 3.23 t
10 :
R2 = 0.94
- - - Fitted z= 3.857 t
R = 0.95
0
5
10
15
20
25
Square root of elapsed time (min0-I)
30
10
15
20
25
30
Square root of elapsed t~me(mino5)
Figure 4.20: Vertical advance of wetting front versus square root of elapsed time at
different time for sandy soil under application rate of 2.0, 4.5, 4.8 and
6.0 11%
For instance, the maximum depth of wetting front was found to be 64 and 62 cm for 4.8
and 6 llh, respectively, when 28 liters of water was applied. The deeper wetted depth in
smaller discharge rate treatments was caused by the fact that the water entry saturated
zone which is a function of application rate becomes larger as the application discharge
rate increases (Fig.4.13). When the water had spread over a larger surface, the
movement was not as deep for a given quantity of water. Moreover, the slower
application rate had a longer time to infiltrate into the soil profile. These results are in
agreement with previous findings where higher application rates result in smaller
vertical advance of the wetting front (Clair et al., 2003; Ah Koon et al., 1990; and
Schwartzman and Zur 1986; Angelakis 1992; Bresler 1971; Li et al., 2004).
100
90
80
70
60
0
'
50 ;
0
0
L3
Do
q= 2.0 11h
q= 4.5 11h
q= 4 8 Vh
q= 6 0 11h
i
10
20
30
40
Figure 4.21: Vertical advances of wetting front versus cumulative water applied for
sandy soil under different application rates
Figure 4.22 shows the vertical advance of wetting front and the surface wetted radius as
a function of cumulative water applied. At the beginning of irrigation, the surface wetted
radius was greater than the vertical advance. This can be noted for 2, 4.8 and 6 l//h
application rate. In case of 4.5 llh, no data for vertical advance were obtained at the
beginning of irrigation.
100
q=45Vh
90
80
Honzontat
Vert~cal
o
D
70.
D
0
60
D
D
50 D
40
000"
30
00
20.
o 0
10"
0
0
10
15
20
25
30
35
40
function of cumulative water applied for sandy soil under different
application rates
This is probably affected by the change in the size of the water entry saturated zone as a
function of elapsed time. At the binging of irrigation the size of water entry saturated
zone was very small and identical to the area formed at the soil -water drop interface.
Thus, a portion of the water discharging out of the emitter was not able to penetrate the
soil through this area (the vertical flux of water is greater than the saturated hydraulic
conductivity of the soil.) and because of conservation principle, a factor which increased
the horizontal advance of water. The rate of growth of the saturated water entry
decreases with time and then reached a steady value (earlier under slower application
rate) when the vertical flux of water equal the saturated hydraulic conductivity of the
soil. At that moment, water begun to move easily and quickly downward under the
effect of the gravity forces which then become more significant. Many studies indicated
that at the beginning of irrigation, the surface wetted radius would be greater than the
vertical advance (Clair et al., 2003, Ah Koon et al., 1990 and Schwartzman and Zur
1986, Angelakis et al., 1992 ,Bresler 1971, Li et al., 2004)
The surface wetted radii under field condition were studied for sandy loam soil. A total
of 18 tests were conducted using average flow of 1.5, 2.0, 2.5, 3.5 and 4.0 I h . The
average flow rates were all within 10% of the treatment flow rates where the duration of
each test was 300 minutes. It was observed that during each experiment, an area of
saturated zone developed in the vicinity of the trickle source. This area where water
infiltrated into the soil from the surface was initially very small, but its radius became
larger and reached a steady value within about 30 minutes. This saturated area, the size
which is a function of time, is the only place where water can infiltrate into the soil from
the surface (Bresler, 1977). Figure 4.23 shows the surface saturated radius as a function
of emitter discharge rate. High emitter discharge rate results in high saturated radius.
Many studies manifest such relation (Clothier et al., 1983; Shani et al., 1987; Rev01 et al.,
1991; Yitayew et al., 1998 and Salem et al., 2002).It is clear from the Figures of 4.13 (for
sandy soil) and 4.24 for (sandy loam soil) that the saturated wetted radius becomes larger
as the emitter application rates increases and the saturated hydraulic conductivity
decreased. This may explain the increase in the horizontal component of the wetted area
and the decrease in the vertical component of the soil wetted soil depth.
Emitter discharge rate (Ilh)
Figure 4.23: Final saturated radius as related to emitter discharge rate for sandy loam
soil
As described earlier, the relation between saturated radius and emitter application rate can
be used to estimate the saturated hydraulic conductivity. Figure 4.24 shows the
relationship between flux densities, q, versus the reciprocal of steady-state radii, l/r,. The
value of 7.5 cmlh for the saturated hydraulic conductivity k, as estimated from the
described method provides a valid estimation where the estimated k, from laboratory were
found to be 5-8 cmlh.
Figure 4.25 shows the advances of wetting front at different time under different
application rates. The horizontal water movement, after water application, was less than
the vertical advance. For all application rates, the surface wetted radius increases
monotonically as a function of elapsed time.
Figure 4.24: The relationship of flux densities, q, versus the reciprocal of steady-state
radii, l/rs
However, the surface wetted radius reaches a constant value and does not increase
continuously as a function of the elapsed time. This can be seen clearly in the case of
application rates of 1.5 and 2.0 l/h where 21.5 and 24.5 cm was observed as steady values
of surface wetted radius, respectively.
A regression analysis of elapsed time and surface wetted radius was made. This analysis
yielded power equation a described by Eq. 4.1. The results were plotted and shown in
Figure 4.26. As presented in this Figure, the determination coefficients ranged between
0.96 and 0.99. The high value of R~ indicates that the regression equations were
appropriate to explain the relationship between the elapsed time and surface wetted
Distance ,x (cm)
Distance, x (cm)
-50 -40 -30 -20 -10
Distance, x (cm)
10
20
30
40
50
Distance, x (cm)
330 rnin
Figure 4.25: Wetting front advances for sandy loam soil under trickle discharge rates of
2.3,2.6, 3.0 and 4.0 l/h
radius. As mentioned previously, the parameters of Eq. 4.1 depends on the emitter
discharge rate. Furthermore it can be observed that a-values increase as discharge rate
increased. For example, a value was found to be 6.93 under application rate of 1.5 l/h
while it reached 8.73 under 4 l/h application rate. Many studies are in agreement with
these findings (Li et al., 2003, Muluneh et al., 1995; and Huchm et al., 1974).
,,_e-CTo
,a
20
p0
'
d
B
15
9'
p/
q= 1.5Ilh
lo P
Observed
F~tted SWR= 693 t
o Observed
Fitted SWR= 6.79 t 0232
OZo6
f
100
200
300
400
100
Elapsed time (min)
--
Fitted SWR= 8.31 t
02'
100
200
400
R = 0.99
300
400
100
200
300
400
Elapsed time (rnin)
Ela~sed
time Irnin)
300
Elapsed time (min)
200
100
200
300
400
Elapsed time (min)
Figure 4.26: Surface wetted radius as a function of elapsed time for sandy loam soil
under application rates of 2.0,2.5, 3.5and 4.0 llh.
Figure 4.27 shows how the surface wetted radius as a function of the amount of water
applied for all application rates combined. For a given volume of water applied, it can be
noted that increasing of application rate would slightly increase the surface wetted
radius. These can be explained by the fact that as the water application rate increases, the
area of saturated soil near the trickle source increases; this in turn enhances the
horizontal advance of the wetting fronts. It is apparent that the surface wetted radius can
be expressed as a function of volume of water applied by the power relationships, Eq
4.5. The coefficients for this equation, as well as the values of R~ criterion, are
summarized in Table 4.4. The values of a fluctuate from 14.76 to 16.28 showing no clear
tendency can be attributed to the effect of discharge rate on surface wetted radius.
Furthermore, R~ which ranged between 0.97 and 0.99 indicated high correlation between
surface wetted radius and volume of water applied. Statistical analysis of the effect of
both emitter discharge rate and volume of water applied on the surface wetted radius are
shown in Appendix A (Table A.2). Three application rates 2.5, 3.5 and 4 l/h and three
volume of water applied 5, 10 and 12 liters were chosen to this analysis. The other
application rates 1.5 and 2.0 l/h were excluded for this analysis because of the lack of
the relevant data. Like sandy soil, the effect of application rate on surface wetted radius
was not significant, i.e. F
calcurated
<F
Crit.
On the other hand, the effect of volume of
water applied were found highly significant on surface wetted radius F calculated > F crti. AS
mentioned previously, Armstrong Wilson (1983) reported that surface wetted radius is a
function of volume applied. However, this result is very useful for practical purpose
where using higher application rate, higher amount of water in shorter time, will insure
reduction of irrigation time and energy use.
10
12
14
16
18
20
22
Cumulativeof water applied (liter)
Figure 4.27: Surface wetted radius as a function cumulative water applied for sandy
loam soil under different application rates
Table 4.4: The regression results of the surface wetted radius as a function of the total
water applied for sandy loam soil. SWR=a vb
Flow rate (l/h)
R~
1.5
2.0
2.5
3.5
4.0
All
The vertical advance of the wetting front for sandy loam soil was detected with the
experimental apparatus. Figure 4.28 shows vertical advance of the wetting front versus
elapsed time for 2.3, 2.6, 3.0 and 4.0 1/h as combined. It can be seen that the vertical
advance of wetting front continues as time irrigation progress. This behavior was not
like that with surface wetted radius where its advances reached a steady value. A
regression analysis was made to the elapsed time and the vertical advance of wetting
front for each flow rate trial and all application rates combined. The results are shown in
Table 4.5. It appears that the values of the exponent ranged between 0.35 and 0.5 1. The
exponent value for all application rates combined equals 0.41 or approaching a value of
0.5 as suggested by Li et al., (2003) and Elmaloglou and Nalmamos., (2003). However,
the rates of discharge they used and their soil maybe explain such differences.
Elapsed time (min)
Figure 4.28: Vertical advances of wetting front versus elapsed time for sandy loam soil
According to the Eq 4.4, the relation between the vertical advance of wetting front and
the square root of the elapsed time is expected to be linear. Figure 4.29 shows the result
of plotting the vertical advance of wetting front versus the square root of elapsed time.
Table 4.5: The regression results of the vertical advances of wetting front as a function
of the elapsed time (Sandy loam soil-field experiment). Z=a tb
2.6
3 .O
4.0
All
The linear relationship between vertical advance of wetting front and square root of
elapsed time fits better, which indicate that Eq. (4.4) can be used to describe the vertical
movement of water through the soil profile under surface trickle irrigation. Furthermore,
the slope of the fitted straight line in Figure 4.29 increased as application rate increased.
This concludes that the vertical advance of wetted pattern would increase by increasing
of application rate.
Clearly, the volume of water applied affects the width and the depth of the wetted soil
volume, and therefore influences the optimal emitter spacing and the depth of roots. The
influence of water application rate on vertical advances can be deduced, where the
increase in the application rate results in decreasing the vertical advance of wetting
fronts. This was a consequence of a larger saturated water entry zone (Fig. 4.23) for the
higher application rate. When the water had spread over a larger surface, the movement
was not as deep for a given quantity of water. This can be noted clearly in Figure 4.30.
For application rates of 2.3, 2.6,3 and 4 llh, the vertical advance of wetting front was
found to be 26, 27 22 and 24 cm, respectively, by using the same volume of water i.e. 4
liters. This tendency is assured for higher amount of water (1 0 liters) where the vertical
advance of wetting front was found to be 39, 38, 35.8 and 33.5 cm for 2.3, 2.6, 3.0 and 4
llh, respectively. These are in agreement with previous findings where higher water
application rates result in smaller rates of vertical advance of wetting front.
gm
>
U
m
U
.-
20
q= 2.3 Ilh
---
Observed
Fitted z= 2.48 P
F=0.94
1
0
Square root of elapsed time (min O.5)
10
15
20
25
5)
25
P
,6
0
,d
q= 3.0 l/h
Observed
- - - 2=2 53 t
R2=0.99
5)
10
15
20
5)
Figure 4.29: Vertical advance of wetting front versus square root of elapsed time at
different time for sandy loam soil
Figure 4.30: Vertical advances of wetting as a function of cumulative water applied for
sandy loam soil under different application rates
Figure 4.31 shows the distance of the wetting front from the trickle source which is at
point (0, 0) as a function of the amount of applied water for the application rates of 2.3,
2.6, 3.0 and 4 llh. For initial addition of water, the surface wetted radius surpassed the
vertical advance of the wetting front. This can be noted in the case of application rate of
3.0 l/h and 4.0 llh. For other application rates, no data for vertical advance of wetting
front was available for comparison at the beginning of irrigation.
This is probably, as mentioned previously; affected by the change in the size of the
water entry saturated. Continuous water application would continue increasing vertical
advance of the wetting front and decelerating the wetted surface advance. As discussed
previously, the wetting front moves because of the potential water gradient. Gravity
force adds to the vertical potential gradient, thus result in greater vertical movement than
lateral movement.
50
0
o o
-6
-8
0
0
0
0
0
0
0
0
q= 2.3 I/h
o O
O
O
Vertical
Horizontal
s
m
5
u
40 -
0
D
D
30-
zs
2
o
0
20 -8
0
0
10
0
0
q= 2.6 ilh
p
0
Vertical
Horizontal
. . . . . . . . . . . . . . . . . . . . . . .
0
0
10
15
20
25
Volume of water applied (liter)
q= 4.0 I/h
Vertical
Horizontal
0
0
10
15
20
25
function of cumulative water applied for sandy loam
Figure 4.32 shows the horizontal advance of water as a function of volume applied for
both soils at four different water application rates 2.6, 3, 3.5 and 4 llh. For any amount
of water application, the surface wetted radius in the sandy loam is greater than that in
the sandy soil. To consider an example of the effect of soil type on the surface wetted
radius advance, let us take treatment of 3.5 llh. It can be seen that the surface wetted
radius were 27 cm and 32 cm for sand and sandy loam soil, respectively at 18 liters of
application water. For 4.0 1/h, the surface wetted radius was 29 cm and 31 cm,
respectively, for the same amount of water application. The gravity as additional force
dominating the water movement in the sand decreasing the rate of the horizontal
advance and resulting in smaller horizontal advance in the sand than in sandy loam soil
in which the capillary forces have the main significant influence on the water movement.
The data presented in Figures 4.13, 4.223 and 4.32 demonstrate clearly that the
application rate and the saturated hydraulic conductivity of the soil have a remarkable
effect on the saturated surface wetted radius. Increasing the application rate and
decreasing the saturated hydraulic conductivity result in increase in the saturated wetted
radius which in turn results in an increase in the horizontal component of the wetted
area. This finding is in concurrence with the literature. Bresler (1977) Angelakis et al.,
(1993) Yitayew et al., (1998) found that the wetted surface radius increased by
decreasing the soil saturated hydraulic conductivity and increasing application rates.
Figure 4.32 also illustrates the effects of soil water properties and the rate of discharge
on the selection of spacing between emitters. Larger spacing is permitted in soils with
lower saturated hydraulic conductivity. Closer spacing is required for soils having higher
saturated hydraulic conductivity. In addition, the emitter spacing can be increased as the
discharge rate becomes higher. However, since the rate of growth of surface wetted
radius decrease with increase in the discharge application rate, the proper choice of
emitter application rate chiefly depends on some optimization criteria in the engineering
design of the field irrigation system.
q=3.0 Ilh
0
l
Sand
Sandy loam
e e e
e *
e *
e * *
0 0
0
0
0
: e
a0
0
r
D
0
l
q=3.5 Ilh
Sand
Sandy loam
q= 4.0 Ilh
Sand
Sandy loam
Figure 4.32: Surface wetted radius in sandy and sandy loam soil as a function of
cumulative water applied under different application rates
Figure 4.33 shows the vertical water advances versus the volume of water applied for
sand and the sandy loam soils at different water application rates. For both soils, vertical
advance of the wetting fronts increased with increasing amount of water applied and
decreasing water application rate. It can be also seen that the slower application rate of
2.0 l/h in sandy soil results in higher vertical advance of wetting front for a given
amount of water application compared to 2.3 l/h in sandy loam. On the other hand, as
stated previously, increasing the discharge rate and decreasing the saturated hydraulic
conductivity result in an increase in the horizontaI component of the wetted area and a
decrease in the vertical component of the wetted soil depth. This is probably affected by
the changes in the size of the water entry saturated zone for each soil type and rate of
discharge (Fig. 3.13 and Fig. 4.23). This saturated zone becomes larger as the soil
becomes less permeable and as the application rate increases.
100.
o Sand 2.0 llh
90 :
0 Sand 4.5 11h

o Sand 4.8 Ilh
80 70 :
60
50
a Sand 6.0 llh

Sandy loam 2.3 Ilh
Sandy loam 2.6 llh
Sandy loam 3.0 Ilh
A sandy loam 4.0 Ilh
Oo
n
o
0
40 r
30 :
1
1 Om
A
A
.my*
#p.6
.:%B
o
a
A A
2 0 : ~ 2
10 8
0 0
. .
' . . . ~ ' ~ ~ ~ - ' " ' ~ ' - ' - ' ' ' ' - ' .~ ' " ' ' ~
10
15
20
25
30
35
Figure 4.33: Vertical advance of wetting front in sandy and sandy loam soil as a function
of cumulative water applied under different application rates
The results obtained from this study agreed with Bresler's (1978), where the combined
effects of emitter discharge, soil hydraulic properties and application time on the shape
of the wetted soil volume and the position of the wetting front indicate that the finer soil
resulted in more spreading at the surface. In addition, the higher rates resulted in a lesser
wetted soil depth (Fig 4.34).
R A O f A t DISTANCE, r l c n j
Figure 4.34: Wetting front position as a function of discharge rate q and cumulative
infiltration water in liters (the numbers labeling the lines, Bresler; 1978)
The wetted surface is small while the vertical advance of wetting front d long if emitter
discharge is small when the volumes of water applied are the same. The root distribution
ranges are affected by soil wetted patterns. Therefore, a higher application rates may be
better when one is considering the crops to get enough water while infiltration loss is
rare when crops are small (young) if trickle laterals are installed in the center of two
crop rows. This is because crop roots distribute shallowly and far from the emitters in
this stage. A slower application rates may be better when crop is bigger according to the
above findings.
4.1.2.2 Moisture Distribution
The shape and moisture distribution within the wetted soil volume from a point source is
important for an efficient design, operation and management of trickle irrigation
systems. It is within this volume that a plant obtains water to meet its nutrient
requirements. A maximum crop yield per unit of water applied maybe obtained by
placing it in a soil volume where the plant roots have easy access to water and nutrients.
The gravimetric moisture contents within the wetted profile are shown in Figure 4.35
and 4.36 for application rates ranging from 1.50 to 6.0 Ilh. The cumulative amount of
water applied were 8, 12, 12, 19, 24, 34, 29 and 34 liters for 1.50, 2.0, 2.50, 2.70, 3.50,
4.50, 4.80 and 6.0 l/h application rates, respectively. As can be seen from these Figures
the contour plots of the moisture content patterns are irregular and non-uniform. This
could be explained by the natural variability of the soil and by the fact that the times
required for sampling the soil profile for the determination of moisture contents may
induce redistribution of moisture content. The maximum moisture content was found
just below the irrigation source and decreased laterally as the distance from the irrigation
source increased. The rapid decrease in volumetric moisture content laterally away from
the application source may be due to the limited effect of capillary forces in sandy soil.
In fact, the tension values increase negatively at further distance from the point source.
This tension gradient is not a linear function. This is because the unsaturated hydraulic
conductivity K (h) is also a non-linear function of moisture content or tension
(Figure 2.1). Thus the rate at which moisture moves through the soil depends on the
moisture content in the soil. As the moisture content increases, the rate of moisture
Horizontal distance (cm)
-80 1
Vw= 8 liters
t
V,= 12 liters q= 2.0
q= 1.5 I/h
-80
-80
Vw= 19 liters q= 2.7 Vh
I
I
Figure 4.35: Volumetric moisture distribution in sandy soil at different discharge rates of
1.5,2.0,2.5 and 2.7 l/h
V,= 24 liters 9= 3.5

- 80
Horizontal dlstance (cm)

0
10
20
30
40
50
10
20
30
40
50
Figure 4.36: Volumetric moisture distribution in sandy soil at different discharge rates of
3.5,4.5,4.8 and 6.0 l/h
movement increases. Thus, the tension gradient would be quite flat near the point source
and become quite steep near the wetting front. The moisture distribution would also
follow the same trend. The moisture content near the point would be quite similar and
then decrease quite rapidly near the wetted front. This relationship can be seen by
referring to Figures 4.35-4.36.
The moisture contents for the first 20 cm show slight increase along the radii and then
decrease at a faster rate as it approaches the wetted front. The tension gradient would be
different in the horizontal and vertical directions. The tension gradient in the vertical
direction would include the effect of gravity which would help increase the gradient.
Hence, as expected, the moisture content was more distributed in the vertical direction
compared to lateral direction which attributed to the gravity forces and to the higher
conductivity of soil.
In order to minimize the effect of application rate on the distribution of moisture content
in the soil profile the cumulative volume of applied water was held constant i.e., 12 liters
for the treatments of 2.0 and 2.5 l/h and 34 liters for the treatments of 4.5 and 6.0 l/h.
The average moisture content for the treatments 2.0 and 2.5 I/h were 0.177 and 0.1885,
respectively. The result of the t-test indicated that there is no statistically significant
difference in the two means of moisture content at 99% significant level (t- ratio = 1.42
compared to t
= 2.6 1, n=
5 8).
Figure A.2 shows the moisture content distribution at radial distance of 2.5 and 17.50 cm
along the vertical distance (z) i.e., [8 (2.5, z) and 8 (17.5, z)]. This Figure gives good
information about how water is distributed within the wetted soil volume and to
facilitate the comparisons between the treatments. The moisture content is higher at
locations closer to the source of irrigation in the case of application rate of 2.5 l/h
'
compared to the application rate of 2.0 l/h. At locations closer to the wetting fronts, the
moisture content is slightly higher under application rate of 2.0 l/h compared to the
application rate of 2.5 l/h. This trend is quite similar to the horizontal transects as
depicted in Figures A.3. This Figure shows the moisture content distribution at the depth
of 7.5 and 27.5 cm below the soil surface along the radial distance (r) i.e., [8(r, 7.5) and
8(r, 27.5)] respectively. It can be seen that the moisture content is higher under the
application rate of 2.5 l/h at the depth of 7.5 cm. On the other hand, at the depth of 27.5
cm, the moisture content is higher under application rate of 2.0 l/h at locations closer to
the source of irrigation and to the wetting front.
The effect of the water application rate was also detected for the application rates of 4.5
and 6.0 l/h. The cumulative volume of applied water for these treatments was 34 liters.
The average moisture content for the application rates of 4.5 and 6.0 l/h were both
approximately 0.175. The average moisture content then increased from 0.15 to 0.185 as
the application rates increase from 4.5 to 6.0 l/h. This result is contradictory with the
conservation law in which the average moisture content should be similar as the
cumulative water applied to both treatments were the same. The logical explanation for
such different is that each experiment was conducted under different location, and also,
the prior climatic conditions (the sampling of moisture content were taken at night for
4.5 l/h treatment and in the afternoon for 6.0 l/h treatment) and the length of test
conducted was different. This would result in different amount of water being
evaporated from the soil surface. This would also affect the moisture movement and
could cause the wetted profile to be different. On the other hand, when the trickle
discharge is relatively slow, the time between drops may be sufficiently long to cause
processes of alternate wetting and drying close to the soil surface (Bresler et al., 1971).
The water content, especially close to the water source at the beginning of infiltration
may be affected by hysteresis effect in the soil water content water suction relationships.
The hysteresis effect seems to retard the movement of water and to increase the water
content during redistribution (Youngs, 1961). Moreover, because of possible errors in
determining soil moisture, assuming symmetrical movement around the point of
application and the time of application for both treatments may cause such differences
.Even though the statistical analysis showed a significant difference between the two
means of moisture content at 99% significant level (t- ratio = 6.66 compared to t
critical =
2.59, n= 126), it was found that this difference is not significant at 99% significant level
within the wetted profile at vertical transects of i-e., [O (2.5, z) and O(17.5, z)] and radial
transects i.e., [ 0(r, 7.5) and 0(r, 27.5)]. Figure A.4 shows the moisture content
distribution at radial distance of 2.5 and 17.50 cm along the vertical distance (2) i.e., [O
(2.5, z) and 0 (17.5, z)]. Figure A.5 shows the moisture content distribution at the depth
of 7.5 and 27.5 cm below the soil surface along the radial distance (r) i.e., [O(r, 7.5) and
0(r, 27.5)], respectively. It is evident from these Figures that the moisture content is
same at lateral distance equals to 2.5 cm and with depth close to 20 cm; however the
moisture content was found higher at depths beyond 20 cm with higher application rate.
At lateral distance of 17.50 cm, the moisture content is higher under the slower
application rate at depths within the depth of 30 cm. On the other hand, this trend
becomes conversely beyond the depth of 30 cm whereas the higher moisture content was
found under application rate of 6.0 llh. For lateral transects, the effect of application rate
seems to be negligible on the moisture content and found to be equal within the depths
of 7.50 and 27.50 cm.
Ultimately the results presented previously clearly demonstrate that the application rate
has a remarkable effect on the vertical distribution of moisture content while this effect
is not notable on the horizontal distribution of moisture content. Sandy soil has bigger
pore sizes resulting in more vertical spread of applied water due to the gravity forces. Li
et al., 2004 reported the same results with sandy soil and under different application
rates ranged from 0.5 to 2.0 l k . The possibility of controlling soil moisture content and
the wetted volume of soil by regulating the emitter application rate is of practical interest
in the design of a field irrigation system. For instance, using higher application rate
would be more suitable for shallow root plants where less vertical depth is required
Figure 4.37 shows the moisture distribution within the wetted soil volume for various
application rates in sandy loam soil. The cumulative volume of water applied were 10,
17, 17 and 24 liters for 2.3, 2.6, 3.0 and 4.0 l k . It can be noted from the contour plots
that the volumetric soil water content distribution within the wetter volumes are not
uniform; it steadily decreased within the radial and the vertical distance from the
irrigation source. The lateral spread of moisture content was found to be higher in sandy
loam soil compared to sandy soil. On the other hand, the vertical spread of moisture
content in sandy loam was less than sandy soil. The limited effect of the gravity forces
on the moisture movement in the sandy loam soil as being finer textured soil is the main
reason that caused this type of moisture content distribution. Warrick (2002) attributed
the different wetting pattern during water infiltration within different soil type to the
pore size distribution and the shape of k (h) function. Normally, sandy soil would have a
much greater drop in hydraulic conductivity compared to sandy loam soil because sandy
loam soils have smaller capillaries that can conduct moisture movement at greater
tensions. Therefore, coarser-textured soil with a narrow pore-size distribution, the

wetting front is more abrupt and in a fine-textured soil, the wetting front is more diffuse.
For various application rates, it was found that the maximum volumetric moisture
content was stored within the depth of 35 cm and 25 - 30 cm radii from the emitter.
Horizontal distance [cm)
Horizontal [cm)
Figure 4.37: Volumetric moisture distribution in sandy loam soil at different discharge
rates of 2.3,2.6, 3.0 and 4.0 llh
The effect of application rate can be considered under the treatments of 2.6 and 3.0 l k
where the same volume of water was applied i.e., 17 liters. As shown for the treatments
with application rates of 2.6 and 3.0 llh, the higher application rate result in a quite
larger wetted pattern. For example, the horizontal surface spread for 2.6 l/h was about 27
cm. In the case of 3.0 l/h the horizontal spread was about 3 1 cm.
The wetted depth shows also small difference among both application rates, for example
for the application rates of 2.6 and 3.0 llh the wetted depths are at 42 and 48 cm,
respectively. However, the mean moisture content within the wetted volume for 2.6 and
3.0 l/h were 0.233 and 0.236, respectively.
The result of the t-test indicated that there is no statistically significant difference in the
two means of moisture content at 99% significant level (t- ratio
t ,,it
0.883 compared to
2.61 n= 64). Figure A.6 shows the moisture content distribution at radial distance
of 2.5 and 17.50 cm along the vertical distance (z) i.e., [0 (2.5, z) and 0 (17.5, z)],
respectively. Figure A.7 shows the moisture content distribution at the depth of 7.5 and
27.5 cm below the soil surface along the radial distance (r) i.e., [0(r, 7.5) and 0(r, 27.5)],
respectively. These Figures indicate very close agreement between the moisture content
distributions with similar trends for both application rates.
4.1.2.3 Solute distribution
Whenever soil water moves, solutes tend to move with it (Colthier and Sauer, 1988). In
soil trickle irrigated with water containing dissolved nutrients, the solutes in the
displacing fluid become distributed three dimensionally throughout the soil. The
speculation on the shape and chemical distribution with the chemigated soil volume
from a point source is important to disclose before design, operation and management of
a chemigation system. It is within this volume that a plant obtains the nutrients it needs
to meet its requirements. The results and discussions of the field experiments are
presented to understand the solute distribution under point source trickle irrigation.
The initial chloride content of the sandy soil was determined in the laboratory to be 0.2
meql 1. The inlet concentrations of chloride in the irrigation water (c,) were 54, 13, 35,
13,20, 54, 35 and 20 meqll for the application rates of 1.5,2.0, 2.5, 2.7, 3.5,4.5,4.8 and
6.0 llh, respectively. As stated in the preceding chapter, the samples used to determine
the moisture content distribution were also used to measure the chloride concentrations.
The initial volumetric chloride concentration (c, 8,) was found to be 0.01 meqlliter bulk
soil. The inlet volumetric chloride concentrations which are the maximum possible of
volumetric chloride concentration (c, 8,) at the soil inlet are 22.14, 5.33, 14.35, 5.33,
8.20, 22.14, 14.35 and 8.2 meq/liter bulk soil for the application rate treatments of 1.5,
2.0, 2.5, 2.7, 3.5, 4.5, 4.8 and 6.0 llh, respectively. Figures 4.38 and 4.39 show the
distribution of chloride concentrations for each application rate. The numbers within
each contour line represent the concentration of chloride as expressed in milliequivalents
per liter bulk soil. The patterns of the chloride concentrations distribution are similar to
those for moisture content distribution where the vertical spread is more than horizontal
for all application rates. Kirda et al. (1973) studied the simultaneous transport of water
and chloride in soil during one-dimensional infiltration and reported similar results for
chloride. They also noted that the apparent diffusion coefficients for chloride are less
than 0.01 crnlmin that is slow relative the rate of water movement. This means that once
a soluble non-reactive ion, like C1 -,begins to move with water, its back diffusion is
slow, which means that ion and water are closely coupled. Clothier and Sauer (1988)
and Li et al., (2003) reported the same results with nitrate NO-3.
The effect of inlet concentration on the distribution is significant. The higher the inlet
concentration, the higher the concentration of soil solution will be, and the same hold
true for low inlet solution concentrations. For example, the maxinium chloride
concentration values were found under the application rates of 1.5 and 4.5 l/h while the
minimum values of chloride concentrations were found under application rates of 2.0
and 2.7 l/h. The inlet concentrations of chloride for these application rates were 22.14
meqll bulk soil for 1.5 and 4.5 l/h and 5.33 meqll bulk soil for the application rates of
2.0 and 2.7 l/h. A similar result was also reported by Muluneh et al. (1995) and Li et al.
(2003). Muluneh et a1 . (1995) attributed this trend to the fact that in the active zone
close to the point of application having higher soil water content would mean greater
solute content. So in the case of higher concentration solution application, there will be
more solute concentration close to the point of application. On the other hand, the
concentration of chloride is decreased toward the wetting front and reached its minimum
values at the boundaries of the wetting front. This result is in agreement with Riss et al.
(1989) who found that the interface between previous soil water and current water
applied was more clearly defined with bromide. This finding would provide an accurate
method of locating the wetting front under high antecedent moisture condition.
Horizontal distance Icm)
50-60-
V,= 8 liters
q= 1.5 I/h
-70
1
9 liters
V,= 12 liters
q= 2.7 Ilh
q= 2.5 Ilh
80
q= 2.0 I/h
-7080
80
V,= 12 liters
-0
3
Figure 4.38: Chloride distribution (meq/l bulk soil) in sandy soil at different discharge
rates of 1.5,2.0,2.5 and 2.7 l/h
In order to limit the treatments to flow rate alone, the cumulative volume of applied
water was held constant. This procedure was adopted for the application rates of 2.0 and
2.5 (cumulative water applied 12 liters), and 4.5 and 6.0 llh (cumulative water applied
34 liters).The chloride concentrations were expressed in a dimensionless relative form

rather than miliequivalents per liter bulk soil.
10
20
30
40
50
Horizontal advance (cm)
'
10
20
30
40
50
10
20
30
40
50
-10-
-20111
-40-
Q)
-60-70-
V,= 29 liters q= 4 8 I/h
- 08 -
Figure 4.39: Chloride distribution (meqll bulk soil) in sandy soil at different discharge
rates of 3.5,4.5,4.8 and 6.0 l/h
In this manner, the comparisons between the treatments will be under the same basis.
The chloride relative concentration was calculated as:
CRC
co -
C,
cn
where CRC is the chloride relative concentration (dimensionless)

c= chloride concentration in soil (meqll bulk soil)
c,= initial chloride concentration in soil (meqll bulk soil)
c,= inlet chloride concentration (meqll bulk soil )
The samples used for statistical analysis were chosen to be within the horizontal
distances of 2.5, 12.5, 17.5 and 22.5cm and within the vertical depths of 2.5, 7.5, 12.5,
17.5, 22.5, 27.5, 32.5 and 37.5 cm. This region was selected to exclude the extreme
concentrations and to ensure the data points to be located within the wetted volume. The
average chloride relative concentration was found to be 0.422 and 0.45 for the
appIication rates of 2.0 and 2.6 llh, respectively. The statistical analysis using paired ttest showed non significant difference between both averages of chloride distribution (tratio= 1.194 compared to t ,,it
= 2.74,
n=32). This result agrees with the findings of Li et
al., 2003 who found the application rate had insignificant effects on the radial .and
vertical nitrate concentration within the range of 15 cm from the source.
In order to facilitate the comparison, the relative chloride concentrations were plotted
along the radial and vertical transects as shown in Figures A. 8 and A.9. Figure A.8
shows the distribution of the chloride relative concentrations for application rates of 2.0
and 2.5 l/h along the vertical axis (z) at two different horizontal distances from the
source: r= 2.5 and 17.5 cm, i.e. CRC (2.5, z) and CRC (7.5, z). This Figure shows the
distribution of chloride under application rate of 2.0 llh agree well with those under
application rates of 2.5 l h . The results of paired t- test revealed these treatments have no
significant difference at the 99% level for chloride distribution (t- ratio
compared to t ,,it
= 3.49,
n= 8 at r= 2.5 cm and t- ratio =2.58 compared to t ti,
2.076
= 3.49, n=
8 at r= 7.5 cm).
Figurer A.9 shows the horizontal distribution of chloride concentrations at two depths,
z= 7.5 and 27.5 cm below the soil surface. The results show that chloride concentration
under application rate of 2.5 l/h is higher at the horizontal distances close to the point
source. At the edge of wetting fronts this trend does not follow, where the lower
application rate results higher chloride concentration compared to the higher application
rate. The statistical analysis using paired t-test indicated no significant difference at the
99% level between the chloride concentrations as measured under the application rates
of 2.0 and 2.5 l/h (t- ratio = 1.75 compared to t
,,,,=4.60, n= 5 at z= 7.5 cm and t- ratio =
0.32 compared to t c,,~ =4.60, n= 5 at z= 27.5 cm).
The second treatments used to study the effect of application rate on the solute
distribution in sandy soil were selected to be under application rates of 4.5 and 6.0 l/h.
The cumulative water applied for both application rates was 34 liters and the inlet
chloride concentrations were 22.14 and 8.2 meq/l bulk soil for 4.5 and 6.0 l k application
rates, respectively. The average chloride relative concentration was found to be 0.64 and
0.56 for the application rates of 4.5 and 6.0 llh, respectively. The statistical analysis
using paired t- test showed significant difference at the 99% level between both
application rates on the chloride distribution (t-ratio= 4.38 compared to t
,,it
= 2.74,
n=
32). These results are in disagreement with the findings of Li et al, 2003 where the
application rates had insignificant effect on solute distribution.
The relative chloride concentrations were plotted along the radial and vertical transects
as shown in Figure A.lO. This Figure shows the chloride distribution in a dimensionless
relative form at two different horizontal distances from the source: r= 2.5 and 17.5 cm
i.e., CRC (2.5, z) and CRC (7.5, z). At r= 2.5 cm, the chloride distribution showed very
good agreement between both application rates and the statistical analysis showed no
significant effect of application rate at the 99% level on relative chloride concentration
where t- ratio = 3.19 compared to t c,,~ =3.49, n= 8. In case of vertical transects at r= 17.5
cm, the results of paired t-test showed significant difference between the chloride
distributions at the 99% level under both application rates and t- ratio
to t
6.45 compared
=3.49, n=8 at r= 17.5 cm). Figure A.11 illustrates the distribution of the relative
chloride concentration in soil along the radial transect at 7.5 and 27.5 cm from the soil
surface. An extremely high chloride concentration was found at the distance close to the
edge of wetting front r= 22.5 and z= 27.5 cm. However, it appears that the distributions
of the relative chloride concentration are similar under both application rates. The results
of paired t- test show no significant difference between the chloride distributions under
both application rates (t- ratio
ratio
= 0.877
compared to t
1.88 compared to t
,,,, =4.60, n= 5 at z= 7.5 cm and t-
,,,, =4.60, n= 5 at z = 27.5 cm). The discrepancies between
both treatments can be attributed to the effect of moisture content redistribution as

indicated previously. Some previous reports for one -dimensional salt and water (Biggar
and Nielsen, 1962; Wanick et al., 1971, Kirda, 1972, Alemi, 1980) indicated that, after
infiltration, salt is displaced deeper by the same amount of water added at a slower rate
than at faster rate, and concluded that the effect of hydrodynamic dispersion on mixing
was less at the slower rate.
The initial chloride content of the sandy loam soil was determined in the laboratory to be
3.0 meql 1. The chloride concentratibns of the irrigation water (inlet concentration) were
30,25,47 and 16 meqll for the application rates of 2.3,2.6,3.O and 4.0 llh, respectively.
The initial volumetric chloride concentration (c, 8), was found to be 0.1 8 meqlliter bulk
soil. The inlet volumetric chloride concentrations which are the maximum possible of
volumetric chloride concentration (c, 0,) at the soil inlet are 14.40, 12.0, 7.68 and 22.56
meqlliter bulk soil for the application rate treatments of 2.3, 2.6, 3.0 and 4.0 llh,
respectively. Figure 4.40 shows the distribution of the chIoride concentrations within the
wetted profile for the sandy loam under the application rates of 2.3, 2.6, 3.0 and 4.0 l/h,
respectively. The cumulative amount of water applied were 10, 17.3, 17.5 and 24 liters
for the application rates of 2.3,2.6, 3.0 and 4.0/h, respectively. The higher input solution
concentration, the higher the concentration of soil soil solution will be, and the same
holds true for low input solution concentrations. The concentration decreases as the
wetting front is approached and reaches its initial value near the wetting front boundary.
On the other hand, patterns of chloride distributions similarly follow the patterns of the
moisture content distribution. Similar results were also reported by Kirda et al. (1973),
CIothier and Sauer (I988), and Li et al. (2003). Like sandy soil, the effect of inlet
concentration on the distribution of chloride concentration in sandy loam soil is
significant, whereas a greater inlet concentration produced a higher chloride
concentration in the soil. Chloride distribution for each application rate shows the
vertical spread is more than horizontal. As mentioned earlier, this result is in accordance
with the findings of Muluneh et al. (1995) and Li et al. (2003)
The effect of application rate on the distribution of chloride concentration can be

detected by examining the treatments of application rates of 2.6 and 3.0 l/h. These
treatments were selected because they all applied approximately equal volumes of water
for different duration and they are typical of other treatments of equal volumes. The
cumulative water applied for both treatments were 17.3 and 17.5 liters for the
application rates of 2.6 and 3.0 I/h, respectively. To facilitate the comparisons between
both treatments the chloride distribution data were expressed in a dimensionless relative
form. The samples used for statistical analysis were chosen to be within the horizontal
distances of 2.5, 12.5, 17.5 and 22.5 cm and within the vertical depths of 2.5, 7.5, 12.5,
17.5, 22.5, 27.5, 32.5 and 37.5 cm. The average relative chloride concentration within
the selected region was found to be 0.52 and 0.51 for the treatments with application
rates of 2.6 and 3.0 l/h, respectively.
The results of paired t-test indicated that there was no significant difference between the
two averages at 99% significant level (t-ratio= 0.44 compared to t
2.74, n= 32).
This indicates that the application rates had insignificant effect of chloride distribution
which is in accordance with the findings of Li et al., (2003).
The transect plots as displayed in Figures 4.12 and A.13 permit a more objective
comparison. These Figures show the relative chloride concentration distribution under
the treatment of 2.6 l/h along with those under application rate of 3.0 l/h. Figure A.12
shows the relative chloride concentration distribution at vertical planes r= 2.5 and 17.5
cm away from the point source, respectively. At the vertical transect of 2.5 cm, the
relative chloride concentration is higher under the application rate of 2.6 I/h compared to
3.0 l/h along vertical distances.


0
10 15 20 25 30
35 40 45 50
Figure 4.40: Chloride distribution (meqll bulk soil) in sandy loam soil at discharge rates
of 2.3,2.6, 3.0 and 4.0 l/h
On the other hand, this tendency does not follow at vertical transect of 17.5 cm where
the relative chloride concentration is lower under application rate of 2.6 l/h at the
vertical distances closer to the edge of wetting front.
However, the statistical analysis showed no significant effect of the application rate on
the relative chloride concentration at 99% significant level between both application rate
at the vertical transect of 2.5 cm (t-ratio= 0.01 1 compared to t
,,it
=3.49, n=8). At the
vertical transect of 17.5 cm, the results of statistical analysis revealed there is no
significant effect of the application rate on the relative chloride concentration
distribution at 99% significant level (t-ratio= 0.63 compared to t ,,it =3.49, n=8).
Figure A. 13 shows the distribution of the relative chloride concentration in soil along the
radial transects at 7.5 and 27.5 cm from the soil surface, respectively. Although there are
some discrepancies between the relative chloride concentrations as measured from both
treatments, the statistical analysis using paired t- test showed that these discrepancies are
not significant at 99% significant level (t- ratio = 0.82 compared to t ,,it =4.60, n= 5 at z=
7.5 cm and t- ratio = 0.93 compared to t ,,it =4.60, n= 5 at z = 27.5 cm).These results are
in agreement with other findings discussed above where the concentration of solute
affected by the moisture content distribution within the wetted soil volume.
4.2
Predictive Model for wetting Pattern Dimensions
4.2.1 Surface wetted radius Equation

A nonlinear regression approach using SPSS statistical package was used to find out the
best-fit coefficients for Equation 3.14. The predictor variables were significantly
correlated with the dependent variable, and i.e. correlation is significant at 0.01 levels, as
shown in Table 4.6. The estimated coefficients of the surface wetted radius equation and
statistical tests are presented in Table 4.7. All the estimated coefficients have the
expected signs. The value of the coefficient of determination, R' shows that the variation
on the selected variables explained almost 83% of the variant of the dependent variable
in the model. Moreover, the diagnostic tests revealed that the discharge rate was found
to be insignificant while the other predictor variable passed all the tests successfully, i.e.,
t-value and probability level (P<0.01). In addition the values of standard error were
found to be very small. The adequacy of the surface wetted radius model was evaluated.
The results indicate that RMSE and MBE of the vertical movement of the wetting front
model are acceptable as their values obtained are relatively small 3.19 and 2.39 cm,
respectively. In addition, the value of U Theil's inequality coefficient of the model is
0.061, which reflects a good model performance. Furthermore, Figures A. 14 to A. 17 (in
the appendix) support the above finding that Eq. 4.8 has high predictive power and
ability to track the path of the actual observations.
Table 4.6: Nonlinear model summary for surface wetted radius

Model
Correlation coefficient
R square
Std.Error of regression
Table 4.7: Coefficients and diagnostic tests for nonlinear regression model for surface
wetted radius
Predictor variable
Coefficient
Std. Error
t-Statistic
Probability
The final mathematical form of the simulated surface wetted radius under point source
irrigation is defined as:
where consistent units were used in this approximation r [cm]; A0 [crn3/cm3],q , [mllh],
V [ml]; and k, in [crnlh].
4.2.2 Vertical Advances of Wetting Front Equation
The best-fit coefficients of vertical advances of wetting front equation were determined
by following the same technique used to estimate the surface wetted radius Equation
3.15. The estimated coefficients for the vertical advances of wetting front equation and
statistical tests are presented in Tables 4.8 and 4.9. The most interesting result of the
vertical advances of the wetting front equation is that all estimated coefficients have the
expected signs. This will be discussed, particularly in the section on sensitivity analysis.
Table 4.8.: Nonlinear model summary for vertical advances of wetting front
-
Model
Correlation coefficient
R square
Std-Errorof regression
Table 4.9: Coefficients and diagnostic tests for nonlinear regression model for vertical
Predictor variable
Coefficient
Std. Error
Probability
The diagnostic tests indicate that the effects of predictor variable are statistically
significant on the dependent variable, i.e. t-value at probability level (P<0.01). The
standard errors for all predictor variables were found to be very small. The magnitude of
the coefficient of determination, R ~shows
,
that the variation in the selected variables
explained almost 93 % of the variant of the dependent variable in the model.
The adequacy of the vertical movement of wetting front model was evaluated by root
mean square error (RMSE), mean bias error (MBE) and Theil's inequality coefficient (U)
criteria. The results indicate that RMSE and MBE of the vertical movement of the
wetting front model are acceptable as their values obtained are relatively small 2.91 cm
and 2.24 cm, respectively. In addition, the value of (U) Theil's inequality coefficient of
the model is 0.050, which reflects a good model performance. Furthermore, Figures
A.18 to A.21 support the above finding that Equation 4.9 has high predictive power and
ability to track the path of the actual observations.
According to the results mentioned above, the final mathematical formula for the
simulated vertical advance of the wetting front under point source irrigation follows as:
where consistent units were used in this approximations z [cm]; A 0 [cm3/cm3],q [mlh],
V [ml]; and k, in [crnlh].
4.2.3 Sensitivity Analysis

Sensitivity analysis is an important procedure to provide a better understanding of the
dominant modeling parameters and to reveal how the model behavior would be
influenced by the changing of its parameters. Parameter sensitivities were evaluated by
systematically varying each model parameter, while maintaining all other parameter
values at their original calibrated values. In this section, a sensitivity analysis was
conducted with respect to these parameters and examines the robustness of the abovementioned model.
Equations 4.8 and 4.9 were used to examine the sensitivity of the surface wetted radius
(r) and the vertical advances of wetting front
(2)
to the inputs of the models. The soil
type was represented by the average change of moisture content ( A @ and the saturated
hydraulic conductivity and the hydrologic parameters were represented by the discharge
rate (9,) and total volume applied of water (V,).To evaluate the sensitivity of the
surface wetted radius and the vertical advances of wetting front to the saturated
hydraulic conductivity, we assumed the average change of soil moisture content, the
emitter discharge and the total volume water applied to be constant. Equation 4.8 could
be rewritten to be in the form of Equation 4.10, *whichdescribes the relation between the
surface wetted radius and the saturated hydraulic conductivity. In Equation 4.10, the
surface wetted radius is proportional to the power of -0.034 of the saturated hydraulic
conductivity. The minus sign indicates that the surface wetted radius would be increased
as the saturated hydraulic conductivity decreased.
The response of the surface wetted radius to the change in the saturated hydraulic
conductivity is shown in Figure 4.41. The base data used to display this figure was
considered as a soil with 5.8 cm/h saturated hydraulic conductivity (k,), the average soil
moisture content AO= 0.22 irrigated by a single emitter with an application rate of 2 LA.
This base data were considered to be generalized for the other parts of sensitivity
analysis. The solid line represents the based data, while the opened and the closed
symbols represent the changing in the value of saturated hydraulic conductivity (1/4
5.8 crnlh and 4
5.8 cmh, respectively). An increase in the saturated hydraulic
conductivity k, from 5.8 cm/h to 23.2 cm/h, representing a change to a lighter soil would
result in an approximate decrease of 4.6 % in the surface wetted radius. In the case of
decreasing saturated hydraulic conductivity k, from 5.8cmlh to 1.23 cmlh, which

representing a change to a finer soil would result in an approximate increase of 4.9 % in
the surface wetted radius.
28
26 24 h
k= 1.40 cmlmin
- k= 5.80 cmlmin
k= 23.20 cmlmin
'
1
'
'
'
4
.
5
Elapsed time (hour)
Figure 4.41: Sensitivity analysis of the surface wetted radius to the change in saturated
hydraulic conductivity
The results of sensitivity analysis of the surface wetted radius due to the change of
saturated hydraulic conductivity verify the effect the soil texture on the behavior of
surface wetting movement. In addition, it leads to the statement : "Under the same
conditions of emitter discharge, soil moisture content and irrigation duration, the surface
wetted radius would be less in the light soil (higher saturated hydraulic conductivity)
than in heavy soil (lower saturated hydraulic conductivity)".
The sensitivity of the surface wetted radius and the depth of the wetted soil volume to
emitter discharge were considered. To find out the effect of changing discharge rate on
the surface wetted radius, the other factors must be constant. In this manner, the surface
wetted radius can be related to discharge rate as:
Equation 4.11 describes the response of surface wetted radius to the changing in emitter
discharge rate. It reveals that the surface wetted radius is proportional to the power of
-0.0028 of the discharge rate. Graphical result of sensitivity analysis of surface wetted
radius with respect to the change in discharge rate is shown in Figure 4.42. Again, the
same conditions were adopted to display the results graphically. Doubling the discharge
rate would cause an increase of 0.19% in surface wetted radius. The magnitude of the
change in the surface wetted radius due to the change in discharge rate is very small.
This is an interesting result, which agrees with results obtained by Roth (1974), who
found that the rate of application has little effect on surface wetted radius.
26
24 .
22 20 -
- q= I.O llh
q= 2.0 11h
------- q= 4.0 I/h 1
Elapsed time (hour)
Figure 4.42: Sensitivity analysis of the surface wetted radius to the change in emitter
discharge rate.
Sensitivity analysis was also performed for the average change of moisture content.
Following the same procedure adopted to analyse the effect of the saturated hydraulic
conductivity and emitter discharge rate, the relation that describes the effect of the
average change of moisture content on the surface wetted radius is written as:
In Equation 4.12, the surface wetted radius is proportional to the power of -0.56 of the
average change of moisture content. Again, the minus sign indicates that the surface
wetted radius would be increased as the average change of moisture content decreased.
Figure 4.43 shows the influence of the average change of moisture content on surface
wetted radius. This figure was constructed in the same manner, as discussed earlier. The
solid line represents the base data, while the symbols represent the change in the value
of average change of soil moisture (0.22 & 0.04).
2.0
2.5
3.0
Elapsed time (hour)
Figure 4.43: Sensitivity analysis of the surface wetted radius to the change in average
change of soil moisture content
An increase in the average change of moisture content d8 from 0.22 to 0.26,
representing a change to a heavy soil would result in decrease of 9 % in the surface
wetted radius. WhiIe a decrease in average moisture content AB from 0.22 to 0.18 would
induce in an approximate increase of 12 % in the surface wetted radius. The tendency of
the increase in surface wetted radius as a consequence of the decrease in average change
of moisture content was expected according to the hemispherical model, Ben Asher et al.
(1986), where the relationship between the wetted radius and the average change of
moisture content results from hemispherical model could be presented as:
oc A 0 -0.33
rhernis
4.13
As mentioned in the previous sections, the total applied water (V,) is an important factor
affecting the surface wetted radius. The sensitivity analysis of the model with respect to
the change in the total applied water was performed. It was found that the relationship
between the surface wetted radius and the total applied water follows:
As expected, the increase in water application would induce increase in surface wetted
radius. Figure 4.44 shows the influence of total water applied on the simulated curve of
the surface wetted radius. For instance, doubling the application water would result in a
20% increase in the surface wetted radius. On the other hand, decreasing of the
application water would result in decrease of surface wetted radius whereas decreasing
the applied water from 4 liters to 2 liters would induce a 17 % decrease in the surface
water radius.
Cumulative infiltration water (ml)
Figure 4.44: Sensitivity analysis of the surface wetted radius to the change in cumulative
infiltrated water
Sensitivity analysis was also performed to assess the effect of saturated hydraulic
conductivity k,, emitter discharge rate q,, and applied volume of water V , and total
change of soil moisture content A0 on the depth of wetting zone as predicted by
Equation 4.9.
The sensitivity of the vertical advances of the wetting front to the saturated hydraulic
conductivity could be described by Equation 4.1 5
According to Equation 4.15, higher value of saturated hydraulic conductivity would lead
the wetting front to be deeper and vice versa. For more clarification, Figure 4.45 shows
that the depth of wetting front is increased by an approximate of 31 % as the saturated
hydraulic conductivity increased from 5.8 cmlh to 23.2 crnlh. On the contrary, the depth
of wetting zone is decreased by approximately 24 % as the saturated hydraulic

conductivity decreased from 5.8 cmlh to 1.45 crnlh. However, the increase of vertical
advances of the wetting front by increase in the saturated hydraulic conductivity may
interpret the elongated of vertical movement in the light soil compared to heavy soil.
The sensitivity analysis of the model with respect to the change in emitter discharge rate
was performed. The effect of emitter discharge rate on the vertical advances of wetting
front as formulated by Equation 4.16 is in inverse proportion. This finding is similar to
the results reported by Bresler et al., (1971) and Schartzman et al., (1985)
Elapsed time (hour)
Figure 4.45: Sensitivity analysis of the vertical advance of wetting front to the change in
saturated hydraulic conductivity.
Unlike the surface wetted radius, the vertical advance of wetting front is more affected
by emitter discharge rate, Figure 4.46 shows the sensitivity of the vertical advances of
wetting front to the change in emitter discharge rate. As can be seen from Figure 4.46,
doubling emitter discharge rate would result in an approximate 6.75 % decrease in the
wetted depth. In contrast, increasing of the vertical advances of the wetting front by 7.25
% would be expected as the emitter discharge rate decreases from 4 l/h to 2 l/h. The
results of the present analysis confirm our belief that the depth of wetting zone can be
controlled by the proper selection of emitter discharge.
0.0
0.5
1.O
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Elapsed time (hour)
emitter discharge rate
The influence of the total average change of soil moisture content A 8 on the vertical
movement of wetting pattern was investigated. According to the Equation (4.9), the
relationship between the total average change of soil moisture content A 8 and the
vertical movement of wetting pattern could be described as:
Examining Equation 4.17, it can be noted that the relation between the total average
change of soil moisture content A 0 and the vertical movement of wetting pattern is
inversely proportional. The sensitivity analysis results as showed in Figure 4.47 showed
a 7 % increase in the vertical advances of wetting front as of the total average change of
soil moisture content A 0 decreased by 18%. Contrary to the above, the increasing in the
total average change of soil moisture content A 0 of 18 % would induce decrease in 6 %
vertical advances of the wetting front. According to ~ ~ u a t i 4.17,
o n the value of the total
average change of soil moisture content A 0 is increased as the value of the total porosity
decreased. This is acceptable reason to explain the increasing of vertical movement of
wetting pattern in light soil compared to the heavy soil. Ben Asher et al. (1986) noted
the same effect of the total average change of soil moisture content on the depth of
wetted soil volume.
The sensitivity analysis was made to check the effect of the total applied water V , on of
the vertical advances of wetting front as estimated from Equation 4.9. Based on
Equation 4.9 the relationship between the depths of wetting front and the total applied
water V,,, may be written as:
Elapsed time (hour)
average change of soil moisture content
Examining Equation 4.18 leads one to note that the vertical advances of wetting front is
proportional to the total water volume applied with a power value of about 0.365. Figure
4.48 shows that doubling total water volume applied V,, would induce in an approximate
29 % increase in the vertical advances of wetting front. On the other hand, decreasing by
half the total water volume applied V, , would decline the vertical advances of wetting
front in approximate of 22%.
From the above, it is apparent that the behavior of the outputs of predicted model,
(surface wetted radius and vertical advances of wetting front) as affected by their input
parameters, (saturated hydraulic conductivity, total average change of moisture content,
emitter discharge rate and the total volume water applied) were we11 reasonable.
Cumulative applied water (liter)
cumulative infiltration water
This was also confirmed by other results Shwartzman et al. (1986), Bresler (1978) and
Ben Asher et al. (1986). In addition, the results of sensitivity analysis could be used as
good criteria for designing purpose. For instance, to increase the surface wetted radius
by lo%, total volume water applied, V, must be increased by 42%. Such an increase in
total volume of water applied, V, would normally result in 14 % increase in the wetted
depth, z.
4.2.4 Model's Verification
One objective of this study was to evaluate a simple procedure based on easily
obtainable inputs for predicting the components of the wetting pattern geometry under
point source trickle irrigation. In the previous sections we found out interesting
Equations 4.8 and 4.9 based on statistical evaluation. To verify the reliability of these
equations, data other than those used in the identification and estimation stage of the
analysis are needed. The results of our study were used as database for the purpose of
the verification process, i.e. the data of the wooden box and field experiments. Verifying
a model basically involves comparing model predictions with observations and the
common approach is to plot predicted and observed data on a graph for visual
impression. A more quantitative approach involves statistical analysis to compare
predicted and observed values. The results of verifying Equations 4.8 and 4.9 are
discussed in the following sections.
4.2.4.1 Wooden Box Experiments

The visually recorded surface wetted radii during water infiltration for sand soil under
laboratory condition were used to test the proposed relationships i.e., Equation 4.8. The
inputs used to calculate surface wetted radius are discharge rates (3, 5.5 and 7 lh),
saturated hydraulic conductivity (30 cm/h) and total change of water content (0.21). The
irrigation time was 380, 300 and 180 minutes for 3, 5.5 and 7 l/h application rates,
respectively. The results of comparison between observed surface wetted radius and
those predicted by the equation for all trials are plotted against a 1:1 line as shown in
Figure 4.49. In this figure, for a perfect model prediction all of the points would fall on
the line across the graph signifying a slope of one (Observed= Predicted). It is clear that
the observed-predicted data pairs lie very close to the 1:l line, which represents a
perfect agreement.
The values of the coefficient of determination, R ~ root

,
mean square error (RMSE),
mean bias error (MBE) and U Theil's inequality coefficient criteria (0.92,2.01 cm, 0.93
cm; and 0.0448, respectively) reveal that the model obtained best goodness of fit and
that the variation on the selected variables explained almost 92% of the variant of the
dependent variable.
o Observed-Predicted data
Fitted line y = 1.044 x I?= 0.95
-------
10
15
20
25
30
35
Observed surface wetted radius (cm)
Figure 4.49: Correlation of observed and predicted surface wetted radius for various
application rates for river sand
Furthermore, a comparison between the resultant slope of the best-fit line with
interception equal zero with the scattered of observed and simulated data on the
regression line shows a good simulation capability of the model for its purpose.
Figuer 4.50 shows the comparison between the mesured wetted surfcae radius
propagation with model predictions for infiltration of water under 3, 5.5 and 7 l/h
application rate. Model predictions were based on Equation 4.8 with k, = 30 cmlmin. For
each trial, the measured data, represented by open symbol, compared extremely well
with the predictions. This could be noted more clearly by examining the values of
performance indicators. The root mean square error (RMSE) values were found to be
2.24, 2.04 and 0.92 cm with respect to the trickle irrigation rates of 3, 5.5 and 7 llh,
respectively. The mean bias errors (MBE) between predicted and measured data were
calculated and values of 1.77, 0.27 and -0.42 cm were obtained. The U Theil's
coefficients obtained were 0.057, 0.058 and 0.021 for 3, 5.5 and 7 l/h application rate,
respectively.
For more clarifications, Figure A. 22 shows the correlation between measured and
predicted surface wetted radius for each application rate. The correlation coefficient (R),
which indicates the strength of the relationship between two variables, was found to be
98 %, 97 % and 99 % with respect to the trickle irrigation rates of 3, 5.5 and 7 lfh,
respectively.
The performance of the model for predicting the surface wetted radius can be also
achieved by matching 1:1 line (interception equal zero) with the best regression line.
The intercept values of the regression line were found to be 1.1, 1.02 and 0.97 for 3, 5.5
and 7 l/h application rate, respectively. This indicates very good matching between two
lines as they are very close to one. The discrepancies between the model results and the
observed data, which are very small, may have been caused by the variability of bulk
density as a result of packing the soil in the container.
Observed
Predicted
Elapsed time (min)
rn
.3
25 -
20 -
U
U
a,
P
a,
0
0
00
15-80
0
10
i0
Observed
Predicted
50
in
O . . . . - ' . - - . '
0
50
100
'
150
200
250
300
350
Elapsed time (min)
'in5
Observed
Predicted
f
0
20
40
60
80
100 120 140 160 180 200
Elapsed time(min)
Figure
Observed and predicted surface wetted radius for river sand soil under
5.5 and 7 l k application rates
As a conclusion, the results of the laboratory comparisons between the simulations and
laboratory results were generally well predicted by the suggested models. In addition,
they verify that the movement of wetting front is dependent upon the selected inputs
which mean the predicted model can successfully describe the surface wetted radius of
the soil wetting zone under point source trickle irrigation.
4.2.4.2 Field Experiments
The field experimental results were used to verify the ability of the suggested models to
predict the surface wetted radius and the vertical advances of the wetting front. If the
results of comparisons between the observed and predicted data indicated high
coincidence, it could then be reliably recommended in practice. The following sections
present the results of comparisons between the observed and predicted wetting pattern
components of sandy and sandy loam soil under field condition.
4.2.4.2.1
Sandy Soil
The soil properties, which were used as inputs to simulate the surface wetted radius and
the vertical advances of wetting front, were 0.20 cm3/ cm3 and 20.8 cm/h for the average
change of moisture content and saturated hydraulic conductivity, respectively. The
discharge rates ranged from 1.5 to 6 l/h while the duration of irrigation ranged between
four to six hours.
Surface wetted radius

The surface wetted radius of the sandy soil calculated based on the model described
earlier was tested against the measured data. The comparison between the measured
surface wetted radius for all trials combined and those predicted from the suggested
model, Equation 4.8, are shown in Figure 4.5 1. The performance indicators of the model
show a satisfactory agreement between predicted and measured data, where the values of
RMSE, MBE and Theil's coefficient were found to be 2.20 cm, 1.67 cm and 0.052,
respectively. In addition, the value of correlation coefficient R= 0.98 and the slope of
regression line 1.079, which is close to 1, indicating that the predicted model performed
quite well.
0 Observed- Predicted data
Fitted line y= 1.079 x F?= 0.97
10
15
20
25
30
35
40
Figure 4.51: Correlation of observed and predicted surface wetted radius for sandy soil
(all application rates)
Observed surface wetted radius versus those predicted by the model for selected
application rates are plotted in Figures 4.52,4.53 and 4.54. For most selected application
rates, it is clear that the predictions of the surface wetted radius fitted quite well with the
measured values and being able to track the path of the actual observations.
35:--
30 r
q= 1.5 Vh
25 -
l
l
20 :
10
-,o
* 8
15
ex
o
l
50
100
'
150
200
250
Observed
Predicted
'
300
350
400
'
'
450
1
Elapsed time (min)
q=2.10 Ilh
0
l
Observed
Predicted
Elapsed time (min)
Predicted
0
50
100
150
200
250
300
350
400
450
Elapsed time (min)
Figure 4.52: Observed and predicted surface wetted radius for sandy soil under 1.5,2.10
and 2.6 1/h application rates
Elapsed time (min)
1.
'
- "
'
"
: q=4.0 Ilh
0
0
,8
e
Observed
Predcted
Elapsed time (min)
Observed
predicted
Elapsed time (min)
Figure 4.53: Observed and predicted surface wetted radius for sandy soil under 3.5,4.0
and 4.5 l/h application rates
0
0
0
0
'
q= 6.0 Ilh
'8
0
50
100
150
200
250
Observed
Predicted
300
350
400
450
Elapsed time (min)
Figure 4.54: Observed and predicted surface wetted radius for sandy soil under 6.0 l/h
application rate
Values of statistical parameters including root mean square error (RMSE), mean bias
error (MBE) and Theil's inequality coefficient (U) are summarized in Table 4.1O.This
table shows that the best value of RMSE (0.88cm) was found at 2.6 l/h, while the worst
value (3.53 cm) was found at 6 llh. On the other hand, the best value of MBE (0.23) cm
was given by the application rate of 4.0 l/h. The worst value of MBE (3.31 cm) was at
application rate of 6 1/h. U Theil's inequality coefficient, which is used to test how well
the model prediction is, compared with the observed data, was found to be 0.02, which
indicates the lowest value, with application rate of 2.6 llh. The highest value of Theil's
inequality coefficient (0.085) was found for application rate 6 l/h.
Referring to both the Figures 4.52 through 4.54 and Table 4.10, it can be noted that the
identical fit between the observed and predicted surface wetted radius is more notable
especially at the beginning of irrigation while the gaps between the observed and
predicted data widened in some application rates as the irrigation duration increased.
The discrepancies between the predicted and observed surface wetted radius were
relatively very small for application rates of 2.10, 2.60 and 4.0 l k . This could be
extracted from the values of the performance indicators as displayed in Table 4.10. The
values of the mean bias error were found to be 0.78, 0.53 and 1.03 cm for application
rates of 2.10, 2.60 and 4.0 l/h, respectively. In addition, the values of U Theil's
inequality coefficient were 0.027, 0.024 and 0.023 for application rates of 2.10, 2.60 and
4.0 l/h, respectively. On the other hand, the discrepancies were found to be relatively
high for application rates of 3.5, 4.5 and 6.0 l/h, compared to the previous one.
According to the performance indicators, these discrepancies would be accepted as long
as the performance indicators are reasonable, as indicated from Table 4.10. The values
of the mean bias error were found to be 2.10, 2.59 and 3.31 cm for 3.5, 4.5 and 6.0 l/h
application rates, respectively. The values of U Theil's inequality coefficient for these
application rates were found to be 0.045,0.065 and 0.085 for application rates of 3.5,4.5
and 6.0 llh, respectively. The results of linear regression analysis showed that the model
verified well with the observed data in Table 4.10. The goodness of fit with
determination coefficients which ranged from 0.99 to 0.95 supported these results. The
slope of the best-fit line with interception equal to zero fluctuated from 1.02 to 1.15,
indicating slight overestimation of the surface wetted radius. However, the differences
between the predicted and observed surface wetted radius can be accepted since the
experiments were conducted under field conditions.
Table 4.10: Values of the statistical parameters used in comparison for surface wetted
radius for sandy soil (field experiment)
Vertical Advance of Wetting Front
The suggested model, Equation 4.9, which describes the vertical advances of wetting
front z was tested against the observed data obtained from the field experiments for
sandy soil. The comparisons between the observed and predicted vertical advances of
wetting under all application rates are shown in Figure 4.55. The results showed a
satisfactory agreement between the predicted and observed data with determination
coefficient, R~ = 0.92. This is also notable from the value of the regression line slope
0.98, which is close to 1 (i.e. slope of 1.1 line).
Observed-Predicted data
Fitted line y= 0.97 R *= 0.919
Observed Depth of wetting ront (cm)
Figure 4.55: Correlation of observed and predicted vertical advance of wetting front for
sandy soil (all application rates)
The performance indicators of such comparison assure the previous result which reveals
that the predicted model has the power to simulate the vertical advance of wetting front,
where their values were found to be 5.40 cm, 0.27 cm and 0.063 for RMSE, MBE and
Theil's inequality coefficient (U), respectively.
Graphical comparisons between the actual and the predicted vertical advance of wetting
front for five application rates 2.0,4.5,4.8 and 6.0 l/h are shown in Figure 4.56. There is
excellent agreement between the observed and the predicted data; this can be noted
under 2, 4.5 and 6 l/h application rates. The performance indicators, as shown in Table
4.1 1, reveal that the experimental and predicted data of vertical advances of wetting
front are in general agreement.
100
200
300
400
Elapsed time (min)
500
100
200
300
400
500
Elapsed time (min)
Observed
Predicted
Elapsed time (min)
Elapsed time (rnin)
Figure 4.56: Observed and predicted vertical advance of wetting front for sandy soil
under various application rate
For the selected application rate the values of RMSE, MBE and U Theil's inequality
coefficient are in the range of 2.65 to 7.84 for RMSE, -0.17 to 3.58 for MBE and 0.027
to 0.08 for U Theil's inequality coefficient. The best (lower) values of RMSE and U
theil's were found under 4.5 l/h application rate, while the best value of MBE was found
under 2.0 I/h application rate. On the other hand, the worst (largest) values of RMSE,
MBE and Theil's inequality coefficient were found under application rate of 4.8 l/h.
Table 4.11 shows the slope values of the best fitted line and the determination
coefficients of the regression equation for the observed and predicted data. As can'be
observed from this table, these values are closer to their optimum value, i.e. 1 for both
slope values and determination coefficients which indicated that the suggested model
gives good simulation of vertical advance of wetting front. Many reasons can be
attributed for such discrepancies such as soil variability and model error. However, these
results encourage using the suggested model in practice as it has been verified versus
real data obtained under field condition
Table 4.1 1: Values of the statistical parameters used in comparison for vertical advance
of wetting front for sandy soil (field experiments)
Flow rate ( l h )
RMSE (cm)
MBE (cm)
R*
Sandy Loam Soil

In the previous sections, the aptitude of the suggested models show high reliability to
simulate the surface wetted radius and the vertical advances of the wetting front for
sandy soil under laboratory and field conditions. For more verification, the suggested
models were used to simulate both the surface wetted radius and the vertical advances of
wetting front with sandy loam soil under field condition. The saturated hydraulic
conductivity and average change of moisture content, which represent the soil
properties, were 5.8 crnh and 0.24, respectively. The discharge rates ranged from 1 to
4.2 I/h while the duration of irrigation ranged from 210 to 420 minutes.
The predicted and observed values of the surface wetted radius for all application rates
combined are plotted in Figure 4.57. Consequently, in this graph, the perfect agreement
between the predicted and observed data would be located on a 45 O line (1 :1 line). The
predicted surface wetted radius agrees well with the field results where the data points
are seen closer to the equity line. The correlation coefficient (R) and the slope of equity
line to such relation were 0.94 and 0.89, respectively indicating that the predicted model
can be useful for track the surface wetted radius. This result is assured also by
examining the performance indicators of the model. The low values of U Theil's
inequality coefficient 0.054, the root mean square error and the mean bias error between
the observed and simulated data, 2.1 1 and -1.77 cm, respectively demonstrate that the
suggested model has a high predictive ability to simulate surface wetted radius.
10
15
20
25
30
35
Figure 4.57: Correlation of observed and predicted surface wetted radius for sandy loam
(all application rates)
Comparisons were made between the predicted and observed surface wetted radius for
each trial, separately. Figures 4.58 shows the observed and proposed data plotted versus
elapsed time. It was observed from these figures that the predicted model is highly
adequate to simulate the surface wetted radius. This is more clearly shown in cases of
application rates of 1.5, 2.0 and 3.5 l/h. On the other hand, there are some discrepancies
between the observed and predicted data can be seen under application rate of 2.5 Ih.
Beside the visual graphical comparisons, the accurateness of the fit between the
predicted and observed data can also be identified by the performance indicators which
include root mean square error RMSE, mean bias error MBE and Theil's inequality
coefficient U.
35
30 .
q= 2.0 Ilh
q= 1 50 11h
25
0
o
B e ' '
0
0
.
20 .
8
o
o0
15
Observed
Predicted
10
D*
- 0
i'
Observed
o
* Predicted
5 .
.-'
0
0
50
100 150 200 250 300 350
Elapsed time (min)
Elapsed time (min)
q= 3.5
Nh
0
'
0
8
0
00
100 150 200 250 300 350
50
Observed
Predicted
100 150 200 250
Elapsed time (min)
Elapsed time (min)
35
30 25 2015
-oO
lo
q=4.0 Ilh
0
0
O
0
Observed
Predicted
5 0
i.
0
50
100
150
200
250
300
350
Elapsed time (min)
Figure 4.58: Observed and predicted surface wetted radius for sandy loam soil under 1.5,
2.0, 2.5, 3.5 and 4.0 l/h application rates.
These indicators as summarized in Table 4.12 indicate that there is a good accuracy of
the model simulation to the observed data. The lowest values of the root mean square
error was 1S O cm, U Theil's inequality coefficient of 0.045 as obtained under 1.5 l/h
application rate. On the other hand, the highest value of these indicators were found
under 2.5 l/h application rate, 2.63, -2.42 cm and 0.067 for RMSE, MBE and Theil's
inequality coefficient, respectively.
In general, the difference between the predicted and observed data of the surface wetted
radius was small although the slope of the best fit line with interception equal zero
exhibits slightly underestimated may be due to model or experimental error. Table 4.12
shows that the discrepancies between the predicted and observed data fluctuated from
- 5% to -12 %. The proportion of the total variance of the observed data explained by the
predicted data was found close to one with coefficient of determination R2 equal to 0.96,
0.97, 0.97, 0.98 and 0.98 for application rates of 1.5, 2.0, 2.5, 3.5 and 4.0, respectively,
signifying model performance to be fairly good.
radius for sandy loam soil
Flow rate (llh)
RMSE (cm)
MBE (cm)
R2
Vertical Advance of Wetting Front
Comparisons were made between the observed and predicted data of vertical wetting
front advance. Figure 4.59 is a plot of the total data of vertical wetting fiont advance as
predicted by the suggested model versus the measured values. The data points are seen
close to the equity line with slope of 0.85 and 0.97 determination coefficients. It is
evident that the discrepancy between the measured and predicted data as shown in
Figure 4.108 may due to the natural variability of the soil. However, the performance
indicators lead to accept the applicability of the model for predicting the vertical wetting
advances where the values of RMSE, MBE and Theil's inequality coefficient were
found to be 5.0cm7 -4.07 cm and 0.082, respectively.
Comparisons between observed and predicted vertical wetting front advances under
2.30, 2.60, 3.00, and 4.00 l/h application rates are shown in Figure 4.60. It is apparent
from these figures that the predictions show the same patterns, like observed data.
Higher agreements between the predicted and observed data can be noted in the case of
application rates of 3.0 and 4.0 l/h. The deviations from measured values are slightly
higher at the later stage than at the beginning of irrigation.
The discrepancies between observed and predicted data could be attributed to the natural
variability of the soil and model error. However, the performance indicator values, as
summarized in Table 4.13, indicate that there is a good fit of the model simulation to the
observed data.
Observed- Predicted data

fitted line y =0.85 x R '=0.97
- 1:l line
Observed depth of wetting front (cm)
Figure 4.59: Correlation of observed and predicted vertical advance of wetting front for
sandy loam (all application rates)
The RMSE, MBE and U Theil's inequality coefficient values ranged from 3.53 to 6.24
cm, -2.82 to -5.72 cm and 0.057 to 0.099, respectively meaning that the suggested model
can be used to simulate the vertical wetting front advances.
A simple linear regression for each treatment with interception equal zero was fitted to
the predicted and observed data of depth of wetting front pairs, and, the slope A , and
the determination coefficient R~ were obtained. If the predictions are in perfect
agreement with the observed values, the slope should be 1, and R~ should be 1. The
closer the slope is to 1, and R~ is to 1, the better is the general predictive power of the
model. A slope of less than 1 indicates that the model under predicts the observation.
0
0
q=2.6 Ilh
q=2.3 I/h
0
0
0
0
0
0
0
O
.
D
D
([I
100
20/o:
Observed
Predicted
200
300
100
Observed
Predicted
200
300
Elapsed time (min)
Elapsed time (min)
-5
q= 4.0 I/h
40
\C
L
t:
0
0 .
Q)
a,
0
100
200
300
Elapsed time (min)
0
0
([I
m
'2
8
0
20
Observed
Predicted
400
30
Y-
s
m
>
Cn
.-s
400
I0
100
200
Observed
Predicted
300
400
Elapsed time (min)
Figure 4.60: Observed and predicted vertical advance of wetting front for sandy loam
soil under various application rate
Table 4.13 shows the slope values and the determination coefficient of the fitted
equation for each treatment. As can be seen the values of slope fluctuated from 0.82 to
0.89 indicating the percent of difference varied from -18% to- 11%. The determination
coefficient R~ of the predicted and observed for linear regression are 0.99, 0.99,0.95 and
0.99 for 2.3, 2.6, 3.0 and 4.0 l/h application rates, respectively. In general, these results
indicate a reasonable agreement between the observed and the predicted model and
suggested that the model performed well considering the discrepancies may be resulting
from model and experimental error.
Table 4.13: Values of the statistical parameters used in comparison for vertical advance
Flow rate (lh)
RMSE (cm)
MBE (cm)
R~
2.3
4.96
-4.46
0.088
0.84
0.99
The quantitative discrepancies that have been observed in some cases can be the result
of one or several of the following:
1 Inadequacy of the adopted assumptions as it simplifies a very complex process.

2 Inability to create uniform initial conditions in the field.
3 Lack of precision in estimating the soil water parameters i.e., saturated hydraulic
conductivity and saturated moisture content.
4 The difference of the atmospheric conditions could account for some of these
discrepancies where the predicted equations were developed based on
experiments were conducted under laboratory conditions.
5
The natural variability of the soil also could account for some of the observed
differences.
4.2.4.3 Verification with Published Data
To test the models more thoroughly, a series of experiments selected from published
data may support the proposed models for determining the wetting pattern (i.e., surface
wetted radius and vertical advance of wetting front) under point source trickle irrigation.
The selected experiments were conducted by Colthier and Scotter (1982), Risse et al.,
(1989), Yitayew et al., (1998), Palomo et al., (2002) and Li et al., (2004). Two
experiments were conducted under laboratory condition i.e., Colthier and Scotter (1983)
and Li et al., (2004), while the others were conducted under field conditions. The
differences between these experiments are more related to the input data, i.e., saturated
moisture content; saturated hydraulic conductivity, the discharge rate and the frequency
of irrigation. These differences create advantages which could help for verification
purpose. The details of these experiments were discussed extensively in the literature.
However, the salient experimental details are repeated here.
Constant Flux from a Hemispherical Cavity
Colthier and Scotter (1982) presented experimental results of 3-D infiltration

experiment. Manawatu fine sandy loam was packed in' an acrylic box 20 by 20 cm in
cross section and 30 cm in depth. Taking advantage of radial symmetry of water flow,
water was applied in one corner of the box with a discharge rate of 90 mllh. The cavity
formed had a radius of 4mm which remained filled with water during the experiment.
On the adjoining box faces the position of the wetting front was drawn at various times
during each experiment. The saturated hydraulic conductivity and saturated moisture
content of this soil were 7.2 cm/h and 0.37, respectively. The value of saturated moisture
content was determined by Clothier and Scotter (1982) while the value of saturated
hydraulic conductivity was reported by Clothier et al., (1985). Although the experiment
was conducted under laboratory conditions, its findings may be used to verify using the
suggested models for practice purpose. However, the comparisons between the observed
and predicted data provide a consistency test for the suggested models.
The comparison between the observed surface wetted radius data and those predicted
from the proposed model is shown in Figure 4.61. As can be seen from the figure, there
is high accordance between observed and predicted data of surface wetted radius
whereas the observed data was closely matched by the predicted. The reliability of the
model predictions can be estimated also from the Figure A.23-A where the predicted
data was plotted versus the observed. The figure shows also the equity line and best
fitted line for the observed and predicted surface wetting radius. The points are closely
lie on and around the 1:l line with slope value found to be 1.002. The determination
coefficient R~ of the predicted and observed for the linear regression is 0.996 indicating
high precision of the predicted model to simulate surface wetted radius under described
conditions. The calculated performance indicators (i.e, MBE, RMSE and U Theil's
coefficient) support the findings discussed above, where their values are close to their
optimum, 0.1 1 cm, 0.28 cm, and 0.015 for MBE, RMSE and U Theil's coefficient,
respectively.
Elapsed time (min)
Figure 4.61 : Observed and predicted surface wetted radius for fine sandy loam soil
The vertical advance of wetting front was also examined to assess the predicted model,
Equation 4.9. The results of the model evaluation are presented in Figure 4.62. The
computed vertical advance of the wetting front agrees well with the observed data
especially at initial times. However, at later times the observed show a faster movement
vertical advance of the wetting front compared to the predicted data. Discrepancies can
be due to uncertainty in assumed hydraulic parameters and to soil uniformity. The
predicted data of vertical advance of wetting front as obtained from the model versus
observed data is presented in scattergram as shown in Figure A. 23-B. Graphical
depiction shows that the points are uniformly scattered around the line 1:l line with
slope found to be 0.87. The coefficient of determination R* of the linear regression was
found to be 0.987. Since the value of R2 could be interpreted as the proportion of
variation in observed value that is explained by the fitted model, this value indicates that
the model prediction in vertical advance of wetting front was reasonably good. The
performance indicators, i.e., MBE, RMSE and U Theil's coefficient for simulated and
observed results were calculated. The values of these parameters were found relatively
close to their optimum values where -1.09 cm, 1.89cm and 0.07 were found for MBE,
RMSE and U Theil's coefficient, respectively.
The findings of the maximum wet in the horizontal and vertical directions at the selected
time which was reported by Clothier and Scotter (1982) enhancing our trust in the
results of the suggested model we created. This coincidence implies that the suggested
model is successfid and can be a useful for predicting the lateral and vertical wetting
front throughout the soil profile under surface point source trickle irrigation.
[7
P
I
Observed
Predicted
rn
1
I
t - - .
'
100
200
300
400
500
600
700
Elapsed time (min)
Figure 4.62: Observed and predicted vertical advance of wetting front for fine sandy
loam soil
Laboratory Study of Li et al., 2004
Laboratory experiments were undertaken to find empirical model that predicts wetting
patterns from a surface point source for sandy soil (94.8 % sand, 2.4 silt% and 2.8%
clay). The completely mixed air dried soil was passed through a 2 rnrn sieved and
packed in a 15' wedge -shaped plexiglass container to an approximate bulk density of
1.46 gm/cm3. The average stable'infiltration rate determined by double ring method was
1.8 -2.4 crnh. The field capacity at pressure head of -100 cm was 0.22 cm3/cm3and the
saturated water content was 0.42 cm3/cm3. Five experiments were conducted with
application rates of 0.5, 0.7, 1.0, 1.40 and 2.0 l/h. During each experiment, the positions
of the moving wetting front on the soil surface and in the vertical plane were recorded
visually at several times. The data collected from these experiments were used to check
the accuracy and consistency of Equations 4.8 and 4.9. Figure 4.63 shows the observed
and predicted surface wetted radius as a hnction of elapsed time for different
application rates, 0.5, 0.7, 1.0, 1.4 and 2.011h. In this figure the open circles represent the
observed locations of surface wetted radius, while the solid circles represent the
predicted. As can be seen, the experimental and the predicted surface wetted radius
agree in general. There are trends of slight over-estimation of the surface wetted radius
with the predicted model simulation, as shown in the figure. The main reason for such
discrepancies may be due to the variability of bulk density as a result of packing the soil
in the container. Both the input data, saturated hydraulic conductivity ks and saturated
moisture content $ are strongly related to the bulk density (Hill 1990). The differences
between observed and predicted data are further displayed in Figure A.24. In this figure,
for perfect model prediction all of the points would fall on the line across the graph
signifying a slope of one (observed =predicted). For all treatments, it was found that all
predicted data lie relatively close to the equity line. In addition, the best fitted line for
each treatment was so close to the 1:1 line. This can be revealed from the slope values of
the best fit equation where these values ranged from 1 .O1 to 1.07. The determination
coefficients, R~ for the linear regression fluctuated from 0.967 to 0.998. Since the value
of R~ could be interpreted as the proportion of the variation in observed value that is
explained by the fitted equation, the obtained values of
indicated that the suggested
model performed well. Additional statistical tests were performed to validate and
investigate the capability of the model in describing the surface wetted radius. Table
4.14 shows the performance indicators used to measure the accuracy of the suggested
model i.e. mean bias error (MBE), root mean square error (RMSE) and U Theil's
coefficient.
The MBE, RMSE and U Theil's coefficient statistics have as a lower limit the value of
zero. As can be indicted from Table 4.14 these statistics are generally small and close to
their optimum values. The maximum values of these indicators are 1.16cm71.21cm and
0.010 for MBE, RMSE and U Theil's coefficient, respectively. Even though these values
represent the worst findings, they are still under their optimum limit.
The observed data of vertical advances of wetting front were used to evaluate the
predictive ability of Equation 4.9. The results of the model evaluation for five
application rates of 0.5, 0.7, 1.0, 1.4, and 2.0 llh, respectively, were plotted versus
elapsed time as presented in Figure 4.64. As can be noted, the predicted model tracks the
observed points and the path of the actual observations very closely. At larger times the
simulated data are slightly lower than observed ones. This discrepancy can be due, as
mentioned earlier, to the soil uniformity. For meaningful comparisons, the predicted
results were plotted versus the observed in Figure A.25. Both predicted data and best
fitted line located so close to the identity line (where predictions equal observations).The
slope values of the best fit line (ranged between 0.95 and
0.988) indicated higher
capability of the suggested model to simulate the vertical advances of wetting front. The
determination coefficients R~ were found to be in range between "0.94 to 0.96", i.e., the
fitted equation explained from 94 % to 96 % of the variability in observed data. The
other statistical criteria, i.e, MBE, RMSE and Theil's inequality coefficients were also
used to test how the results predicted by the model mimic the observed results, Table
4.14. The maximum calculated values of MBE, RMSE and Theil's coefficient were 0.43
cm, 1.77 cm and 0.055, respectively. The calculated coefficients values are close to their
optimum values.
As a conclusion, the results of the surface wetted radius and vertical advances of wetting
front which extracted from laboratory experiments of Li et al., 2004 almost coincided
with the predicted results. This coincidence implies that the suggested models are
successful and can be used as useful tools for predicting the surface wetted radius and
the depth of wetting front through the soil profile under surface point source trickle
irrigation.
- .
30
q= 0.5 Ilh
25
20
- - - . -
. .
q= 0 7 1Ih
s
8
0
15 -
8
0
10
0
Observed
Predicted
:.O
rb
>
5 .
Observed
Predicted
o ~ " ' " ' " " " " " " '
0
200
400
600
800
Elapsed time (min)
Elapsed time (min)
100
200
300
Elapsed time (min)
400
100
Observed
Predicted
200
300
400
Elapsed time (min)
8
s
0
Observed
s Predicted
80
120
Elapsed time (min)
Figure 4.63: Observed and predicted surface wetted radius for sandy soil under
application rate of 0.5,0.7, 1.O, 1.4 and 2 lh.
200
400
600
800
200
I00
200
300
Elapsed time (rnin)
80
600
800
Elapsed time (min)
Elapsed time (min)
400
400
50
100
150
200
250
300
Elapsed time (min)
120
Elapsed time (min)
Figure 4.64: Observed and predicted vertical advance of wetting front for sandy soil
under application rate of OS,O.7, 1.O, 1.4 and 2 Ik
radius and vertical advance of wetting front for sandy soil
Vertical advance of wetting front
Field Study of Risse et al., (1989)
Risse et al., (1989) conducted a series of experiments in a Cecil sandy loam soil to
determine the wetted radius for a trickle emitter. An orchard planted with four -year-old
Black peach trees was chosen for the study. The statistical design was a randomized
complete block design of four blocks with two replications of each flow rate within each
block. Each test applied a total volume of 37.8 L with application rates of 3.78 and 7.57
L/h which resulted in irrigation durations of 10 and 5 hours, respectively. The emitter
flow rates were tested and verified to be within 10% of their rated discharge. The extent
of water movement from a trickle emitter was determined using both matric potentials
and a bromide tracer. A single quantitative value for the horizontal wetted radius was
taken to represent the average wetted radius within 30 cm depth of soil. Table 4.15
shows the measured wetted radius values for each tree. The average wetted radius for the
3.78 l h treatment was 37.1 cm while the average wetted radius for 7.57 l k was 35.3 cm.
Natural variability of soil, different conditions and the interpolation error derived from
the plotting process were the main reasons for the differences within observed data.
The inputs required to determine the wetted radius by using Equation 4.8 are application
rates q, volume of water applied V,, saturated hydraulic conductivity k, and average
moisture content A0. The authors presented the input data of application rates and
volume of water applied. However, we estimated k, and 8, using (Schaap et aI., 2001),
ROSETTA software package for each site based on their soil particle distribution.
ROSETTA contains a hierarchy of pedotransfer functions that can be used depending on
the soil characterization data that are available. The lowest order model makes
predictions of hydraulic parameters based on the soil textural class. Table 4.16 shows the
inputs used in the prediction procedure and the predicted wetted radius for each tree. It
also gives the percent error between the observed and predicted values. The average
percent error for both treatments is only 8.7 %, but when broken down by treatments, the
3.78 I h has an average difference of 6.37% while the 7.57 l h treatment has an average
error of 11% when compared to measured wetted radii.
The discrepancies between predicted and observed data may be attributed to the
variances within the observed data. For example, the wide range of observed wetted
radius in case of treatments of 3.71 l/h and 3.56 l k were 30.00 and 45.25 cm. This wide
range cannot be explained using a texture variation of 3% and flow rate variation of less
than 0.4 I/h. In general, the results of comparisons showed that using Equation 4.8 offers
a reasonable estimate of the wetted radius without having to obtain any data except
application rates q, volume of water applied V,, saturated hydraulic conductivity k, and
average moisture content A9.
Table 4.15: Summary of observed wetted radius for Cecil sandy loam soil (adapted from
Risse et al., 1989)
Texture
Tree
Number
Flow rate
l/h
% Sand
%Clay
Wetted radius (cm)

Observed
(Potential)
Observed
(Bromide)
Table 4.16: Comparison of the observed and predicted wetted radius for Cecil sandy
loam soil
Tree
Number
16
Flow
rate (llh)
3.94
Soil properties
Wetted radius (cm)
KS( c d h )
A9
Observed *
Predicted
2.83
0.202
30.50
39.32
22.40
* Average wetted radius as calculated from bromide tracer and water potential data
** Percent of error based on predicted data.
Field Study of Yitayew et al., (1998)
These experiments were conducted at the Campus Agricultural Center University of

Arizona, Tucson, Arizona, on three different soils in different three locations to compare
the surface wetted regimes under a point source. At each site three different constant
flow rates 2, 3, and 4 l/h of water were applied to the soil surface as a surface point
source. The soil physical properties of the experimental soils are summarized in Table
4.17.
Table 4.17: Physical properties of three soils studied by Yitayew et al., (1998)
Sand (%)
Silt (%)
Clay (%)
K, (crnh)
Loam
51
38
11
1.51
Silty clay
40
41
19
1.15
Soil
Loamy sand
The authors calculated the saturated hydraulic conductivity of the three soil types by
using Wooding's equation (1968).
where Q is the steady-state volumetric discharge rate (L~/T)and r, is the saturated radius.
Running Equation 4.8 to determine the surface wetted radius required the hydraulic
parameter of 0,. However, 8, for each soil were estimated using ROSETTA (Schaap et
al., 2001), a pedotransfer function software package that uses a neural network model to
predict hydraulic parameters from soil texture and related data. Inputting the data of
particle size distribution for each soil to ROSETTA resulted in the following parameter
estimates of $, 0.3883, 0.3914 and 0.4008 for Loamy sand, Loam and Silty clay,
respectively.
Figures 4.65 through 4.67 show the extent of the experimental and simulated surface
wetted radii at different time with 2.0, 3.0 and 4 l/h application rates for loamy sand,
loam and silty clay soils. As can be seen from the figures, the computed surface wetted
radiuses agree well with the observed data especially for short duration. On the other
hand, at longer duration the observed data show a faster movement of surface wetted
radius compared with predicted data. The comparisons were found more favorably with
loamy soil compared with the other soils. Figures A.26 through A.28 show the computed
surface wetted radius versus the observed data where the circle symbol represents the
observed -predicted paired data, the solid line represents the equity line (slope equal I)
and the dashed line represents the best fitted line. It is clear that all data points lie very
close to the equity line. On the other hand, the regression lines of the predicted versus
observed values were fitted close to the line of identity. The slope the regression lines
were found close to one; they ranged from 0. 89 to 1.03. These results indicate that the
predicted data located within minus 11% to 3% of observed data. The determination
coefficients of such relationships were found to be in range between 0.90 and 0.99. Both
the slope values of the linear regression equations and their determination coefficients
demonstrate that there is a good fit of the model simulations to the observed data.
Beside these, the performance indicators, i.e, MBE, RMSE and U Theil's coefficients
were also used to test how well the results predicted by the model mimic the observed
results. The calculated values of these parameters are further displayed in Table 4.18. As
can be extracted from this table, the values of MBE were found to range oscillated from
-1.50 cm to 0.88 cm. It can also be noted that the maximum value of RMSE and U
Theil's coefficient, 3.12 cm and 0.063 for RMSE and U Theil's coefficient, respectively,
are close to their optimum value.
Although there are some discrepancies found between predicted and observed data
which may be attributed to the soil variability and modeling error, the suggested model
showed high reliability and capability for determining the surface wetted radius. These
results encourage the use of the suggested model in practice as it has been verified
against real data obtained under field conditions.
18
2o
q= 2 11h Loamy Sand
Observed
Predicted
10
20
30
40
50
60
70
80
Elapsed time (min)
--5
24
22 :
20 2 18 ;
5 16"4
U
0
12 :
35 1 0 :
w 8 6 4 2:
0
0
q= 3 Ilh Loamy sand

0
o
0
'
8
0
oa.0
'
'
10
20
'
30
"
Observed
Predicted
'
'
40
"
50
'
60
"
70
"
80
Elapsed time (min)

26
24
22
20 3
18 E 16 0
14 :
Q)
35 12 :
108:
6 4 2 -
q= 4.0 Vh Loamy sand
8
8
5'
*
ob.-.-'..0
Observed
Predicted
10
20
"
30
"
40
'
'
50
'
60
"
70
80
Elapased time (min)
Figure 4.65: Observed and predicted surface wetted radius for loamy sand soil under
application rates of 2.0, 3.0 and 4.0 l/h
-E 30
q= 2 ~h ~ o a m
soil
S 25
m
$j 2 0 I
u
C
15-
*o
O
1m
10
!o
0
5 -
Observed
Predicted
Elapsed time (cm)

0
q= 3 I/h Loam soil
8
0
Observed
Predicted
50
. . . ' . - . . ' . .
100
150
200
250
300
Elapsed time (min)
40
-6 35 h
.-9
u
30
q= 4.0lfh Loam soil
o
0
2 25
a, 2 0 t:
i 15 1 8
al
10
o
0
:O
I5 ,
V)
oi
0
'
50
'
'
100
150
Observed
predicted
'
200
250
. . .
300
Elapsed time (min)
Figure 4.66: Observed and predicted surface wetted radius for loamy soil under
application rates of 2.0,3.0 and 4.0 l/h
u,
3
-0
B
$a,
g
%
V)
45
40
35
30
25
20
15
q=2 Ilh Silty clay

0
.?j
0 0
0
10
5
0
Observed
Predicted
. . . - ' . - - - ' . . - - '
50
100
150
200
250
300
350
Elapsed time (min)
Observed
Predicted
Elapsed time (min)

0
q= 4.0 I/h Silty clay
0
0
0
0
0
0
50
100
150
Observed
Predicted
200
250
300
350
Elapsed time (min)
Figure 4.67: Observed and predicted surface wetted radius for silty clay soil under
application rates of 2.0,3.0 and 4.0 I/h
Table 4.1 8: Values of the statistical parameters used in comparison for surface wetted
Texture of soil
q (11 h)
MBE (cm)
RMSE (cm)
Loamy sand
Loamy sand
Loamy sand
Loam
Loam
Loam
Silty Clay
Silty Clay
Silty Clay
Field Study of Palomo et al.,-(2002)
The experiments were conducted in an olive orchard at La Hampa, the experimental

farm of the Instituto de Recursos Naturales y Agrobiologia, located at Coria del Rio near
Seville in southwest Spain (37"17'N, 6"3'W, and elevation 30 m). A 0.5 ha plot
containing 29-year-old olive trees planted at a spacing of 7 mx5 m was selected for the
experiments. The irrigation system consisted of a single pipe placed on the soil surface
in each tree row, with five 3 l/h drippers per tree, 1 m apart. The soil of the plot was a
sandy loam of variable depth. A hard limy sandstone pan impedes the penetration of
both roots and water at a depth which varies within the orchard from 0.9 to 2 m. From
the surface to the pan, the soil texture is quite homogeneous, with average values of
14.8% clay, 7.0% silt, 4.7% fine sand and 73.5% coarse sand. The bulk density (pb) of
the soil was found to be in range between 1.5 and 1.3 gm/cm3 with average value of 1.4
gm/cm3. The authors presented data of the measured hydraulic conductivity (k) versus
moisture content (8) in the range near saturation. The evolution with time of the wet
bulb diameter produced from single emitter in the soil surface was measured in and
perpendicular to the pipe direction. In addition, the maximum wetted depth (z) was
estimated using a tensiometer installed at the depth of 0.3 m.
The results from this study were used to verify using Equations 4.8 and 4.9 to determine
the surface wetted radius and the vertical advance of wetting front. Running equations
4.8 and 4.9 required the hydraulic parameters, i.e., saturated hydraulic conductivity (k,)
and average moisture content (Ae). Because there was no available measured value of
the saturated hydraulic conductivity, it was decided to be estimated using extrapolation
technique. The average value of the extrapolated saturated hydraulic conductivity was
obtained to be 8.39 cm/h. Considering the average value of the bulk density (pb) 1.4
gm/cm3;the computed average moisture content become 0.24.
Figure 4.68 shows the location of experimental and simulated surface wetted radius at
different times. Excellent agreements between the simulated and experimental surface
wetted radiuses are obtained for the whole time range. However, at longer duration the
simulated data are slightly lower than observed ones. Statistical analysis was performed
to evaluate the performance of Equation (4.8) in predicting surface wetted radius. The
values of MBE, RMSE and Theil's coefficient (U) were -0.13 cm, 0.37 cm and 0.012,
respectively. The ideal vaIues of MBE, RMSE and Thei17scoefficient (U) are zero. The
performance of the model was further displayed in Figure A.29. In this figure, for a
perfect model prediction all of the points would fall on the line across the graph
signifying a slope of one (observed = predicted).
20
40
60
80
100
120
Observed
Predicted
140 160
180
Elapsed time (min)
Figure 4.68: Observed and predicted surface wetted radius for coarser sandy soil under
application rates of 3.0 l/h
As it appears, the observed- predicted pairs are uniformly scattered closely to the 1:l
line with slope value found to be 0.99. The determination coefficient R* of the predicted
and observed for the linear regression is 0.99. This statistically reasonable agreement
between the observed and predicted modeled results suggests that the model performed
well, thus the model was satisfactorily validated.
The measured time for the wetting front to reach 2-30 cm was 156 min. At this time, the
predicted maximum vertical advance of wetting front was 30.79 cm which indicates
excellent agreement between the measured and predicted value.
The agreement between the experimental and simulated values of r and z was very good,
particularly for z. This shows that the suggested models are valid for the conditions of
the orchard.
Model Evaluation Hydrus-2D
The
experimentally determined moisture characteristic curve
and functional
relationships between diffusivity and soil water content D(0) for sandy and sandy loam
soil as fitted with van Genuchten (1980) using RETC fitting procedure are shown in
Figures 4.69 and 4.70.
Results of the RETC fitting procedure for the hydraulic Van Genuchten parameters,
which used as inputs to the Hydrus-2D model, by using the measured 8 (h) and D (0)
data are presented in Table 4.19. Parameters describing the soil hydraulic properties for
the two soils used in this study (0, is residual water content, 0, is saturated water content,
a, n and 1 are empirical factors and K, is the saturated hydraulic conductivity).
It should be noted that the measured 0 (h) and D (8) were determined under laboratory
conditions; also it was assumed that any hystersis was neglected.
Table 4.19 Soil hydraulic parameters as fitted with Hydrus-2D

Soil type
0,
0,
a (cm-')
ks(crn/min)
Sand
0.042
0.37
0.145
1.77
0.335
0.0001
Sandy loam
Determination of the components of the dispersion tensor Dxx, Dzz and Dxz requires
knowledge of the molecular diffusion coefficient of the solute in free water Dw and the
longitudinal and transversal dispersivities DL and DT. We used Dw =0.0012 cm2/min,
which is the value of the diffusion coefficient of chloride ion in water at 24 C0 (Stewart,
2003), to mimic chloride diffusion. As an approximation, we took the longitudinal
dispersivity equal to one-tenth of the profile depth 2, as this is supported by previous
studies that addressed characterization of solute transport in soils (e.g. Beven et al.,
1993; Cote et al., 2001). This gives DL =0.1 m. We also set DT = DL/ 10=0.01 m (Bear.
0.50
0.40
0.45
0.35
0
Observed
0.40
Observed
0 30
0.35
0.25
K
a,
K
a,
K
-2
0.25
0.20
a,
a,
2
C
([I
0.30
0.15
0.20
0.15
0 10
0.10
0.05
0.00
0
0.05
4000
8000
Pressure head (cm)
12000
0.00
0
4000
8000
12000
16000
Pressure head (cm)
Figure 4.69: Soil water retention curve fitted by van Genuchten (1980) for sandy and
sandy loam soil .
Moisture content
Moisture content
Figure 4.70: Soil water diffusivity as a function of soil water content as fitted by van
Genuchten (1 980) for sandy and sandy loam soil
4.3.1 - Wetting Front Movement
The wetting front is the region where a steep moisture gradient exists. With water
infiltration into relatively dry soil, this front can be determined as the visual limit of
water penetration. If this wetting front can be predicted by Hydrus-2D model, it may
serve as an indicator for the ability of the model to predict water movement in soil under
point source trickle irrigation. For comparison, the location of the wetting front in the
corresponding numerical solution was defined as a line of maximum gradient in water
content
(vei), (Bresler et al.,
1971 and Clair et al, 2003). The wetting front was easily
traced visually with the experimental apparatus. Two axes described the geometry of the
soil-wetting pattern. A radial axis r (at the soil surface) and indicated as surface wetted
radius, and a vertical axis, z, at the middle of the soil, having the origin, point (0, 0), at
the trickle irrigation source.
4.3.1.1 Sandy soil
Figures 4.71 and 4.72 show the actual and the computed surface wetted radius as a
function of elapsed time in minutes for sandy soil under 1.5, 2, 2.7, 3, 3.5, 4.5,4.8 and 6
l/h application rates. In these figures the open circles represent the actual location of the
wetted front, while the solid circles represent the predicted. The figures indicate good
simulation capability of the model in the horizontal direction, especially in the earlier
stage of irrigation. The discrepancies between the numerical and the experimental
surface wetted radius are generally small and may be attributed to the lack of precision
in estimating soil hydraulic characteristics and measurement errors. Haverkamp et al.,
(1977) reported that a relative difference of only 2% in surface water content caused a
relative difference of 24% in the wetting front position.
The results of comparison between observed surface wetted radius and those predicted
by Hydrus-2D model for each trial are plotted against a 1:l line as shown in Figures
A.30, A.31 and A.32. For a perfect model prediction all of the points would fall on the
line across the graph signifying a slope of one (Observed= Predicted). It is clear that the
observed-predicted data pairs lie very close to the 1:l line. This can be also noted by
fitting the predicted data with observed. The slope of the linear fitted line was obtained
to be in range between 0.88 tol.10, which means the percent difference between the
predicted and observed data fluctuates fiom -12 % to +10.0 %. The coefficient of
determination, R2 for the linear regression equations of the predicted and observed data
are 0.93, 0.99, 0.94, 0.97,0.99, 0.99, 0.99, and 0.99 for application rates 1.5,2.0,2.5,2.7,
3.5, 4.5,4.8, and 6.01/h, respectively. Since the value of R2 could be interpreted as the
proportion of the variation in observed value that is explained by the predicted model,
these values indicate that the model predictions are very good. Additional measures were
used to evaluate the model performance, such as: mean bias error, MBE, root mean
square error, RMSE and Theil's coefficient, U. The MBE, RMSE statistics have as a
lower limit the value of zero, which is the optimum value for them, as it is for Theil's
coefficient, U. These statistical parameters are summarized in Table 4.20. The MBE
(ranged between 1.2 1and 2.83 cm) and RMSE (ranged between 1.49cm and 3.03 cm) are
small, which indicates slight deviations between the numerical and experimental results.
The values of U Theil's (ranged between 0.035 and 0.075) put more evidence the fact
that the predicted values are close to zero which, in turn, reflect that the model has a
good performance to simulate wetting front movement under point source trickle
irrigation under the mentioned conditions.
The measured and simulated vertical advance of wetting fronts throughout the sandy soil
profile under application rates of 2.0, 4.5, 4.8 and 6.0 l/h are shown in Figure 4.72 and
4.73. The open squares represent the observed data while the solid squares represent the
predicted one. These values were measured visually along the vertical distance under the
point source (r=O). Again, the simulated wetted vertical distances were defined as the
depth of the wetted soil where maximum gradient in water content (VOi) occurred. As
can be seen from these figures excellent overall agreements were achieved. The
simulated vertical advances of wetting front were nearly identical to the observed values.
For example, under 4.5 l/h application rate, the simulated and measured values at 280
min were 56.0 cm and 56.6 cm, respectively. Similarly, under application rate of 6.0 llh,
the discrepancy between simulated and predicted value was 0.3 cm at 280 min. The
discrepancies can be attributed to the impact of surface topography where there was no
guarantee that point source infiltration area would be exactly circular. On the other hand
the natural variability of the soil properties and the lack of determination soil properties
which determined under laboratory conditions may cause such these discrepancies. The
saturated radius which used for Hydrus-2d simulations was assumed to be constant.
However, in fact as stated previously, this zone is not stable at the beginning of
irrigation, i.e., the saturated radius increases with time. This may also induce such
different between observed and simulated data. The performance of the model was
examined graphically by plotting the predicted versus the observed values. Figures A.33
and A.34 show that the regression line of the predicted versus observed values was close
to the line of unity (y = x) so that a cluster of points in a band along this line indicates a
reasonable fit of the data. The determination coefficients of R~ for the linear regression
equation ranged between 0.92 and 0.99 while the slope values are in the range between
0.87 and 1.05. The adequacy indicators of the model were calculated. Table 4.21
displays the values of the mean bias error MBE, the root mean square error RMSE and
U Theil's coefficient for each treatment. The values of the mean bias error fluctuate
from -2.95 cm to 4.21 cm, indicating an over all agreement between predicted and
observed data. Also the values of RMSE and U Theil's can be considered small as their
values are close to zero. This support the results mentioned above that Hydrus-2D has
the power to simulate the wetting front movement for sandy soil under point source
trickle irrigation.
q=1.5 I/h
e 0
0
-0
Observed
Predicted
c
0
100
200
300
400
500
100
200
300
400
500
Elapsed time (min)
Elapsed time (min)
q=2 5 I/h
q=2.71/h
o o o O
o O * . . *
0 0
0 0
*a8
**
**
. O
* O
*o
8O
t'
100
200
Observed
Predicted
300
400
Elapsed time (min)
500
100
200
Observed
Predicted
300
400
500
Elapsed time (min)
Figure 4.71: Observed and computed surface wetted radius for sandy soil under
application rates of 1.5,2.0,2.5 and 2.7 l/h
q=4.5 I/h
....
oooO
O
0
o0
.*
-8
8
0
100
200
300
400
500
Elapsed time (min)
100
200
Observed
Predicted
300
400
500
Elapsed time (min)
Observed
Predicted
Observed
Predicted
0
0
100
200
300
400
Elapsed time (min)
500
100
200
300
400
Elapsed time (min)
Figure 4.72: Observed and computed surface wetted radius for sandy soil under
application rates of 3.5,4.5,4.8 and 6.0 l/h
radius for sandy soil
Discharge rate (llh)
MBE (cm)
RMSE (cm)
Observed
Predicted
Elapsed time (min)
100
200
300
400
Elapsed time (min)
Figures 4.73: Observed and computed vertical advance of wetting front for sandy soil
under application rates of 2.0 and 4.5 l/h
a Observed
m Predicted
100
Elapsed time (min )
200
300
400
Elapsed time (min)
Figures 4.74: Observed and computed vertical advance of wetting front for sandy soil
of wetting front for sandy soil
Discharge rate (lh)
MBE (cm)
RMSE (cm)
2.0
-2.95
4.39
0.076
0.163
0.95
0.01
4.5
4.3.1.2 Sandy loam soil
Figure 4.75 shows the observed and predicted surface wetted radius as a function of the
elapsed time for sandy loam soil under 2.3, 2.6,3.0 and 4 l/h application rates. Again the
open circles represent the actual location of the surface wetted radius, while the solid
circles represent the predicted. The results show that the experimental and predicted
surface wetting radiuses are in genera1 agreement. This can be noted dearly in case of
the slower discharge rate 2.3 l/h where the simulated data nearly matched the observed
data points. For more clarification, the predicted data were plotted against experimental
data, Figures A.35 and A.36. Consequently in these figures the perfect agreement
between computed and observed data would be located on a 1:l line. Favorable
agreements with the observed data toward slower application rates whereas the slope
values of the best-fit line were 0.959, 0.94, 0.88 and 0.89 for 2.3, 2.6, 3.0, and 4.01/h
application rates, respectively. The slope values of the best fit regression equation reveal
that the percent difference between the predicted and observed data oscillated from -12
% to -4.0 %. The determination coefficients ranged between"0.93 to 0.99", which means
the obtained fitted equations have high accuracy to describe the relation between
predicted and observed data. Furthermore, the performance indictors MBE, RMSE and
U Theil's coefficient were found relatively small, whereas the maximum values of these
indicators are -2.65 cm, 2.92 cm and 0.07, respectively (Table 4.22). Figures 4.76 and
4.77 show the measured and calculated vertical component of the wetting fiont under
2.3, 2.6,3 and 4.0 l/h versus elapsed time. The highest agreements between the observed
and calculated were found for the higher application rates, i.e., 3 and 4 l/h. The
discrepancies between the calculated and measured positions of the vertical wetting front
may be due to variation in the size of the surface source of water during infiltration and
to the natural variation of soil properties.
Figures A.37 and A.38 show the plots of the calculated vertical advance of wetting front
versus the observed data. These figures illustrate that the model captured the observed
data moderately. For 4 l/h application rate, the results showed that data points more
perfectly scattered along the equity line than do under the other application rates. The
slope values of the regression equations, used to describe the best-fit relation between
the observed and predicted data, are displayed within the mentioned figures. As can be
concluded from these values, the differences between the calculated and observed data
are in the range between -18% and - 8%. In addition, the regression equations give high
degree of correlation between the calculated and the observed data as indicated by the
coefficient of determination which ranged between 0.96 to more than 0.99. The
maximum values of MBE, RMSE and U Theil's coefficient were found to be -5.68 cm,
6.01 cm and 0.095, respectively (Table 4.23). Although these values are relatively higher
than those for the previous comparisons, it can be acceptable according to the complex
mechanisms of water transport under the complicated boundary and initial conditions
fi-om a surface point source.
q= 2.6Ilh
30-
0 0
V)
25-
0 0 0 .
2
$ 20:
O "
0 .
Observed
Predicted
100
200
300
400
500
Elapsed time (min)
Elapsed time (min)
q= 4.0Ilh
q= 3.0Ilh
0
O .
0 0
O
.*
.
.
O .
o @
. 0 .
0
.
. o O :
.*
0.
0
0
>*
Observed
Predicted
Observed
Predicted
0
0
100
200
300
Elapsed time (min)
400
500
100
200
300
400
500
Elapsed time (min )
Figure 4.75: Observed and computed surface wetted radius for sandy loam soil under
application rates of 2.3, 2.6, 3.0 and 4.0 l/h
radius for sandy loam soil
MBE (cm)
Discharge rate (lh)
RMSE (cm)
..
.
.
.
0
@.6 Ifh
0
0
.
0
0 .
i
0
Observed
Predicted
Elapsed time (min)
Observed
Predicted
100
200
300
400
500
Elapsed time (min)
Figure 4.76: Observed and computed vertical advance of wetting front for sandy loam
soil under application rates of 2.3and 2.6 l/h
Observed
Predicted
100
200
300
400
500
Elapsed time (min)
100
200
Observed
Predicted
300
400
500
Elapsed time (min)
Figure 4.77: Observed and computed vertical advance of wetting front for sandy loam
soil under application rates of 3.0 and 4.0 l/h
of wetting front for sandy loam soil
Discharge rate (llh)
MBE (cm)
RMSE (cm)
4.3.2 Moisture Content Distribution
Quantitative comparison between the theoretical and experimental moisture content

results is of prime importance from both a practical and theoretical point of view. The
observed moisture content distribution data obtained from field experiments with the
sandy and sandy loam soils are compared with the calculated solution using Hydrus-2D
model (1999).
4.3.2.1 Sandy Soil Profiles
Figures 4.78 through 4.85 show the measured and simulated moisture content
distribution for each trial. Each figure contains contour plots of the measured and
simulated moisture content profiles. The contours in the measured profile were drawn
using kriging interpolation algorithm. However the data are relatively sparse, one should
not attach too much significance to contour details (Skaggs et al., 2003). Both measured
and simulated results show that in highly permeable coarse-textured soils such as the
sand used in this study, water drains easily and quickly because gravity dominates. The
irregular shape of the contour lines in case of observed data may be attributed to
variability in soil. The other factor which may be the main reason for such discrepancies
is the input data, i.e., the values of K,, a, n and 1, determined in laboratory, which
normally differ with field data. In addition as mentioned previously the saturated radius
is dynamic and not stable with time. Hydrus 2D model does not calculate this dynamic
boundary condition. This is additional reason may induce such discrepancies.
Nevertheless, it is clear from the contour plots, the predicted pattern of wetting are in
good agreement with the spatial distribution of the moisture content.
The mean bias error MBE, the root mean square error RMSE and U Theil's coefficient
for the simulated and measured volumetric water contents provide a quantitative
measure of the goodness of fit between the observed and simulated data. Table 4.24
shows the values of model performance indicators for each trial. As can be noted from
the Table, the values of model performance indicators are relatively approaching to their
optimum. For instance, the MBE values range from 0.0006 to 0.040 m3 m" while the
values of RMSE oscillates between 0.017 and 0.06.In addition, the Theil's coefficient
values were found to be between 0.120 and 0.054. Although there are some relatively
high values of these indicators, it can be acceptable according to the complex
mechanisms of water transport under the complicated boundary and initial conditions
from a surface point source.
The transect plots in (Appendix, Figures A.39 through A.54) permit a more objective
comparison. These figures show the experimental data along with the corresponding
theoretical curves. The close symbols represent experimental measurements while the
smooth curves are simulated results. Figures A. 39 through A. 49 show the moisture
content distribution at depths z = 7.5, 17.5,27.5 and 37.5 cm below the soil surface. The
values displayed within each figure represent the performance indicators, i.e., MBE,
RMSE, and U Theil's coefficient. As mentioned earlier, these figures show that the
predicted soil moisture content agree quite well with the measured data. For example,
under the slower discharge rate, i.e, 1.5 l/h and at z = 7.5 cm these values were found to
be -0.0077 m-3 m", 0.01 1 m-3 m-3 and 0.021 for MBE, RMSE and Theils's coefficient
(U), respectively. On the other hand, at z
37.5 cm the MBE, RMSE and Theils's
coefficient (U) were -0.015me3m-3, 0.024 m-3 m-3, 0.164, respectively. Under higher
application rate, i.e, 6.0 l/h and at z= 7.5 cm, these values were -0.038 m3 mm",0.061
m3m-3 and 0.086 for MBE, RMSE and Theils's coefficient (U), respectively. For deeper
d
for MBE,
depth z= 37.5 cm, these values were -0.007 m3 m", 0.049 m3 ~ n - ~ a n0.101
RMSE and Theils's coefficient (U), respectively.

The measured and predicted moisture content distribution at selected vertical planes
r= 2.5, 12.5, 17.5, 22.5 and 27.5 cm away from the trickle source are given in Figures A.
47 through A. 54. These figures show the experimental moisture distribution curves
agree in general with those predicted by the Hydrus-2D model. For example, with
application rate 1.5 l/h and at vertical plane where r = 2.5 cm the performance indicators
were -0.0089 m m", 0.017 m mv3and 0.040 for MBE, RMSE and Theil's coefficient U,
respectively. At vertical plane where r= 17.5 cm the values of performance are close to
their optimum values 0.0002 m3 m'3, 0.02 m3 m"and 0.092 for MBE, RMSE and U
Theil's coefficient, respectively. Under 6.0 l/h application rate the performance
indicators were -0.014 m3 m-3, 0.051m3 m-3 and 0.083 for MBE, RMSE and Theil's
coefficient U, respectively at r= 2.5 cm. For vertical plane closer to wetting front where
r= 17.5 cm the performance indicators were 0.002 m m - 0.066 m m - and 0.142 for
MBE, RMSE and U Theil's coefficient.
As a conclusion, the results show the experimental moisture distribution agrees in

general with those predicted by the numerical solution of Equation 2.49. The
disagreements that exist between the calculated and experimental seen in some of the
plots may have been caused by the natural variability in the soil profile. The Hydrus-2D
model described the water content distribution quite well at relatively high levels of
moisture contents; however, it did not do as well at lower moisture contents. Another
reason demonstrated the discrepancies between the measured and the predicted moisture
distribution within the soil profile is the redistribution of water after infiltration during
the time required for sampling the soil profile for determination moisture content.
Table 4.24: Values of the statistical parameters used in comparison for moisture
Statistical parameters
Flow rate l/h
MBE (m3mJ)
RMSE (m3 m")
U' Thiel's
Distance (cm)
Distance (cm)
10
Observed
20
30
40
50
Predicted
Figure 4.78: Observed and predicted moisture distribution under application rate of 1.5 l/h for sandy soil, elapsed time 320 min
Distance (cm)
Observed
Distance (cm)
Predicted
Figure 4.79: Observed and predicted moisture distribution under application rate of 2.0 l/h for sandy soil, elapsed time 360min
Distance (crn)
Predicted
Observed
Distance (cm)
10
20
30
Distance (cm)
40
L
C
10
20
30
40
50
Observed
Figure 4.81 : Observed and predicted moisture distribution under application rate of 2.70 l/h for sandy soil, elapsed time 420min
.Distance (cm)
Observed
Predicted
Distance (cm)
Observed
Distance (cm)
Predicted
Distance (cm)
DIstance (cm)
10
Observed
20
I
30
I
40
Predicted
Distance (cm)
Distance Icml
Predicted
Figure 4.85:
340 min
4.3.2.2 Sandy Loam Soil Profiles
Figures 4.86 through 4.89 show the measured and simulated soil moisture distribution in
sandy loam soil for the application rates 2.3, 2.6, 3 and 4 l k respectively. Unlike the
sandy soil, the pattern of moisture distribution took less ellipsoidal shape. In several
plots, the measured moisture distributions were irregular and non-uniform while the
simulated moisture distribution appeared to be much smoother. This could be explained
by the natural variability of the soil and by the fact that the simulated data were based on
the assumptions that the soil profile is isolated and isotropic. The simulated values of
moisture content are closer to measured values especially in the horizontal direction
while in the vertical direction the measured value are larger than simulated value. For
example, under application rate 2.6 l/h, the simulated moisture distributions were in
excellent agreement with the measured data in horizontal direction while in the vertical
direction the measured values were larger than predicted. However, the discrepancies
can be noted more clearly at low moisture content values, i.e. at the edge of wetting
zone. It seems the main reason for disagreement is due to the low estimate of the
saturated hydraulic conductivity that has been adopted for the calcu1ations.
The
adequacy of the Hydrus-2D model was evaluated by the mean bias error MBE, the root
mean square error root RMSE and Theil's inequality coefficient (U) criteria. Values of
these indicators for each application rate are given in Table 4.25.
Table 4.25 indicates that MBE and RMSE for each trial are acceptable as their values
obtained are relatively small. In addition, the values of Theil's inequality coefficient U
for all trials are close to zero. In order to facilitate the comparison between the simulated
and experimental data, the moisture content was plotted as a hnction of the distance for
every single space coordinate.
Table 4.25: Values of the statistical parameters used in comparison for moisture
Flow rate I/h
MBE (m3 m")
RMSE (m3 m")
U'Thiel's
In Figures A.55 through A.58 the moisture content distributions are presented along the
radial plane at depths z= 7.50, 17.50,27.50 cm, i.e., [8 (r, 7.50), (r, 17.50), (r, 27.50) and
ry
(r, 37.50)] at the four application rates. The values displayed within each figure
represent the performance indicators, i.e., AME, RMSE, and U Theil's coefficient. The
simulated data were nearly identical to the observed values at the depths closer to the
dripper but at further depths the discrepancies were increased. This can be also extracted
from the values of performance indicators. For example, under application rate of 2.3 l h
and at depth of 7.50 cm the values of
AME, RMSE, and Theil's coefficient U were
0.0008 m3m3, 0.01 9 m3m3and 0.028, respectively. On the other hand, at depth of 27.5
cm, these values increased to be -0.056 m3m-3, 0.074 m3m-3 and 0.177 for
AME,
RMSE, and Theil's coefficient U, respectively. Figures A.59 through A.62 show the
moisture content distributions along vertical plane at distances of 2.5, 12.50, 17.50 and
22.50 cm away from trickle source, i.e.
[0 (2.50, z), (12.50, z), (17.50, z) and (22.50,
z)]. The results show the measured moisture content distributions agree in general with
those predicted by the model solution. As can be noted the measured moisture content at
depths near to soil surface are less than those predicted by Hydrus-2D. These
discrepancies can be attributed to the effect of surface evaporation where the
experiments conducted in summer under high temperature condition ranged between
40
and 47' C. The values 'of performance indicators AME, RMSE, and Thei17s
coefficient U showed excellent agreement between measured and predicted data of

moisture distribution within the soil profile. For example, the values of AME, RMSE
and Theil's coefficient U oscillated between -0.031 and 0.023 m3m", 0.024 and 0.063
m3 m-3 and 0.043and 0.093, respectively.
Overall, considering the difficulties in estimating the dynamic water conditions in the
field there was generally good agreement between the measured and calculated values,
especially in the lateral direction. In the deeper downward direction the simulated
moisture content distributions were less than the measured. The differences were more
apparent at lower moisture contents and seem to agree with the literature. Skaggs et al.
(2003) found that the Hydrus-2D model described moisture content distribution quite
well at relatively high levels of moisture content; however, it did not do as well at lower
moisture contents. Another reason for the slight discrepancies in moisture distribution in
some of the plots may be due to the redistribution of water after infiltration during the
time required for sampling the soil profile for the determination of moisture content.
Distance (cm)
Distance (cm)
Observed
Predicted
Figure 4.86: Observed and predicted moisture distribution under application rate of 2.3 llh for sandy loam soil, elapsed time 270 min
Distance [cm)
Distance (cm)
0
0A
I0
1 5 2 0 2 5 3 0 3 5 4 0 4 6 5 0
'---
't
Observed
Predicted
Figure 4.87: Observed and predicted moisture distribution under application rate of 2.6 l/h for sandy loam soil , elapsed time 400min
Distance (cm)
Distance (cm)
Predicted
Figure 4.88: Observed and predicted moisture distribution under application rate of 3.0
I/h for sandy loam soil, elapsed time 350min
Distance (cm)
Distance (cm)
0
1 0 1 5 a 0 2 5 3 0 3 6 4 0 4 5 5 I ]
Predicted
Figure 4.89: Observed and predicted moisture distribution under application rate of 4.0
I/h for sandy loam soil, elapsed time 360min
4.3.3 Solute distribution
To evaluate the reliability of the Hydms-2D model to simulate solute distribution within
the soil profile under point source trickle irrigation, it is desirable to be compared with
experimental results. That one can gain confidence in their consistency. The field
infiltration experiments provide good opportunity for such comparisons. This section
presents the field results of the solute movements from point source trickIe irrigation as
compared with to Hydms-2D model.
4.3.3.1 Sandy soil
Figures 4.90 through 4.97 show the measured and the simulated chloride concentration
distribution within the soil profile for 1.5, 2.0,2.7, 3.5,4.5,4.8, 6.0 l/h application rates,
respectively. The chloride concentration was expressed in milliequivalents of chloride
per liter of soil water. Each figure contains contour plots of the measured and simulated
chloride concentration profiles. The contours in the measured profiles were drawn using
a kriging interpolation algorithm. However, because the data are relatively sparse, one
should not attach too much reliance to the contour details. Again, the contour lines of the
model data are smoother than those of measured values. The figures show relatively
good agreements between the observed and simulated chloride concentration
distributions in most of the treatments, especially near the soil surface. The comparison
between simulated and measured solute distribution data are further presented in Table
4.26. The Table summarizes the three statistics: Mean bias error (MBE), root mean
square error (RMSE), and Theil's coefficient (U) against application rates. The values of
MBE, RMSE and Theil's coefficient (U) fluctuate from -3.24 to -0.081, 1.71 to 8.92 and
0.09 to 0.12 meqll, respectively, indicating slight deviation between the numerical and
measured values. The MBE showed more agreement compared to Theil's coefficient
(U). This result maybe due to the high sensitivity of Theil's coefficient (U) compared to
MBE. The transect plots provide a better visual comparison between the measured and
calculated distributions. Figures A. 63 through A. 70 show the solute distribution in
various horizontal direction at various depths z = 7.50, 17.50, 27.50 and 37.50 cm below
the soil surface. The results in these figures show that the observed and simulated solute
distributions are nearly constant especially at the shallow depths with a slight decrease
as the horizontal distance increases. The observed Chloride concentrations are slightly
higher than the simulated especially near the soil surface. This may be due to the effect
of high evaporation which was not accounted in the Hydrus- 2D model. In general, the
model results agree reasonably with the measured ones. The statistical indicators, which
are displayed within each section show that the greatest agreement occurred directIy
under the emitter while the lowest agreement can be noted at locations further away
from the emitter source. For example under application rate of 1.50 llh with Chloride
concentration 54 meqll, these values were 0.20 meql 1, 9.58 meqll and 0.128 for MBE,
RMSE and U Theil's, respectively at depth 7.50 cm. For deeper location, under depth
27.50 cm the values of statistical indicators increased to be -5.79 meqll, 10.33 meqll and
0.15 for MBE, RMSE and U Theil's, respectively. Figures A.71 through A.78, show the
data along the vertical axis (z) at four different horizontal distances from the emitter
source r
2.50, 12.50, 17.50 and 22.50 cm, i.e. c (2.5, z), c (12.5, z), c (17.5, z) and c
(22.5, z). As mentioned previously there is a tendency of a slight increase of measured
chloride concentration compared to the simulated ones. For locations close to the emitter
source the model performance was quite good, but less acceptable as the distance from
the emitter increased. For example, under application rate of 3.5 l/h with Chloride
concentration of 20 meqll the values of MBE, RMSE and U'Thei17s increased from
-0.68, 1.70 meqll and 0.058 to be 1.31meq/173.57 meq/l and 0.15, respectively as the
section distance increased from 2.50 cm to 22.50 cm.
While there is some disagreement between the predictions and observations, overall the
predictions are very good, particularly considering that the simulations were done
without any fitting to the solute transport data, and without characterizing in detail the
soil at each experimental site.
Table 4.26: Values of the statistical parameters used in comparison for solute
Flow rate l/h
MBE (meqll)
RMSE (meqll)
U'Thiel's
4.3.3.2 Sandy loam soil
The results of the model evaluation for the application rates of 2.3, 2.6, 3.0, and 4.0 llh
are presented in Figures 4.98 through 4.1 01. The model data seem smooth, while the
lines of equal chloride concentration, in cases of measured data were not smooth, which
was expected. There was a slight tendency of over-estimation of chloride concentration
in the horizontal direction as shown in the figures, which may be due to water and solute
loss in the actual measurement/analysis with gravimetric method. It is clear from these
figures that the chloride concentrations moved somehow deeper through the profile than
was predicted by the model. This may be explained as a result of anion exclusion. In
soils with a high negative charge, the soil water near the clay surface will have a higher
anion concentration in addition to being more mobile. This phenomenon causes chloride
to move faster through many soils than nonexcluded and noninteracting ions (Bresler,
1973, Cassel et al., 1975; van der Pol et al., 1977; Muluneh et al., 1995). The effect of
anion exclusion on the rate of Chloride movement in the soil may explain part of the
discrepancy between measured and calculated data. This discrepancy may be related to
minor deviation from the experimental assumptions, to soil homogeneity, and to
measurement errors. The comparison between the simulated and observed chloride
distribution data in the soil profile is presented in Table 4.27. This table summarizes the
three statistics: mean bias error (MBE), root mean square error (RMSE) and U Theil's
coefficient versus each application rate. According to the results indicated in Table 4.29,
the highest agreement between observed and simulated data resulted under the greatest
application rate 4.0 l/h where the values of MBE, RMSE and U Theil's coefficient were
-O.Olmeq/l, 1.82 meqll and 0.07, respectively. The accuracy of model maybe more
related to the input concentration where the lowest agreement resulted under highest
input concentration and vice versa. Figures A.79 through A.82 show the horizontal
distribution of chloride content at selected depths, z= 7.50, 17.50 and 27.50 cm (and
37.50 cm for 4.0 I/h application rate) below the soil surface. The observed results show a
similar pattern of chloride distribution like predicted ones but with slight increasing at
the deepest zone. On the other hand, the observed and predicted Chloride content are
nearly constant with a slight decrease as the horizontal distance increases. However, at
the edge of the wetted volume the Chloride concentration reach minimum concentration.
The performance model indicators within each plot show that the values of MBE, RMSE
and U Theil's coefficient are acceptable. For example, under application rate of 2.3 l/h
(c= 30 meqll) the maximum values of AME, RMSE and U Theil's coefficient were 1.87
meqll, 5.53 meqll and 0.154, respectively. The comparisons between observed and
predicted data of chloride distribution within the soil profile were further described
through Figures A.83 to A. 86. These figures show the chloride distribution in four
vertical planes at r= 2.50, 12.50, 17.50, and 22.50 cm away from the trickle source.
These figures show that both the observed and the predicted chloride content decrease
with depth. The agreement between the observed and predicted data of Chloride content
in general is quite good, except at the distance further away from the trickle source
where the discrepancies between the observed and predicted chloride content are high.
For example, under 3.0 l h (c= 47 meq 11) the values of AME, RMSE and U Theil's
coefficient increased from 1.44 meqll , 4.10 meq/l and 0.05 to 1.93 meqll, 6.53 meqll
and 0.138 respectively as the distance of the transect increased from 2.50 cm to 22.50
cm. This result would indicate that the Hydrus-2D model performed well for the
locations close to the trickle source.
Table 4.27: Values of the statistical parameters used in comparison for solute
Flow rate I/h
MBE (meq/l)
RMSE (meqll)
U'Thiel's
Distance (cm)
10
20
30
Distance (cm)
40
Observed
50
Predicted
Figure 4.90: Observed and predicted chloride distribution under application rate of 1.5 l/h for sandy soil
Distance (cm)
Distance (cm)
Observed
Predicted
Figure 4.91: Observed and predicted chloride distribution under application rate of 2.0 llh for sandy soil
Distance (cm)
Distance (cm)
0
10
20
30
40
-1
-2
-3
-4
Observed
-4
Predicted
KC
Distance (cm)
Distance (cm)
20
30
40
Observed
Predicted
Distance (cm)
Observed
Distance (cm)
Predicted
Figure 4.94: Observed and predicted chloride distribution under application rate of 3.5 I/h for sandy soil
10
Distance (cm)
20
30
40
Distance (cm)
Observed
Figure 4.95: Observed and predicted chloride distribution under application rate of 4.5 lk for sandy soil
Distance (cm)
10
20
Distance (cm)
30
20
Observed
30
Predicted
Distance (cm)
Distance (cm)
10
20
30
40
Predicted
Distance (cm)
Distance (cm)
0
1 0 15 20
25
30
35
40
45
Observed
50
-4
Predicted
Figure 4.98: Observed and predicted chloride distribution under application rate of 2.3 l/h for sandy loam soil
Distance (cm)
Distance (cm)
Distance (cm)
- 45
- 50.
Distance (cm)
Observed
I
-4 5
- 501
Predicted
Distance (cm)
Distance (cm)
Predicted
CHAPTER 5
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
Summary
Trickle irrigation is a popular practice in many parts of the world. It allows high water
application efficiency continuous high level of readily available soil water content and
direct application of fertilizers. A clear advantage of trickle irrigation is that soluble
fertilizers can be applied directly to the plant root zone with both high uniformity and
application efficiency. The speculation on shape of water and solute distribution within
the wetted volume from a point source is important to discover before the efficient
design, operation and management of a trickle irrigation system. It is within this volume
that a plant obtains water to meet its nutrient requirements. To obtain maximum crop
production per unit of water applied, water must be located in a soil volume where the
plant roots have easy access to water and nutrients. The purpose of this work was to
study water and solute movement within the soil profile irrigated from a trickle point
source. There were three main components: laboratory and field evaluation of the water
and solute movements from a point source trickle irrigation using gravimetric method,
developing a new method to determine the major components of wetting pattern
produced from a point surface trickle irrigation i.e., surface wetted radius and vertical
advance of wetting front and finally evaluating The HYDRUS-2D model for predicting
water and solute movement under point source trickle irrigation.
5.2
Conclusions
Many conclusions can be drawn from this study:

1. The lateral wetting front immediately below the point source in the plexi glass
experiments and the surface wetted radius in wooden box and field experiments
moved very fast at the beginning and then slowed down as duration of irrigation
increased while the vertical advances of the wetting front progresses steadily
through the irrigation. This can be expected because of the'additional gravity
component and high K (8) within narrow wetness range associated with lighter
textured soil.
2. Both the main components of the soil wetted pattern under point source trickle
irrigation, the horizontal and vertical movements of the wetting front, have good
relationships with the elapsed time. These relationships can be approximated by
a power equation of the form of a T ~where
,
T is the elapsed time, a and b are
empirical parameters. It was found that during irrigation a linear relation ship
exists between the vertical movement of the wetting front and the square root of
elapsed time
3. In the laboratory experiments, water application rates caused a notable effect on

horizontal advance of the wetting front where the increase of water application
rates caused an increase in horizontal advance of the wetting front. However, in
the field experiments, the surface wetted radius was not affected by the discharge
rate. The discrepancies between the laboratory and field experiments results can
be attributed to many factors which have very significant effect on the water
movement and distribution in any soil, like bulk density, soil structure, surface
condition and initial moisture content.
4. The relationships between lateral advance and total water applied can be
described in the form o f , a V
b,
where V is the total water applied, a and b are
empirical parameters, suggesting that for a certain volume of water applied, a

corresponding lateral advance of wetting front can be predicted. This relationship
can be applied also in case of vertical advance of wetting front.
5. The lateral spread of water movement was found to be higher in sandy loam soil
compared to the sandy soil. On the other hand the vertical spread in sandy loam
was less than sandy soil. The limited effect of gravity forces on the water
movement in the sandy loam soil as being finer textured soil is the main reason
that caused this type of water movement, which was apparently quite different
from the distribution for the sandy soil.
6. The trickle irrigation discharge rate affects the transient saturated radius on the
soil surface. The saturated radius reaches steady values earlier in sandy soil than
sandy loam soil. The saturated hydraulic conductivity as predicted by Wooding's
equation for steady state infiltration from circular saturated area matched with
the measured values obtained by using constant head method where 23 and 26%
of variations were found for sandy and sandy loam, respectively. A wider range
of discharge rates is useful and helps to minimize error in estimation of saturated
hydraulic conductivity of being studied.
7. The soil moisture content after irrigation is higher at distances closer to the
source of irrigation. The statistical analysis showed that there is no significant
effect of water application rate on the soil moisture content in the region within
the lateral distance of 17.5 cm and vertical distance of 27.5 cm.
8. The patterns of the chloride concentrations distribution were similar to those for
moisture content distribution where the vertical spread was more than horizontal
for all application rates. This means that once a soluble non-reactive ion, like C1 -
,begins to move with water, its back diffusion is slow, which means that ion and
water are closely coupled.
9. In most treatments, the effect of water application rate on the solute distribution
in soil was insignificant within the selected region with vertical transactions
ranged from 2.5 to 22.5 cm and horizontal transactions ranged from 2.5 to 37.5
cm.
10. The soil type, the volume of water applied to the soil, and emitter discharge rate,
are the major factors affecting the wetted zone geometry. Two expressions,
~ e - 0 . 5 6~ ~ 0 . qw
2 6 -0.03k -0.03
s
d,
z=
Ae-0.38
~qw -0.'ks0.'9,
~ .
relating
~ the
~
surface wetted radius and vertical advances of wetting front to the saturated
hydraulic conductivity and the average change of soil moisture content, the
volume of water applied to the soil, and emitter discharge rate were developed
using four published laboratory experiments results, i.e. Taghavi et al, 1984;
Anglelakis et al, 1993; Hammami et al, 2002; and Li et al, 2003. The results of
validation tests which were performed by determination coefficient (R~),root
mean square error RMSE, mean bias error MBE, U Theil's inequality coefficient
and graphical comparisons, reveal that both equations have high predictive
power and ability to track the path of the actual observations.
11. Sensitivity analyses of those equations indicated that contrasting to the surface
wetted radius the vertical advance of wetting front is more affected by emitter
application rate. Unlike the surface wetted radius, the vertical advances of the
wetting front would be increased as saturated hydraulic conductivity increased.
This may explain the various shapes of the wetting pattern between various soil
textures. Furthermore, the volume of water applied was found to be the crucial
factor in both equations where its increase would induce increase in the surface
wetted radius and vertical advances of the wetting front.
12. The suggested equations were verified by comparing the outputs with observed
data obtained from laboratory and field experiments and with other published
experiments. As stated in the text the results of these comparisons encourage the
capability of using these equations in practice.
13. The quantitative discrepancies that have been observed in some cases can be the
result of one or several of the following:
i) Inadequacy of the adopted assumptions as it simplifies a very complex
process.
ii) Inability to create uniform initial conditions in the field.

iii) Lack of precision in estimating the soil water parameters i.e., saturated
hydraulic conductivity and saturated moisture content.
iv) The difference of the atmospheric conditions could account for some of these
discrepancies where the predicted equations were developed based on
experiments were conducted under laboratory conditions.
v) Then natural variability of the soil also could account for some of the
observed differences.
14. The numerical HYDRUS-2D was used to simulate water and solute movement
in soil under point surface trickle irrigation. The simulation of water movement
was conducted for surface wetted radius, vertical advance of wetting front and
the distribution of moisture content in the soil profile. Simulation positions of the
wetting front were in agreement relative to the observed measurements of the
wetting front. Specifically, in the lateral the experimentally determined wetting
front was closely estimated by the model. However, in the downward direction
the simulated wetting front advanced much slower than the observed especially
at later stage of infiltration.
15. Considering the difficulties in estimating the dynamic water conditions in the
field there was generally good agreement (especially in the lateral direction)
between the measured and simulated values. In the deeper downward direction
the simulated moisture content distributions were less than the measured. On the
other hand, the Hydms2D model described the water content distribution quite
well at relatively high levels of moisture contents; however, it did not do as well
at lower moisture content.
16. The agreement between experimentally and numerically computed solute

distributions, which where reasonably good for the simple experimental
technique used, show that the solute transport during infiltration may be
quantitatively predicted by the previously described technique (HYDRUSZD).
17. The disagreement that exist between the simulated and experimental in results
some of the plots may have been caused by the natural variability of soil profile.
Another reason demonstrated the discrepancies between the measured and the
simulated moisture distribution within the soil profile is: the redistribution of
water after infiltration during the time required for sampling the soil profile for
determination moisture content. The other factor which may be the main reason
for such discrepancies of water and solute movement is the input data where they
were determined in the laboratories which normally are quite different than under
field conditions.
Recommendations for Further Work

1. As stated in the text, all the soils that were used in this work were classified as
homogenous sandy and sandy loam soil and had very dry initial water content.
Different textured soils could be tried to study the water distribution for other type of
soil and for a layered soil profile. Different initial water content for the soil profile
could also be tried to determine the effect of initial water content of the soil on the
water movement, the moisture pattern and the water content distribution at any time
in the soil profile.
2. Based on the results of this study, it is suggested that adjustment in emitters and
lateral spacing affected by the soil property must be made. Wider spacing in the case
of sandy loam soil and closely spaced in sandy soils because of the greater vertical
than horizontal movement of irrigation water in the soil profile.
3. Trickle irrigation designers can use the simplified models, Equations 4.8 and 4.9, to
determine the number of emitters per tree or the distance between emitters for row
crops and to determine the depth of the wetting zone. The designers can also
maximize the emitter outlet size based on the soil type since the surface wetted
radius is more a function of the volume of water applied than the rate at which it is
applied. Using emitters with higher discharge rates will result in higher application
efficiency and in fact will reduce the irrigation time and probably the energy use to
run the system.
4. As obvious from the results of this study, the saturated hydraulic conductivity under
field condition can be obtained using the relationships between saturated water entry
and emitter application rate. Wider range of discharge rates is useful and helps to
minimize error in estimation of saturated hydraulic conductivity being studied.
5. Further studies should include the sink term representing water and nutrient uptake
by plants, and field situations where the interaction of roots with soil moisture and
evapotranspiration play an important role in the water and nutrient distribution

patterns in the root zone.
6. Based on the results of this study, it is recommended in future studies to use

HYDRUS-2D model under different soil and water salinity conditions in cropped
field. However, for good results the soil hydraulic and solute parameters should be
determined on nondestructive sample or in- situ under field conditions rather than
under laboratory condition.
REFERENCES
Abramowitz, M and I. A. Stegun,1964. Hand book of mathematical functions. Natl
Bur Standards Appl Math Series vol 55. US Government Printing Office
Washington, DC
Ah Koon, P. D., P. J. Gregory, and J. P. Bell, 1990. Influence of drip irrigation
emission rate on distribution and drainage of water beneath a sugar cane and a
fallow plot. Agric. Water. Manage. 17: 267-282.
Alemi, M. H. 1980, Distribution of water and salt in soil under trickle and pot
irrigation regimes. Agric. Water. Manage. 3: 195-203.
A1 -Qinna M.1 and A. M Abu-Awwad, 2001. Wetting Patterns under Trickle Source
in Arid Soils with Surface Crust, Journal of Agricultural Engineering Research,
80(3): 301-305.
Angelakis, , A. N, T. N. Kadir, and D.E. Rolston, 1993. Time-dependent soil-water
distribution under a circular trickler source. Water. Management 7:225-235.
Armstrong, C. F. and T. V. Wilson, 1983. Computer model for moisture distribution
in stratified soils under a trickle source. Transactions of the ASAE, 26: 575582.
Assouline, S., 2002. The effects of microdrip and conventional drip irrigation on
water distribution and uptake. Soil Sci. Soc. Am. J. 66, 1630-1 636.
Bar-yosef, B. and M. R.SheikhoIslami, 1976. Distribution of water and ions in soils
irrigated and fertilized from a trickle source. Soil. Sci. Soc. Am. Proc. 40:575582.
Bear, 5.1972, Dynamics of fluids in porous media. Elsevier, New York
Benami, A. and A. Offen, 1983. Irrigation engineering. Irrigation engineering
Scientific Publication: 161-163.
Ben-Asher, J., CH.Charach, and A.Zeme, 1986. Infiltration and water extraction
from Trickle Irrigation Source: The effective hemisphere model. Soil Sci. Soc.
Am. J50: 882-887.
Beven KJ, Henderson DE, Reeves AD 1993, Dispersion parameters for undisturbed
partially saturated soil. J Hydro1 143: 19-43
Biggar, J.W. and Nielsen, D.R., 1962. Miscible displacement. 111. 'Theoretical
considerations.Soi1 Sci. Soc. Am. Proc., 26: 2 16--221.
Black, C. A. 1965, Method of soil Analysis Partl. By American Society of

Agronomy, INC.
Brandet, A., E. Bresler, N. Diner, I. Ben-Asher, J. Heller, and D.Goldberg.1971,
Infiltration from a Trickle Source: I. Mathematical Models. Soil. Sci. Soc.
Am. Proc.35: 678-682.
Bresler, E. 1973, Simultaneous transport of solutes and water under transient
unsaturated flow conditions. Water Resource. Res. 9: 975-986.
Bresler, E., and D. Russo, 1975. Two-dimensional solutes transfer during
nonsteady infiltration: Laboratory test of mathematical model. Soil Sci.
Soc. Am. Proc. 585-587.
Bresler, E., J. Heller, N. Biner, I. Ben -Asher, and D.Goldberg,l971 .Infiltration
from a Trickle Source: 11. Experimental Data and Theoretical Predictions.
Soil Sci. Am Proc.35: 683 -689.
Bresler, E. 1975, Two-dimensional transport of solutes during non-steady
infiltration from a trickle source. Soil Sci Soc.Am.Proc.39: 604-613.
Bresler,E. 1978, Analysis of trickle irrigation with application to design problems.
Irrigation Science, 78(1): 3-17.
Brooks, R. H. and A. T. Corey. 1964, Hydraulic properties of porous media.
Hydrology paper 3. Colorado State Univ. Fort Collins. Co.
Bruce, R.R., and Kulte, A.1956, The measurement of soil moisture diffusivity. Soil
Sci. Soc.Am.Proc.20:458-462.
Buckingham, E. 1907, Studies on the movement of soil moisture. Bu11.38, USDA,
Bureau of Soils, Washington, DC.
Carnahan, B. E., H. A. Luther, and j. 0. Wilkes. 1969, Applied numerical methods.
John Wiely & Sons ,New York.
Cassel, D. K., M. Th. Van Genuchten, and P. J. Wierenga.1975. Predicting anion
movement in distributed and undistributed soils. Soil Sci. Soc. Am. Proc. 33:
827- 832.
Chu, S. T. 1994, Green- Ampt analysis of wetting pattern for surface emitters,
Journal and Drainage Engineering, 120,414-421.
Claire M., L Cote Keith, B. Bristow Philip, J. Charlesworth Freeman, J. Cook Peter
and Thorburn. 2003, Analysis of soil wetting and solute transport in subsurface
trickle irrigation. Irrig Sci. 22: 143-156.
Clothier B. E. and R. A. Wooding, 1983, The soil water diffusivity near saturation.
Soil. Sci. SOC.Am. J 47:363-640.
Clothier, B. E., 1984, Solute travel times during trickle irrigation. Water
Resources Res. 20: 1848-1852.
Clothier, B. E,, and D. R Scotter. 1982, Constant -flux infiltration from a
hemispherical cavity, Soil Sci. Soc. Am. Proc., 46: 696-700.
Clothier, B. E., and D. R Scotter. 1982, Two- and three- dimensional infiltration
analysis under field conditions. Soil Sci.45 (3): 765-787.
Clothier, B. E., D. R Scotter and E.R. Harper, 1985 Three dimensional infiltration
and trickle irrigation. Trans Am. Soc. Agric. Engin. 28: 497-501
Colthier, B. E. and T. J Sauer. 1988, Nitrogen transport during drip fertigation with
urea. Soil Sci. Soc. Am. J 52: 345- 349.
Cote CM, Bristow KL, Ford E-J, Verburg K, Keating B, 2001. Measurement of
water and solute movement in large undisturbed soil cores: analysis of
Macknade and Bundaberg data. CSIRO Land and Water, Technical Report
0712001
Darusman Khan AH, Stone LR, Lamm FR. 1997, Water flux below the root zone vs.
drip-line spacing in drip-irrigated corn. Soil Sci. Soc. Am. J. 61 : 1755-1 760.
Dasberg, S., and E. Bresler, 1985, Drip irrigation manual. International Irrigation
Information Center (IIIC) Bet Dagan, Israel Puble 9.
Dasberg. S, Or D, 1999. Drip irrigation. Springer-Verlag
Elmaloglou, St and N. Malamos. 2003, Method to estimate soil- water movement
under a trickle surface line source, with water exraction by roots. Irrig and
Drain 52: 273-284.
Gardenas, A.I., J.W. Hopmans, B.R. Hanson and J. Simunek, 2005. Twodimensional modeling of nitrate leaching for various fertigation scenarios
under micro-irrigation. Agricult. Water Manag. 67, 89-1 04.
Garnder, w. R., and M. S. Mahyugh. 1958, Solution and tests of the diffusion
equation for the movement of water in soil. Soil Sci. Soc. Am. Proc. 22: 197201.
Goldberg, D., Gormat, B., and Y-Bar.1971, The Distribution of Roots, Water and
Minerals as a Result of Trickle Irrigation. Journal of American Society for
Horticultural Science 96(5): 645-648.
Goldberg, D., Rinot, M., and N., karu.1971, Effect of Trickle Irrigation Intervals
on Distribution and Utilization of Soil Moisture in Vineyard. Soil Sci Soc
Am Proc 35(1): 127-130. -192.
Goldberg,D., B. Gotant, and D. Rimon. 1976, Drip irrigation: principles, design, and
agricultural practices. Drip Irrigation Scientific Publications, Kfar shrnaryahu,
Israel, 1 12p.
Grander, W. R. 1958, Some steady state solution of unsaturated moisture flow
equations with application to evaporation from a water table . Soil. Sci. 85:
228- 232.
Gupta R. K., R. P. Ruda, and W.T. Dickinson. 1995, Modeling of saturated regime
as affected by emitter application rates. Micro irrigation for changing world:
conserving resource/Preserving the environment, proceedings of the fifth
international microirrigation congress, ASAE.
Hachum, A.Y., J.F.Alfao, and L.S.Willardson. 1976, Water Movement in Soil from
Trickle Source. ASCE J.Irr. &Drainage.102: 179-192.
Hammami, M., D Hedi. B., Jelloul and M.Mohamed. 2002, Approach for predicting
the wetting front depth beneath a surface point source: Theory and numerical
aspect. Irrig and Drain, 51 :347-360.
Hanks. R. J., and C . L. Ashcrofi. 1980. Applied soil physics. Springer-Verlag.
Berlin. 159 p.
Haq, K. A. 1973, Moisture distribution from trickle irrigation. M.S. Thesis. Univ. of
Hawaii. Honolulu, Hawaii, 56 p.
Haverkamp R, M. Vauclin, J. Touma, P. J. Wierenga and G. Vachaud, 1977. A
comparison of numerical simulation models for one-dimensional infiltration.
Soil Sci. Soc. Am. J. 141: 285-293.
Healy, R. W., and A. W. Warrick, 1988. A generalized solution to infiltration from a
surface point source. Soil Sci. Soc. Am. J 52: 1245-1251.
Hillel, D. 1982, Introduction to Soil Physics. Chapters 1-7. pp 1-134.
Hills, R. G., P. J. Wierenga, D. B. Hudson, and M. R. Kirkland. 1991, The
second Las Cruces trench experiment: Experimental results and twodimensional flow predictions. 27: 2707-2718.
Hossein D, Y Tahei and R Velu, 2004. Assessment of evapotranspiration estimation

models for use in semi-arid environments. Agric Water Manag 64: 91-106.
Howell, T. a., Stevenson, D. S., Aljibury, F. K., Giltin, H. M., Wu, I.P., Warrick, A.
W., and Raats, P. A. C, 1983. Design and operation of trickle (drip) systems,
In: Design and operation of farm irrigation systems, Jensen, M.E. Ed., An
ASAE monograph, American Society of Agricultural Engineers, 2950 Niles
road, St. Joseph, Michigan, pp. 663-713.
James. A. T., Kunkel and D. E. Anderson. 1999. Unsaturated zone hydrology for
scientists and engineers. Prentice-Hall, Inc
Jensen, M. E, 1983 Design and operation of farm irrigation systems, An ASAE
monograph, American Socity of Agricultural Engineers, 2950 Niles road, St.
Joseph, Michigan, pp.829.
Jobling, G.A, 1974. Trickle Irrigation Design.Part 1.Eng.Inst.Lincoln.Canterbwy.
Jury, W. W. R. Grander, and W. H. Grander, 1991. Soil Physics. John Wiley &
Sons, Inc. New York.
Keller, Jack and D. Bliesner,l990. Sprinkler and Trickle Irrigation. van Nostrand
Reinhold, New York.
Kirda, C., 1972. Simultaneous transport of chloride and water during infiltrati.on
andredistribution, Ph.D. Thesis, University of California, Davis, 223 pp.
Kirda, C., D. R. Nielsen and J.W. Biggar, 1973. Simultaneous transport of chloride
and water during infiltration. Soil Sci. Soc. Am. J. 37, 339-345.
Koorevaar, P., G. Menelik, and C. Dirksen. 1983. Elements of soil physics:
Developments in soil science 13. Elsevier Science
Lafolie, F., R. Guennelon and van Genuchten M. Th, 1989. Analysis of water flow
under trickle irrigation 11: Experimental evaluation. Soil Sci. Soc. Am. J. 53:
1318-1332.
Levin, I., P. C. van Rooyer and F. C. van Rooyer, 1979. The effect of discharge rate
and interrnitte water application by point source irrigation on the soil moisture
distribution pattern. Soil Sci.Soc. Am. J. 43: 8-16.
Levin, I., R. Assaf, and B. Braudo, 1974. Soil moisture distribution and depletion in
an apple orchard irrigated by tricklers. Proceeding of 2nd Intern. Drip Irrig.
Cong., PP. 252-257.
Li, J, J.Zhang and L Ren, 2003. Water and nitrogen distribution as affected by
fertigation of ammonium nitrate from a point source. Irrig. Sci 22: 1:12-30.
Li, J., J. Zhang, M. Rao, 2004. Wetting patterns and nitrogen distributions as
affected by fertigation strategies from a surface point source. Agri. Water
Manag. 67: 89-104.
Lockington D, Parlange J Y, Surin A, 1984. Optimal prediction of the saturation and
wetting fronts during trickle irrigation. Soil. Sci. Am. J. 48:488-494.
Mostaghimi, S., J. k. Mitchell, and W. D. Lembke,l98la. Effect of discharge rate on
distribution of moisture in heavy soils irrigated from a trickle source. ASAE
paper no. 81-208 1, ASAE, St. Joseph, MI.
Mostaghimi, S., J. k. MitcheII, and W. D. Lembke, 1981b. The effect of pulsed
trickling on moisture distribution patterns in heavy soils. ASAE paper no. 812553, ASAE, St. Joseph, MI.
Mostaghimi, S., J. k. Mitchell, and W. D. Lembke,l98lb. Soil moisture movement
in cropped soil under trickle irrigation. ASAE paper no. 83- 2026, ASAE, St.
Joseph, MI.
Mualem Y, 1976. A new model for predicting the hydraulic conductivity of
unsaturated porous media. Water Resources Research 12: 5 13-522.
Muluneh, M. and Akbar, A. K, 1995. Field Evaluation of water and solute movement
from a point source. Micro irrigation for changing world: conserving
resource/Preserving the environment, proceedings of the fifth international
microirrigation congress, ASAE. (609-614).
Naylor,T.H, 1970. Computer Simulation Experiment with Models of Economic
system. John Wiley Company.
Neuman, S.P. 1973. Saturated - unsaturated seepage by finite elements, J.Hydro1.
Div. ASCE 99(12): 2233-2250.
Palomo, M. J. F. Moreno, J. E. Fernhndez' , A. Diaz-Espejo and I. F. Giron, 2002.
Determining water consumption in olive orchards using the water balance
approach. Agric. Water Manage. 1:15-35.
Peter. J., Thorburn., J-Freeman ., Cook., L. Keith and Bristow, 2003. Soil-dependent
wetting from trickle emitters: implications for system design and management
1rrig.Sci. 22: 121- 127.
Philip J. R, 1985. Steady absorption from apherodical cavities. Soil Sci Soc Am J
49:828-830.
Philip, J. R, 1957. The theory of infiltration: I. the infiltration equation and its
solution. Soil Sci. 83: 345-357.
Philip, J. R, 1984. Travel times from buried and surface infiltration point source.
Water Resources Res. 20: 990-994.
Philip, J. R, 1986. Linearized unsteady multidimensional
Water.Resour Res 20: 990-994
infiltration.
Philippe. R, Colthier, B. E. Pascal, K, and Vachaud G, 1996. The free- water pond
under trickle source: a field test of existing.Theories. Irrig Sci 16:169-173.
Raats, P. A. C, 1971. Steady infiltration from point sources, cavities and basins. Soil
Sci. Soc. Am. Oroc. 35:689-694.
Ragab, R, 1984.Simulating infiltration into sand fiom a trickle line source using the
matric flux potential concept. Soil Sci 137:120-127.
Rattan, L.,K.M.Shukla.2004. Principles of soil physics. Marcel Dekker, Inc.
Revol, P., B. E. Clothier, G. Vachaud, and J. L. Thony, 1991. Predicting the field
characteristics of drip irrigation. Soil Technol. 4: 125-134.
Risse, L. M., J. L. Chesness, 1989. A simplified design procedure to determine the
wetted radius for a trickle emitter. Transactions of ASAE 32(6), 1909-1914.
Roth, R.L, 1983. Moisture movement from a point source. Ph. D dissertation Univ.
of Arizona, Tuscon AZ.
Roth, R.L., 1974. Soil Moisture Distribution and Wetting Pattern from a Point
Source. Proceedings Second International Drip Irrigation Congress,
Agu. 1974. Sandiego, California, pp.246-25 1.
Russo, D, 1988. Determination soil hydraulic properties by parameter estimation
on the selection of a model for the hydraulic properties. Water Resour. Res.
24: 453-469.
Salem A. Al-Jabei, R. Horton, and Dan B. Jaynes, 2002. A point- surface method for
rapid simulataneous estimation of soil hydraulic and chemical transport
properties.Soi1 Sci Soc. Am. J. 66: 12-18.
Schaap, M.G., F.J. Leij, and M.Th. van Genuchten,2001. ROSETTA: A computer
program for estimating soil hydraulic properties with hierarchical pedotransfer
functions. J. of Hydrology, 25 1, 163-1 76.
Schwartzman, M, B. Zur,1986. Emitter spacing and geometry of wetted soil
volume. J Irrig drainage Engr ASCE 112:242-253.
Sen.H. S,D. Paul, b. K. Bandyopadhyay, and N. B. Dash, 1992. A simple numerical

solution for two-dimensional moisture distribution under trickle irrigation. Soil.
Sci 154:350-356.
Shani, U., R.J. Hanks, E. Bresler, and C.A.S. Oliveria, 1987. Field method of
estimating hydraulic conductivity and matric potential-water content relations.
Soil Sci. Soc. Am. J. 5 1:298-302.
Simunek, J. 1 and M. Th van Genuchten,l994. The Chain-2D code for simulating
two-dimensional variably saturated water flow, heat transport of solutes
involved in sequential first order decay reaction. Versionl.1 Research
Report No. 126, S.S. Salinity Laboratory, USDA, ARS, Riverside, CA.
Simunek, J., M. Sejna and M.Th van Genuchten, 1999. The Hydrus2D Software
Package for Simulating Two-Dimensional Movement of Water, Heat, and
Multiple Solutes in Variable Saturated Media. Version 2.0.IGWMC-TPS53, International Ground Water Modeling Center, Colorado School of
Mines, Golden, Colorado, pp. 1-25 1.
Simunek, J., M. Vogel and M. Th van Genuchten,l992. The SWMS-2D code for
simulating water flow and solute transport in two-dimensional variably
saturated media. Version1 .1 Research Report No. 126, S.S. Salinity
Laboratory, USDA, ARS, Riverside, CA.
Skaggs, T.H., T.J Trout, J. Simunek and P.J.Shouse, 2004. Comparison of Hydrus2D
simulations of drip irrigation with experimental observations. J. Irrig. Drainage
Eng. 130 (4), 304-3 10.
SPSS, 2003. Sigma Plot for Windows. Ver. 11.5. Chicago, Ill.: SPSS, Inc.
Stewart. S, 2003. Diffusion in Biofilms. Journal of Bacteriology. Vol. 185.No 5:
1485-1491
SURFER Version 7,1999. Reference Manual, Golden software Inc., Golden,
Colorado.
Taghavi, S. A., M. A. Marino and D. E. Rolston, 1984. Infiltration from trickle
irrigation source. J. of Irrig. And Drainage Eng., ASCE 11094).
Taylor. S. A and G. L. Ashcraft, 1972 Physical edaphology the physics of irrigated
and nonirrigated soils. W.H.Freeman and company.
van der Poll, R. M., P. J. Wierenga, and D. R. Nielsen, 1977. Solute movement in
lyred soils. Soil Sci. Soc. Am. J. 41 : 10-13.
van Genuchten, M. T, 1980. A closed form equation for predicting the hydraulic
conductivity of unsaturated soils. Soil sci. Soc. Am. J. 44: 892-898.
van Genuchten, M.Th., F.J. Leij, and S.R. Yates, 1991. The RETC code for
quantifying the hydraulic functions of unsaturated soils. U.S. Salinity
Laboratory, Agricultural Research Service, U.S. Dept. of Agriculture,
Riverside, California, USA.
Vogel, T. and M. Cislerova, 1988. On the reliability of unsaturated hydraulic
conductivity calculated from the moisture retention curve. Transport in Porous
Media 3: 1-15.
Warrick, A. W, 1974. Time dependent linearized infiltration, I. Point sources. Soil
Sci. Am. J. 38(3):383-386.
Warrick, A. W., J.W. and D.R. Nielsen. 1971. Simultaneous solute and water transfer
for an unsaturated soil. Water Resour. Res., 7: 1216-1225.
Wooding, R.A, 1968. Steady infiltration from a shallow circular pond. Water
Resour.Res. 4, 1259-1273.
Yitayew M., A. A. Khan, A.W. Warrick, 1998. In situ measurements of soil
hydraulic conductivity using point application of water. Applied Engineering in
Agriculture, 14(2):115-120.
Yitayew, M. and J. Waston., 1986, Field methods for determining unsaturated
hydraulic conductivity. ASAE paper No.86-2570.St.Joseph.
Youngs. E. G. 1961. The hysteresis effect in soil moisture studies. Int. Cong. Soil. Sci.,
Trans. 7 th ( Madison. Wis.) Elsever Publishing Co., Amsterdam.
Youngs, E. G., P. B. Leeds-Harrison; A. Alghusni, 1999. Surface Ponding of Coarsetextured Soils under Irrigation with a Line of Surface Emitters. J. Agric. Engng Res.
73: 95-100.
Zazueta,F.S, G.A.Clark, A. G. Smajstrla and M. Carrillo, 1995. A simple
equation to estimate soil-water movement from a drip irrigation source.
Conserving resource/Preserving the environment, proceedings of the fifth
international microirrigation congress, ASAE. (58 1-856).
Zemel, A., J. Ben -Asher, Ch. Charach, and M. Silberbush, 1987. Solute transfer
and extraction from trickle irrigation source: The effective hemisphere
model. Water Resources. Res. 23 :2091-2096.
Zur. B,1996. Wetted soil volumes as a design objective in trickle irrigation. Irrig Sci
16:lOl-105.
APPENDICES
Figure A.l: Type of emitters used in this study.( On line, turbo-key, button, and non
pressure compensating emitter) a. Laboratory experiments. B. Field
experiments.
Appendix B statistical analysis of the effect of volume of water applied and emitter
discharge rate on surface wetted radius
Table A. 1 : ANOVA of the effect of water volume applied and emitter discharge rate on
surface wetted radius for sandy soil
Source of Variation
SS
Volume of water applied 143.56
df
Ms
F cdcu~ated
P-value
F critical
71.78
3 1.94
0.00
4.10
Discharge rate
Error
Total
182.16
17
Table A.2: ANOVA of the effect of volume of water applied and emitter discharge rate
on surface wetted radius for sandy Loam soil
SS
df
Ms
Discharge rate
0.57
0.28
Error
0.19
0.04
Total
33.14
Source of Variation
F calculated
P-value
F critical
Volume of water
applied
Appendix C
Results of moisture and solute distribution pattern under point source trickle irrigation
(
field
study-comparisons with
predictive
model
and
comparisons with
Hydrus -2D model}
Figure A.2: Water content distribution at r = 2.5 and 17.5cm as a function of vertical
distance. Data are for two different flow rates 2.0 and 2.5 l/h involving 12 1
of applied water
2.5
7.5
12.5
17.5
22.5
Horizontaldistana (cm)
Figure A.3: Water content distribution at z = 7.5 and 27.5 cm as a function of horizontal
of applied water
2.5
12.5
22.5
32.5
42.5
52.5
62.5
Depth (cm)
Figure A.4: Water content distribution at r = 2.5 and 17.5 cm as a function of vertical
distance. Data are for two different flow rates 4.5 and 6.0 I/h involving 34 1 of applied
water
2.5
7.5
12.5
17.5
22.5
Horizontal distance (m)
27.5
25
75
125
175
225
275
Horizontal distance (an)
Figure A.5: Water content distribution at z = 7.5 and 27.5 cm as a function of horizontal
of applied
Figure A.6: Water content distribution at r = 2.5 and 17.5 cm as a function of vertical
of applied water.
2.5
7.5
12.5
17.5
22.5
27.5
Horizontal dirtame (cm)
2.5
7.5
12.5
17.5
22.5
27.5
Figure A.7: Water content distribution at z = 27.5 cm as a function of horizontal

distance. Data are for two different flow rates 2.6 and 3.0 I/h involving 17 1
of applied water.
2.5 7.5 12.5 17.5 22.5 27.5 32.5 37.5
2.5 7.5 12.5 17.5 22.5 27.5 32.5 37.5
Depth (cm)
Depth (cm)
Figure A.8: Chloride relative concentration at r=2.5 and 17.5 cm as a fwnction of vertical
of applied water.
2.5
7.5
12.5
17.5
22.5
Horizontaldistance (cm)
Figure A.9: Chloride relative concentration at z-7.5 and 27.5 cm as a function of

horizontal distance. Data are for two different flow rates 2.0 and 2.5 l/h
involving 12 1of applied water.
0.8
2.5
7.5
12.5
17.5 22.5
27.5
32.5 37.5
Depth (cm)
Figure A.10: Chloride relative concentration at r= 2.5 and 17.5 cm as a function of

vertical distance. Data are for two different flow rates 4.5 and 6.0 Vh
2.5
Horizontal distance (an)
7.5
27.5 crn
12.5
17.5
Holizontaldistance (an)
Figure A. 11: Chloride relative concentration at z= 7.5 and 27.5 cm as a function of

involving 34 1 of applied water
2.5 7.5 12.5 17.5 22.5 27.5 32.5 37.5
Depth (cm)
2.5 7.5 12.5 17.5 22.5 27.5 32.5 37.5
Depth (cm)
Figure A.12: Chloride relative concentration at r= 2.5 and 17.5 cm as a function of

vertical distance. Data are for two different flow rates 2.6 and 3.0 I/h
0
.-
g 0.8
6
.-
0.8
al
8 0.6
g 0.6
.-9
so4
2 .
0.4
m
v
P)
7J
.-
-b 0.2
0.2
0
0
2.5
7.5
12.5
17.5
2.5
22.5
7.5
12.5 17.5 22.5
Horizontaldistance (cm)
- Figure A.13: Chloride relative concentration at z= 7.5 and 27.5 cm as a function of
involving 17 1of applied water
C.l Comparison of simulated resulted as determined using predictive model with
observed data
1
q=3.3 Ilh
2000
4000
Elapsed time (min)
6000
2000
&=0.85 cmlmin
4000
6000
Elapsed time (min)
Figure A.14: Simulation of the surface wetted radius using results from Taghavi et al.,
(1984).
. .
50
q= 2.1 I/h
n
40
I$ = 0.85 cmlh
.
-
fn
30-
73
a,
t:
20-
C
V)
.
0
'
.
0
& = 0.85 cmlh
q= 7.8 11h
9
0
10
ol.
0
'
400
800
Observed
Predicted
- .
1200
1600
100
Elapsed time (min)
35
observed
Predicted
200
300
400
500
Elapsed time (min)
40
q= 9.0Ih ks= 5.8 cmlh
q= 12.30 11h
E 30 -
35
0
w
V)
= 5.8 cmlh
i
0
5 25 -
u
al 20 ..-.
C
a,
15-
'5
V)
100
0'0
Elapsed time (min)
30
Observed
Predicted
'
60
90
120
150
180
Elapsed time (min)
Figure A. 15: Simulation of the surface wetted radius using results from Angelakis et al.,
(1993).
q= 0.90Ilh I$= 1.85 cmlh
lo5
Observed
Predicted
Elapsed time (min)
q= I.40Vh IS= 1.85cmlh
50
100
Elapsed time (min)
l
0
l
0
0
0
O
0
l
q= 4.90llh I$= 1.85cmlh

0
l
20
O
l
'
0
l
300
l
0
250
'
200
Elapsed time (min)

L
150
a _ . . . , . .
40
60
-.
q= 7.80Ilh ks =I .85cmlh
Observed
Predicted
'....I__
80
,.._,
100 120 140
Elapsed time (min)
Elapsed time (min)
Figure A. 16: Simulation of the surface wetted radius using results from Li et al., (2003).
q= I.O llh I$.= 5.8 cmlh
0
0
'
O
0
'
Observed
Predicted
Elapsed time (min)

35
30-
.-2 2 5 -
q= 2.0 1/h
k=5.8 cmlh
2 20-
a,
15
o
0
0
0
10-
5 -
O I . . . . f .
0
100
Observed
Predicted
. . l . . . . l . - . . l . . .
200
300
400
500
Elapsed time (min)
q= 4.0 Ilh I$.= 5.8 crnlh

o
3
.w
2 25 -
15-
10
O
0
0 :
0
0
20t:
o
0
0
0
Observed
Predicted
. - n - - j
0
0
100
200
300
400
500
Elapsed time(min)
Figure A. 17: Simulation of the surface wetted radius using results from Hammami et
al., (2002)
70
q=2.1011h ks= 0.85 cmlh
60
O
t
50
0)
C
.t:
3 40
o
8s 30
Y-
m
>
n
mm
20
.-0
>
10
0
0
1000
2000
3000
4000
5000
Elapsed time (min)
q= 3.30 11h ks= 0.85 cmlh
.
0
- 8
0
B
0
1000
Observed
Predicted
2000
3000
4000
Elapsed time (min)
Figure A. 18: Simulation of the vertical advance of wetting front using results from
Taghavi et al., (1984)
50
0
--5
q=2.10Ilh
40 -
k= 0.85cmlh
"-
600
Observed
Predicted
1200
1800
Elapsed time (min)
Elapsed time (min)
q=12.30Ilh ks= 5.8 cmlh
2 40 C
0)
.-C
S
30
Observed
Predicted
0
Elapsed time (min)
30
60
90
120
150
180
Elapsed time (min)
Angelakis et al., (1993).
q= 0.60 Ilh & = 1 85 cmlh
&= 1.85 cmlh
q= 0 90 11h
0
0
0
0
0
0
0
0
0
0
0
i . . .
200
l .
Observed
Predicted
400
. . . . . '
600
800
1000
100 200 300 400 500 600 700
Elapsed time (min)
Elapsed time (min)
.
0
&= 1 85 cmlh
q= 1.40 Ilh
Observed
Predicted
-q= 2.011h
k= 1.85 cmlh
0
0
0
0
0
0
0
0
Observed
Predicted
50
100
Elapsed time (min)

3
.-5
a,
0
30 -
q= 4.90 Ilh
&= 1 85 cmlh
q= 7.8 Ilh
&= 1.85
0
0
15 -
0
0
30
60
Obsreved
Predicted
0 ) 0
0
0
20-
300
0
0
5m l o - :
m
250
Elapsed time (min)
25 -
.o
r
200
Y-
150
90
Elapsed time (min)
120
:
150
Observed
Predicted
...
.
-..'....'.-..'. 0
20 40 60 80 100 120 140
. I
- I
Elapsed time (min)
Figure A. 20: Simulation of the vertical advance of wetting front using results fiom Li et
al., (2003).
40
35
-.
25 .
20
1510
. . *
q= 1.0 llh k,= 5.8 cmlh
30
0
0
Observed
Predicted
100
200
'
300
'. .-
400
' ~. .~- .- '' . ~
500
600
700
800
700
800
Elapsed time (mim)
Observed
Predicted
100
200
300
400
500
600
Elapsed time (min)

60
So
- q= 4.0 Ilh k ,= 5.8 cmlh
40 :
30 -
20-
Observed
Predicted
10 O L .
--'...-'.--.'....'.
100
200
300
400
'
500
600
700
800
Elapsed time (min)
Harnrnami et al., (2002).

@=0.97
- - - Fitted line y= 1.10 x

- 111 line
1
J
10
15
20
25
30
35
10
15
20
25
Observed surface wetted radius(cm)
10

ne y = 0.976 x F?=0.99
15
20
25
30
Figure A. 22: Correlation of observed and predicted surface wetted radius for river sand
soil under 3, 5.5 and 7 l/h application rates
10
12
14
16
. . .
0 Predicted
------- Fitted line y= 0 . 8 7 ~R?=0.987
8 10 12 14 16 18 20 22 24 26 28
Observed vertical advance of wetting front (cm)
Figure A. 23: Correlation of observed and predicted vertical advance of wetting front for
various application rates for fine sandy loam Figure A. 22: Correlation of
observed and predicted surface wetted radius for various application rates
for fine sandy loam
10
15
20
25
30
25
-0
10
15
20
25
Observed-Predicted data
10
15
20
25
30
Observed surface wetted radius(cm)
30
10
15
20
25
30
Figure A. 24: Correlation of observed and predicted surface wetted radius under 0.5,0.7,
1.O, 1.4 and 2 l k application rates for sandy soil.
25
10
15
20
25
30
35
10
15
20
25
30
Observed vertical advance of wetting

front (cm)
Observed Vertical advance of wetting

front (cm)

front (crn)
Observed Vertical advance of wetting

front (cm)
35
q= 2.0 Ilh
F~ttedline y= 0 98 x R
= 6 94

front (cm)
sandy soil under application rate of 0.5,0.7, 1.O, 1.4 and 2 Ih.
15 -
q= 2.0 Ilh loam sand

I
10
. .
15
20
. .
25
o Observed- Predicted data
------- Fitted line y= 0,892 x F? = 0.99
q= 3.0 I/h Loamy sand
10
15
20
25
I-
0 Observed- Predicted data
20
~i~~
ed line y= 0.913 I? = 0.986
I:I line
U-.
loamy sand soil under application rate of 2.0,3.0 and 4.0 l/h
- - - Fitted line y= 1 038 x
@ = 0.98
q= 2.0 Ilh Loam
10
15
20
25
30
35
40
40
- . . I

- - - fitted line x= 0 941 x = 0.994
35
q= 3.0 Ilh Loam
10
15
20
25
30
35
40
40
C
a,
- - -
35 - o Observed- Predicted data

30 . ------ Fitted line y= 0.9533 x F?= 0.98
q= 4.011h Loam
10
15
20
25
30
35
40
Figure A. 27: Correlation of observed and predicted vertical advance of wetting fiont for
loamy soil under application rate of 2.0,3.0 and 4.0 l/h

------- Fitted lrne y= 0.927 x F?= 0.947
q= 2.0 Ilh Silty clay
Observed surface wetted radius (cm))

45
----.Fitted lme y= 0.914
I%0.94
q= 4.0 11h Silty clay
10
15
20
25
30
35
40
45
silty clay soil under application rate of 2.0,3.0 and 4.0 l/h
10
15
20
25
30
loamy soil under application rate of 2.0,3.0 and 4.0 l/h
10
15
20
25
30
I0
15
20
25
30
Figure A. 30: Correlation of observed and predicted surface wetted radius for sandy soil
10
15
20
25
30
10
15
20
25
30
35
Figure A. 3 1: Correlation of observed and predicted surface wetted radius for sandy soil
under application rates of 2.5, 2.7,3.5 and 4.5 l/h
10
15 20 25
30 35 40
10 15 20 25
30 35 40
Figure A. 32: Correlation of observed and predicted surface wetted radius for sandy soil
under application rates of 4.8 and 6.0 l k
60
q=2.0 I/h
50
Predicted
-4-.--Fitted line
-0
--
8c h
60
Predicted
Fitted line
- 1:l Line
1:l Line
Observed Vertical advance of

wetting front (cm)
-------
10
20
30
40
50
60
70

wetting front (cm)
sandy soil under application rates of 2.0 and 4.5 l/h
Y-
70
q=4.8 I/h
Predicted

wetting front (cm)
Observed Vertical advance

of wetting front (cm)
Figure A. 34: Correlation of observed and predicted vertical advance of wetting front
for sandy soil under application rates of 4.8 and 6.0 l/h
y= 0.949 x R2= 0.93
- 1:l Line
0
10
15
20
25
30
10
15
20
25
30
35
Figure A -35:Correlation of observed and surface wetted radius for sandy loam under
application rates of 2.3 and 2.6 l/h
10
15
20
25
30
35
Predicted
10
20
15
25
30
35
Figure A .36: Correlation of observed and surface wetted radius for sandy loam under
application rates of 3.0 and 4.0 l h
50
Observed vertical advance of

wetting front (crn)
40
30
Predicted
------- Fitted line

1:l tine

wetting front (crn)
Figure A.37: Correlation of observed and predicted vertical advance of wetting front for
sandy loam under application rates of 2.3 an 2.6 l/h
q= 3.0Wh
0 Predicted
- - - Fitted line

wetting front (cm)
q= 4.0 Ilh
0
Predicted

wetting front (cm)
sandy loam under application rates of 3.0 and 4.0 I/h
Depth = 37.5 cm
0.00
10
20
30
40
Figure A.39: Horizontal moisture distribution for sandy soil under application rate of 1.5
l/h at z= 7.5, 17.5,27.5 and 37.5 em, (-) Hydrus -2D model , elapsed time 320 min
MBE=0.010
RMSE=0.027
U Theil' s =0.058
f.
0
U
0.15
0.10
0.05
- Depth = 17.5 c m
0.00
0
I0
20
30
40

0.30
MBE=-0.0089
l/h at z= 7.5, 17.5,27.5 and 37.5 cm, (-) Hydrus -2D model, elapsed time 360 min
0.35 -
MBE= 0.01 1
RMSE= 0.0265
U Theil's= 0.054
Q)
C
a,
C
0.10 0.05
Depth = 17 5 cm
0.00
0
20
40
60
20
40
60
(cm)
l/h at z= 7.5, 17.5,27.5 and 37.5 cm, (-) Hydrus -2D model; elapsed time 290 min
Horizontal distance
MBE=0.0274
RMSE=0.038
U Theil's =0.075
0.1 0 0.05 0.00
Depth= 17.5 cm
20
40
60
l/h at z= 7.5, 17.5,27.5 and 37.5 cm, (-1 Hydrus -2D model; elapsed time 420 min
20
40
60
orizontal distance [cm)
l/h at z= 7.5, 17.5,27.5 and 37.5 cm, (-1 Hydrus -2D model; elapsed time 460 min
Horizontal distanc
0.4 0
0.3 5
RMSE= 0.049
U Theil's = 0.101
0.30
0.25
a,
C
:: 0.20
L
a,
0.1 5
0.1 0
0.05
0.0 0
0 40
+
0.30
0.20
MBE = 0.0089
RMSE= 0.017
U Theil's= 0.040
Q)
C
00
L
Q)
2 0.10 C
Distance = 2 5 cm
I
0.00
0
10
20
30
40
50
60
TO
80
Depth (cm)
0.40
MBE =- 0.0004
RMSE= 0.024
U Theil's= 0.076
E 030m
'
I
Q)
.#-a
OAO -
Distance = 12.5 cm
I
0.DO
0
10
20
30
40
50
60
70
80
Depth (cm)
0.40
0.30
0.20
-9.
MBE =- 0.0002
RMSE= 0.022
U Thei17s=0.092
Q)
.#-a
'
Q)
.#-a
0.10 0.00
0
Distance = 17.5 cm
I
10
20
30
40
50
60
70
80
Depth (cm)
Figure A.47: Vertical moisture distribution for sandy soil under application rate of 1.5
I/h at r= 2.5, 12.5 and 17.5 cm, (-) Hydrus -2D model; elapsed time 320 min
10
20
30
40
50
60
70
Depth (cm)
0.40 0.35 E 0.30-
'
I
-, 0.25 -/
Q)
0.20
-- 0.15
Q)
0.05 0.1 0
MBE= 0.0033
RMSE= 0.020
U Theil's= 0.049
Distance= 12.5 crn

I
0.00
0
10
20
30
40
50
60
70
Depth (cm)
0.40
0.35
E 0.30
Q)
--,
0)
0.25 -
MBE= 0.022
RMSE= 0.037
U Theil's= 0.123
0 20 -/
=
m
m
I
015-
-z 0.10 0.05
0.00
Distance = 17.5 c m
10
20
30
40
50
60
70
Depth (cm)
l/h at r= 2.5, 12.5 and 17.5 cm, (-) Hydrus -2D model; elapsed time 360 min
D E = 0.0088
RMSE= 0.020
U Theil's = 0.0452
- Distance = 2.5 cm
l/h at r= 2.5, 12.5 and 17.5 cm, (-) Hydrus -2D model; elapsed time 290 min
0~05 isl lance = 17.5 crn
l/h at r= 2.5, 12.5, 17.5 and 22.5 cm, (-1 Hydrus -2D model; elapsed time 420 min
Figure AS1 : Vertical moisture distribution for sandy soil under application rate of
3.5 l/h at r= 2.5, 12.5, 17.5 and 22.5 em, (-) Hydrus -2D model; elapsed time 420 min
20
Dept
Figure A.52: Vertical moisture distribution for sandy soil under application rate of
4.5 I/h at r= 2.5, 12.5, 17.5 and 22.5 cm, (-1 Hydrus -2D model; elapsed time 460 min
Depth (cm)
0.35
MBE = 0.0 13
RMSE= 0.022
U Theil's = 0.045
20
40
Depth (cm)
60
80
20
40
60
80
Depth (cm)
4.8 l/h at r= 2.5, 12.5, 17.5 and 22.5 cm, (-) Hydrus -2D model; elapsed time 370 min
20
40
60
100
80
Depth (cm)
0.40
MBE = 0.002 19
RMSE= 0.066
U Theil's = 0.142
0.35
0.30
0.25
0.20
MBE = -0.005
RMSE= 0.05 1
U Theil's = 0.174
Q)
I
I
Q)
0s
$ 0.1 5 -
0.10
-I
0.05
Distance = 17.5
0.05 -
Disatnce = 27.5 c m
0.00
0
20
40
Depth (cm)
0.00
20
40
60
80
100
Depth (cm)
6.0 l/h at r= 2.5, 12.5, 17.5 and 22.5 cm, (-) Hydrus -2D model; elapsed time 340 min
0 50
+
a,
0.40
0.30
& 0.20 0.1 0
MBE= 0.0008
RMSE= 0.019
U Theil7s=0.028
0.00
Depth= 7.5 cm
10
20
40
30
50
Horizontal distance (c m>
0.00
I
0
15
10
25
20
30
35
-&
5
0.40 0.35 0.30 0.25 0.20 a.l ,5 0'100.05 0.00
MBE=- 0.056
RMSE= 0.074
UTheil's=O.l77
Depth = 27.5 cm
I
m
10
15
20
25
30
35
Figure A.55: Horizontal moisture distribution for sandy loam soil under application rate
of 2.3 l/h at z= 7.5, 12.5 and 27.5 cm, (-) Hydms -2D model; elapsed time 270 min
0
0.50
0.25
M B E - 0.001
RMSE= 0.037
U Theil's= 0.062
0.40 -
30
40
0.20
0.1 5
.
I
MBE= -0.049
RMSE= 0.061
U Theil's= 0.247
0.30 +
a
.
-$
20
0.45 -
0.35
10
0.25
U
L
P)
Q)
0.20
0.15
g 0,o
0.05 -
Depth= 37.5 cm
0.00
0
I0
20
Horizontal distance
30
(cm)
4C
10
20
30
40
H orizontal distance (cm )
of 2.6 l/h at z= 7.5, 17.5,27.5 and 37.5 cm, (-) Hydrus -2D model; elapsed time 400
min
of 3.0 l/h at z= 7.5, 12.5, and 27.5 cm, (-1 Hydrus -2D model; elapsed time 350 min
MBE= 0.0208
U Theil's= 0.038
10
20
30
40
of 4.0 I/h at z= 7.5, 17.5,27.5 and 37.5 cm, (-) Hydrus -2D model; elapsed time 360
min
.
20
Distance = 2.5 cm
30
Depth (cm)
Distance = 12.5 cm
20
30
Depth (cm)
Distance = 17.5 cm
MBE= -0.016
RMSE= 0.033
U Theil's= 0.081
20
30
Depth (cm)
Figure A.59: Vertical moisture distribution for sandy loam soil under application rate of
2.3 l/h at r = 2.5, 12.5and 17.5 cm, (-) Hydrus -2D model; elapsed time 270 min
Depth [cm)
2.6 l/h at r = 2.5, 12.5, 17.5 and 22.5 cm, (-) Hydrus -2D model; elapsed time 400 min
0.60
MAE ~0.0046
RMSE= 0.0370
U= 0.0570
0.50 E 0.40
a,
c
0
-& 0.30 -
0.20
0.10
Distance = 12.50 cm
0.00
Depth [cm)
20
40
Depth (cm)
60
MAE = 0.0008
RMSE= 0.0399
U= 0.0769
Depth (cm)
3.0 llhat r=2.5, 12.5, 17.5 and22.5 cm, (-) Hydms -2D model; elapsed time 350min
Deeth (c m)
Depth [cm)
Distance = 22.5 cm
0. oo
0
20
40
60
Depth (cm)
4.0 l k at r = 2.5, 12.5, 17.5 and 22.5 cm, (-) Hydrus -2D model; elapsed time 360 min
60 -
--.
50
E
5
Depth = 7.50 cm
.5
- 40-
MBE = 0.20
RMSE = 9.58
U Theills= 0.128
30-
[I)
E0
20-
2
L
0
-
Ir
10 0
10
15
20
25
30
35
r
40

1
E"
-c
22!
60
50 -
MBE =- 2.35
RMSE =5.16
U Theil1s=0.067
40-
30U
5
20U
.4
- 10-
Depth = 17.50 cm
10
15
20
25
30
35
40
rn
Depth = 27.50 cm
MBE = -5.79
RMSE = 10.33
U Theil1s=0.15
10
15
20
25
30
35
Figure A.63: Horizontal chloride distribution for sandy soil under application rate of 1.5
lk at z = 7.5, 17.5 and 27.5 cm, (-1 Hydrus -2D model; elapsed time 320 min
3
\
:I0 -
8 -
MBE = -0.98
RMSE = 2.70
U Theit's= 0.18
(I)
20
6 -
0
(I)
PL
O
c
4 2 8
8
I
1
l/h at z = 7.5, 17.5,27.5 and 37.5 cm, (-1 Hydrus -2D model; elapsed time 360 min
20
30
Horizontal distance [cm]
35
10
40
Depth = 37.50 cm
l/h at z = 7.5, 17.5,27.5 and 37.5 cm, (-) Hydrus -2D model; elapsed time 290 min
1
l h at z = 7.5, 17.5,27.5 and 37.5 cm, (-1 Hydrus -2D model; elapsed time 420 min
Figure A.67: Horizontal chloride distribution for sandy soil under application rate of
3.5 l/h at z = 7.5, 17.5,27.5 and 37.5 cm, (-) Hydrus -2D model; elaped time 420 min
60
Depth = 17.50 cm
50
-C
23
'--
I=
40-
0
.F
+a
30-
'z
P
U
.4
- 200
s
0
L
MBE = 0.24
RMSE = 12.52
U Theilk= 0.134
10
20
30
40
Horizontal distance (crn)
50
60
Depth = 27.50 c m
=
rn
Depth = 37.50 c m
I
50 -
-.
c
zs
40'--
-$ 3 0 s
s
w
s
w
20-
m
.L
O
-
10 0
10
20
30
40
MBE = -1.25
RMSE = 10.79
U Theirs= 0.124
rn
I
50
4.5 l/h at z = 7.5, 17.5,27.5 and 37.5 cm, (-) Hydrus -2D model; elapsed time 460 min
40
Depth = 17.50 cn
Depth = 7.50 cm
35 C
\
30 -
5
.C
- 25 F
20 -
15(I)
.s-Q
--D
MBE = -0 60
RMSE = 4.49
U Theilk= 0.076
2z
I'
MBE = -0.53
RMSE = 8.33
U Theirs= 0.137
5 -
10 -
Horizontal distance (em)
40
35-.
.. .
Depth = 27.50 cm
7D3 0 IU
-25
c
.C
-0
z"s
20 -
15 -
al
a,
.-0
-
- 10
2z
MBE = -1.09
RMSE = 7 00
U Theil's= 0 120
5 0
10
20 30 40
50
4.8 l/h at z = 7.5, 17.5,27.5 and 37.5 cm, (-) Hydrus -2D model; elapsed time 370 min
Depth = 7.50 c m
MBE = 0.53
RMSE = 3.39
U Theil1s=0 101
7
I
10
20
30
40
50
Depth = 27.50 c m
Depth = 37.50 c m
-....
MBE = -1.24
RMSE = 2.33
U Theirs= 0.069
MBE = 1.08
RMSE = 4.80
U Theirs= 0.14
I
10
20
30
'c
40
50
1
6.0 Ik at z = 7.5, 17.5,27.5 and 37.5 cm, (-) Hydrus -2D model; elapsed time 340 min
60
Distance = 2.50 cm
Distance = 12.50 cm
rn
= 50
\
-=
rn
40 -
30 -
0
U
I
(I)
MBE = -3.64
RMSE = 5.71
U Theilk= 0.08
0
L
20
0
r
MBE = -3 91
RMSE = 6.93.
U Theilk= 0.1 1
10 -
rn
20
40
---
60
80
Depth (cm)
Figure A.71: Vertical chloride distribution for sandy soil under application rate of
---
12 -
0a,
E 10-
!z
.0
-
II!
8 -
I=
a,
0
U
a,
-0
.r
0
II
6 -
MBE = -1 25
RMSE = 1.55
U Thell's= 0.085
II
Distance = 12.50 cm
I
0
0
10
20
30
- -
40
50
60
Depth (cm)
14
12 -
---.
=
10 -
Distance = 22.50 cm
10 20 30 40
Depth (cm)
50 60
2.0 l/h at r = 2.5, 12.5, 17.5 and 22.5 cm, (-1 Hydrus -2D model; elapsed time 360 min
Depth (cm)
Depth [cm)
Finure A.73: Vertical chloride distribution for sandy soil under application rate of
2.5 l/h at r = 2.5, 12.5, 17.5 and 22.5 crn, (-) ~ ~ d-2D
r krnodel;elapsed time 290 min
stance = 22.50 cm
RMSE = 1.82
U Theil's= 0.130
40
60
80
Depth [em)
25 Distance = 2.50 cm
20 -
--.
c
E 15
-s
.-0
F
C
g 10s
0
U
Q)
& 5 -
MBE = -0.68
RMSE = 1.70
U The~l's=0.058
r
0
20
40
60
Depth [cm)
80
Depth = 17.50 cm
.=.
MBE = -1.31
RMSE = 3 79
U Theirs= 0.12
Depth [cm)
60
=.I.
c
\
.-
Distance = 12 50 cm
Distance = 2.50 cm
.
I
.
.
50 -
.-0
40
30-
O
r
L
MBE = -1.20
RMSE = 8.38
U Theil1s=0.095
20MBE = -6.84
RMSE = 10.53
U Theirs= 0.12
10
20
40
60
80
1C
20
40
60
60
100
60
. . ..
--. .
Distance = 22.50 cm
Depth = 17.50 cm
I.
.I
50 -
50-m,'
--.
c
C--
40-
- 40-
s
'2
30
r
w
$ 30C
a,
=1
20al
7b2
0
-
80
Depth (cm)
Depth (cm)
100
s
MBE = -5.04
RMSE = 7.52
U Theilk= 0.098
20
40
60
Depth (cm)
20-
IU
0
L
O
-
r
0
1.1
I
80
10 0
100
MBE = -7.79
RMSE = 9.04
U The~l's=0.123
.
--
20
40
60
Depth (cm)
80
100
4.5 l k at r = 2.5, 12.5, 17.5 and 22.5 cm, (-1 Hydrus -2D model; elapsed time 460min
Distance = 2.50 cm
Distance = 12.50 cm
1 1-
a,
'=
E
s
m
0
c
0
0
15
0
-
-c
0
.L
MB: =-*.*:
RMSE = 3.64
U The~l's=0.067
20 -
MBE=-1.49
RMSE = 3.93
- U Theirs= 0.073
0
0
20
40
60
.A.
80
Depth (cm)
40
Distance = 22.50 cm
Depth = 17.50 cm
I
c
\
30 -
rn
I
s
s
2
-
15-
0
0
10 -
MBE = 0.40
RMSE = 6.78
U Theirs= 0.130
O
r
5 0
20
40 60
Depth (cm)
80
MBE=-3.38
RMSE = 4.49
U Theilk= 0.09
10[
4.8 I/h at r = 2.5, 12.5, 17.5 and 22.5 cm, (-) Hydrus -2D model; elapsed time 370 min
-.
Distance = 2.50 cm
Distance = 12.50 cm
.I.
MBE = -1.54
RMSE = 2.60
U Theirs= 0.07
20
40
60
MBE = -1.66
RMSEz216
U Theilk= 0 067
I.
80
100
20
Depth [cm)
40 60
Depth [cm)
80
100
25
Depth = 22.50 cm
20 c
\
U
-:
...
I
.I.-
15s
0
.-
-F
C
s
a,
u
!z
10 -
0
U
a,
=
L
AZ
0
0
Depth [cm)
MBE = -2.43
RMSE = 2.89
U Theirs= 0.109
I
20
40
60
Depth [em)
..
80
100
5
z
.-0
2
35
30 2520 1510-
Depth = 7.50 c m
MBE = 0.44
RMSE = 5.53
U Theil1s=0.1 1
:5 -
-0
0 ,
0
-
i f 0
s
r
g
.D
G
L
2
u
g
0
cu
.--0
35 30252015105 0
0
10
rn
20
30
40
50
Depth = 17.50 c m
MBE = -0.58
RMSE = 3.25
U Theilk= 0.07
10
20
30
40
50
Ezs- 35
-E6
30-
.+
- 2520+
s
15 s
10 .4
- 5 0
r
0
0
0
Depth= 27.50 c m
MBE = 1.87
RMSE = 4.73
U Theilk= 0.154
10
20
30
40
50
Figure A.79: Horizontal chloride distribution for sandy loam soil under application rate
of 2.3 l/h at z = 7.5, 17.5 and 27.5 cm, (-) Hydrus -2D model; elapsed time 270 min
e
E
30-
25-
E
.-
20-
0-
z 15C
c
Q)
'
I
Q)
-cl
&
-
105 -
Depth = 7.50 c m
rn
MBE = 0.74
RMSE = 4.61
U Theills= 0.11
A=
10
15
20
25
30
35
= 30
\
Cr
5
.-
25 20 -
15-
Q)
2 100
'
0
I
50
E
A=
o
0
0
Depth = 17.50 c m
MBE = 0.01
RMSE = 3.98
U Theilk= 0.1
10
15
20
25
30
35
Figure A.8 1: Horizontal chloride distribution for sandy loam soil under application rate
of 4.0 l h at z = 7.5, 17.5, 27.5 and 37.5 cm, (-1 Hydms -2D model; elapsed time 360
min
1
Depth (cm)
35
Distance = 17.50 cm
30 -
=
\
MBE=0.187
RMSE = 3.142
25
30
Distance = 22.50 cm
MBE = 2.93
RMSE = 5.04
'c-.
.E
F
-.-.
c
2015 -
0
U
.-L
10 -
r
O
5 0 -0
10
20
30
Depth [cm)
40
50
10
20
30
Depth [cm)
40
50
Figure A.83: Vertical chloride distribution for sandy loam soil under application rate of
2.3 l/h at r = 2.5, 12.5, 17.5 and 22.5 cm, (-) Hydms -2D model; elapsed time 270 min
Distance = 2.50 cm
Distance = 12.50 crn
rn
MBE = 0.20
RMSE = 3.89
U Theills= 0.1 1
.
8
10
20
30
40
50
Depth (cm)
10
20
30
40
50
Depth (cm)
30
Distance = 17.50 cm
25 -
--.
m
16 -
20 --
Es
10-
a,
D
L
r
0
rn
a,
0
-
15 -
14-
E 12 -s
.0
F 10-
s
0
.-
rn
--%
5-
MBE = 0.08
RMSE = 4.25
U Theills= 0.128
a,
P_
&
c
rn
rn
8 -
6 -
U TheilPs=0.162
rn
10
20 30
Depth (cm)
40
50
4 - MBE = 2.62
- RMSE = 3.73
I
10
20
30
40
50
Depth (cm)
Distance = 2.50 cm
Distance = 12 50 cm
I
I
MBE = -1.44
RMSE = 4.10
U Theills= 0.05
- MBE = -1.28
RMSE = 2.70
- U Theil1s=0.40
IN
I=
10 20
30
40 50
60
Depth (cm)
Depth [cm)
50
45
Distance = 17.50 cm
Distance = 22.50 cm
45 P
40 -
--.
..
MBE = 1.93
RMSE=6.53
c
a,
20
MBE = -0.97
RMSE = 3.94
U Theit's= 0.06
a,
a,
0
L
0
20 15 -
5 -
0
Depth (cm)
10 -
10 20 30 40
50 60
Depth (em)
1 0 20
30
40 50 60
Depth {cm)
16
Distance = 22.50 cm
Depth [em)
BIODATA OF THE AUTHOR
Ahmed Ibrahim Mohammed Ekhmaj was born in Tripoli, Libya on 24th April 1970.
He obtained his Bachelor of Science (B.Sc.) in Water Science, Department of Soil and
Water, from Faculty of Agriculture, Al- Fateh University, Libya in (199311994). He
was appointed as a demonstrator in the Soil and Water Department, Faculty of
Agriculture, Al-Fatheh University. He received his M.Sc. degree in Water Science
from the same mentioned University in 1998. Upon his graduation he was appointed
as a teaching staff in the Soil and Water Department in the Faculty of Agriculture. In
2001, he was granted a scholarship from Al-Fateh University and enrolled in the
Department of Biological and Agricultural Engineering, Faculty of Engineering,
University Putra Malaysia to pursue his Ph.D Degree.
List of Publications
1. Ekhmaj A.I., M. S. M. Amin, S. Salim and A. A. Zakaria. Wetting fronts and

Moisture Distribution under a Point Source 2003. Advanced Technology
Congress, Bio-Engineering Conference Smart Farming 111. May 20-21 2003
Putrajaya, Malaysia.
2. Ekhmaj A.I., and M.S.M. Amin, 2004. Surface Wetted Area under Point
Source Trickle Irrigation in Sandy Soil, lStASIAN CONFERENCE ON
PRECISION AGRICULTURE. 1 1- 13 May , The Legend Hotel, Kuala
Lumpur. Malaysia
3. Ekhmaj A.I., M. S. M. Arnin and A. A. Zakaria, 2004. Wetting Fronts

Advances under a Point Source Trickle Irrigation in Two Types of Soil , CIGR
International Conference Session 1 on Land and Water Use The 7th InterRegional Conference on Environment-Water, 11-14 October Beijing ,China.
4. Ekhmaj A.I., M. S. M. Amin, and A. A. Zakaria, 2004. Wetting front
movement under trickle irrigation: Field experiments and modeling, 2004,
International Conference on Water Resources and Arid Environment 5 - 8
December Riyadh, KSA.
5. Ekhmaj A.I., M. S. M. Amin, S. Salim and A. A. Zakaria, 2004. Simple

Aprroximations for Determining the Geometry of Soil Wetting Pattern under
Point Source Trickle Irrigation. Agriculture Congress 2004.Malaysia
International exhibition and Convention Center (MIECC) 4-7 October,
Selangor, Malaysia.
6. Amin M. S. M., A.I. Ekhrnaj, S. Salim and Asep Sapei 2005. TIDE-Trickle
Irrigation Water Distribution Pattern Estimator, 2005.Exibition of Invention,
Research &Innovation UPM 2005.
7. Ekhmaj A.I., M.S.M. Amin, S. Salim and A.A. Zakaria. Estimation of
Saturated Hydraulic Conductivity using Emitter Discharge Rate and Saturated
Radius, 2005. Soil Science Conference of Malaysia 2005 (SOILS 2005) Sg.
Petani, Kedah 12-13 Apr, Malaysia.
8. Ekhmaj A.I., M. S. M .Amin, S.Salim and A.A.Zakaria, 2005. Wetted Surface

Radius under Point-Source Trickle irrigation in Sandy Soil. International
Agricultural Engineering Journal 2005, 14(2):67-75
9. Ekhmaj A.I., M. S. M .Amin, S.Salim and A.A.Zakaria, 2005. Wetting fronts

advances under a point source trickle irrigation in two types of soil. Journal of
Applied Irrigation Science. (Accepted).
10. .Amin, M.S.M and A.I.Ekhmaj, 2005. Dipac: Drip Irrigation water
Distribution Pattern calculator. Paper No. 1278, 26th Annual International
Irrigation Show, Phoenix Arizona USA
11. Ekhmaj, A.I. and MIS.M. Amin, 2005. Simulation of Soil Water Movement in
Sandy loam Soil under Field Conditions with Hydrus-2D Model. International
Advanced technology Conferences 2005. 6-8 December 2005. Putrajaya,
Malaysia.

Ahmed Ekhmaj

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Ahmed Ekhmaj

Diunggah oleh

Hak Cipta:

Format Tersedia

UNIVERSITI PUTRA MALAYSIA

WATER AND SOLUTE DISTRIBUTION PATTERN IN SOIL UNDER

AHMED I.M. EKHMAJ.

Abstract of thesis presented to the Senate of Universiti Putra Malaysia in fulfillment of

WATER AND SOLUTE DISTRIBUTION PATTERN IN SOIL UNDER POINT

A basic requirement in the design of a trickle irrigation system is to obtain more

PtWUS I'AKAAN SULTAN A E U L SAMAD

BENTUK PERTABURAN AIR DAN SOLUT DALAM TANAH DIBAWAH

Pengurusi: Professor Mohd Amin Mohd Soom, PhD, P.Eng

I am especially grateful to my mother, a person whose courage, fortitude and patience I

Ir. Desa Ahmad, PhD

School of Graduate Studies

Mohd Amin Mohd Soom, PhD

AINI IDERIS, PhD

3.2.2 Wooden Box Experiments

RESULTS AND DISCUSSIONS

Summary differences between saturated and unsaturated flow 2.6

Parameters for estimation of surface wetted radius (Wooden Box 4.17

Coefficients and diagnostic tests for nonlinear regression model 4.70

Values of the statistical parameters used in comparison for 4.100

Values of the statistical parameters used in comparison for solute 4.1 67

General elevation for unsaturated conductivity (K (6)) versus 2.7

under trickle discharge rates of 1.0 and 3.0 l/h

Sensitivity analysis of the vertical advance of wetting front to the 4.82

under application rate of 0.5, 0.7, 1.O, 1.4 and 2 l h .

rate of 2.0 I/h for sandy soil, elapsed time 360min

rate of 4.5 l k for sandy soil

CDE= Convection-Dispersion Equation

RETC= Retention Curve

Statement of The Problem

Objectives of the Study

Scope of the study

laboratory and field

The signficance of this research can be summarized as follow:

Organization of the Thesis

Water Flow Mechanism

2.1.2 Saturated Water Flow

A distinction is made between saturated and unsaturated hydraulic conductivity.

100. Saturated hydraulic conductivity cannot be used to describe water

2.1.3 Unsaturated Water Flow

Variable over space and time

Zero (close to zero)

Variable over space and time

Positive and constant

Negative and variable

Low and variable

Possible provided temperature

Steady as well as unsteady

Inflow =outflow+ source or

Edgar Buckingham (1907) suggested a modification of Darcy's law in order to describe

2. The unsaturated hydraulic conductivity of soil is a function of water content

2.1.4 Continuity Equation

~ ) t is the time (T)

where S is an extraction term.

The 0- based equation can be written as

Water Movement under Trickle Irrigation

2.2.1 Field Studies

Bulk density (gm/cm3)

2.2.2 Laboratory Studies

2.2.3 Mathematical Model Studies

Schwartzman and Zur (1986) 'developed a semi-empirical approach to predict the

0,,- 8, ,where: 8,, is average water content, and 8, is

Healy and Warrick (1988) developed numerical solutions in a dimensionless form of

a= a constant expressing the relative importance of gravity and capillarity to