Anda di halaman 1dari 22

Engineering Geology: Fundamental Input or Random Variable?

Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

Jeffrey R. Keaton1, P.E., P.G., F.ASCE

Geotechnical Practice Leader, AMEC, 6001 Rickenbacker Road, Los Angeles, CA 90040;

ABSTRACT: Geologists and engineers view the world in complementary but different ways.
Science seeks to explain all observed details, whereas engineering seeks to design with specific
objectives and multiple constraints. National guidance in the United States calls for geotechnical
site investigations to be performed by geotechnical engineers and engineering geologists. Site
characterization should start with Geologic Models which form the basis for Ground Models
(Geologic Models with engineering parameters) and Geotechnical Models (Ground Models with
predicted performance based on design parameters). If the Geologic Model is wrong, then neither
the Ground Model nor the Geotechnical Model can be correct. Fundamental geologic variability
makes some details unforeseeable. Insufficient geotechnical investigations, faulty interpretations,
or failure to portray results understandably contribute to inappropriate designs or failures. If the
geologist does not interpret the geology and explain it clearly, then the engineer will be forced to
interpret it or ignore it. Incomplete or inaccurate geotechnical site characterization can lead to
selection of incorrect models, geotechnical properties, and design values. Furthermore, project
managers responsible only for design and construction may view geologic site characterization as
extra cost if benefits result in improved life-cycle reliability or reduced maintenance costs but do
not improve design or construction.


The purpose of this paper is to promote application of geology in engineering projects, or

perhaps to revitalize its incorporation into engineering projects. The response to contemporary
economic conditions, in engineering as well as other sectors, has been to reduce aspects that
appear to have limited value. The application of geology in engineering projects may have
slipped for several reasons. First, fewer civil engineering programs include geologic content, so
graduates may not be familiar with the role of geology in engineering projects. Second,
engineering site investigations utilize quantitative methods (e.g., cone penetration tests [CPTs])
which may be used as surrogates for stratigraphy but do not contribute to rational geologic
interpretation. Third, geology sections in reports for some projects may use terms which
engineers do not understand and may be inserted, possibly as ““boiler plate””, at the report-
preparation stage without the information being considered in planning or implementing site
investigations or engineering analyses. Fourth, agencies responsible for reviewing engineering
reports may not have qualified reviewers for geology sections or appreciate their importance.
Fifth, geologists trained in traditional programs may not understand how to interpret aspects of
the geology relevant to a specific project or portray them in terms clearly understandable of
engineers. The last point has been the primary source of ““geologist-engineer jokes”” because of
the fundamentally different way each views the world. For example:
An engineer and a geologist who have worked together on several projects are on their way
to a field site. The engineer is frustrated at not being able to get a straight, unqualified answer


Foundation Engineering in the Face of Uncertainty


from the geologist. The engineer driving the car sees a pasture on the side of the road in which an
isolated, brown cow is standing. The engineer realizes that this setting provides an opportunity to
ask the geologist a question so simple that the geologist might give a straight answer.

Engineer: ““I am going to ask you a simple question about that brown cow in the pasture ahead.””
Geologist: ““Okay.””
Engineer: ““What color is that cow?”” The geologist studies the cow is they drive past.
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

Geologist: ““The color of that cow looks like 10YR 3/4.””

Engineer: ““What?!””
Geologist: ““Brown …… on this side.””

The point of this ““geologist-engineer”” joke is that the geologist first provided information
using a term that the engineer did not understand and then qualified the information because of
the observational nature of geology. The origin of the geologist’’s and engineer’’s differing views
of the world is rooted in the contrast between science and engineering, as outlined in Table 1.

Table 1. Selected contrasting elements of science and engineering.

Geological Science Engineering Design

Qualitative, observational Quantitative, data-driven
Largely interpretive Focused on specific design objectives
Site viewed as part of geologic region Site viewed as discrete location
Seeks to explain geologic details Focused on multiple project constraints
Seeks to predict distribution of Seeks to satisfy specific design
formations requirements

In the first Glossop lecture, Fookes (1997) related his own experience by comparing the
““realrocks”” with which he was confronted in practice to the ““schoolrocks”” that formed the basis
of his education and the ““labrocks”” that were described in textbooks. Realrocks, of course, are
those with complexities caused by the environment of deposition or emplacement, tectonic
processes, alteration, weathering, and, in general, the effects of geologic time. In fact, to most
engineers, geologists appear to be obsessed with time. The fundamental approach to mapping
geologic formations uses the time-rock unit which emphasizes the age of deposition,
metamorphism, or emplacement for rocks of sedimentary, metamorphic, or igneous (plutonic)
origin. For example, the symbol Mmc in part of the western United States denotes Mississippian
Manning Canyon Formation; Mississippian is an Epoch of the Carboniferous Period of the
Paleozoic Era –– in other words, it represents a specific part of the geologic time scale that
occurred approximately 300 to 355 million years ago.
In parts of Utah, the Manning Canyon Formation is a black shale. To a geologist, ““black
shale”” implies a deep-water marine environment of deposition with reducing conditions that
typically fixes iron and sulfur into the mineral pyrite. Environmental and engineering geologists
recognize that oxidation of pyrite-bearing rocks generates sulfuric acid and iron in ferrous form
that readily oxidizes to ferric form with associated undesirable water pollution. Engineering
geologists also recognize that ““black shale”” tends to form expansive soil and contributes to slope
instability. Where the Manning Canyon Formation crops out in urban areas along the Wasatch
Front in Utah, it is sufficiently notorious that geotechnical engineers also recognize its name.
A common geologic unit found on most published geologic maps is Qal which geologists
use to denote Quaternary alluvium. Quaternary is a Period of the Cenozoic Era and alluvium is a
general term that denotes sedimentary deposits composed of particles that were transported by
flowing water. Consequently, Qal can be used to represent sand, silt, clay, gravel, or

Foundation Engineering in the Face of Uncertainty


combinations of grain sizes that are stratified or massive. An even more general geologic map
symbol used by traditional geologists to denote young deposits that conceal older geologic
formations is Qu, undifferentiated Quaternary ““overburden””. These geologists typically care
about older formation bedrock and focus on a variety of research interests, such as igneous,
metamorphic, or sedimentary petrology, paleontology, structural geology, and tectonics.
Another branch of geology focuses on the most recent period of Earth history and may use
the symbol ““r”” on geologic maps to denote undifferentiated rock (sometimes called pre-
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

Quaternary ““underburden”” in discussions with bedrock geologists). These Quaternary geologists

are interested in the degree of soil development on parent material as indicator of landscape age,
stability, climate change, and evidence of prehistoric, but recent, seismic and slope activity.
Geologists who focus on interpreting landscape evolution are geomorphologists, whose
vocabulary includes strange-sounding words like hoodoo and yardang (a hoodoo is a pinnacle
commonly capped with a boulder-sized fragment of durable rock that protects underlying erodible
rock from raindrop impact and other effects of precipitation; a yardang is a wind-sculpted
landform in soft bedrock or cohesive sediment having an elongated pattern with the long axes
oriented approximately parallel to the wind direction dominant over recent geologic time).

Geologic Maps and Sections

A geologic map or a geologic section can be defined as an artistic representation of one
interpretation of geologic features and relationships inferred from limited observations of the
distribution of rock types, surficial deposits, and geologic structures often with little or no
subsurface data or laboratory test results. The symbols geologists use on geologic maps denote
mappable units, commonly called formations, and features geologists call structure which consist
of contact surfaces separating the mappable units and features internal to the mappable units,
including sedimentary structure (primarily stratification called bedding), metamorphic structure
(primarily fabric called foliation, schistosity, or cleavage), and tectonic structure (primarily
effects of stress called joints, faults, and folds). In essence, a geologic map is a type of geologic
model. A map, in general, shows the distribution of selected features; for example, a road map
shows highways and streets at varying levels of detail depending on the scale of the map and the
area of coverage. Road maps typically are based on surveyed positions of key intersections and
other locations that have some quantifiable accuracy. The details of map accuracy standards
(USGS, 1999) do not need to be understood in order for maps to be useful, but the concept of
map accuracy becomes important for engineers to consider when using geologic maps.
A geologic map shows the distribution of selected features of the geology; a geologic section
shows the distribution of selected geologic features in a vertical plane. Unlike features shown on
the road map, geologic features are rarely located by surveyed positions, although general use of
global positioning system (GPS) receivers by geologists is increasing. Instead, geologic features
are located by judgment which is based on observation, inference, and experience. The base maps
on which geologic data are recorded almost always have topographic contours and sometimes
have aerial-photograph images, both of which can be helpful to the geologist in estimating
location within the map area. The actual existence and identity of some geologic features are
interpreted on the basis of geologic reasoning using some representation of what has been called
the process of multiple working hypotheses (Chamberlain, 1890) in which alternative
explanations are carried forward as long as contrary evidence is not found.
Geologists use lines and symbols on maps and sections to represent formation boundaries
and structural features that include an expression of confidence or certainty in location and
interpretation (FGDC, 2006); the lines may be solid, long dash, medium dash, or short dash, with
or without question marks to portray combinations of scientific confidence and location accuracy.
FGDC (2006) defines the ““zone of confidence”” as a buffer zone surrounding a line or a circle
surrounding a point with a non-specified distance in ground units. In keeping with the point of the
““geologist-engineer joke””, the numerical value of the zone of confidence is qualified based

Foundation Engineering in the Face of Uncertainty


““on a number of factors: the area's geology, landscape terrain, vegetation cover, and (or)
cultural features; the scale of mapping; the quality and nature of the base map used; and
(or) a particular project's allotted field-mapping time or other logistical constraints. Because
this standard recognizes that the factors affecting the value of the zone of confidence will
vary from region to region (and from map to map), and because different agencies have
differing mapping needs and mandates, a single, universally applicable value for the zone
of confidence is not herein established. Instead, this standard advocates that the
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

responsibility for setting the value of the zone of confidence for a particular geologic map
or mapped area lies with each geoscience organization and each mapping geologist.””
(FGDC, 2006, p. 19).
Judgment-based interpretations of geologic features and structural relationships can be
critically important on engineering projects. Yet in almost no cases is uncertainty or variability
mentioned in the communication of relevant geologic details. So what then is the value of a
geologic map to an engineer responsible for the site characterization that will be the basis for
design of a project?

How Do Engineers Use Geologic Maps?

Traditional geologic maps tend to be used by engineers in one of three ways:
1. Geologic boundaries are accepted as deterministic truth; geologic units guide
geotechnical characterization based on subsurface data and laboratory test results.
2. The overall range of geologic conditions on a site is estimated and somehow applied to a
site as a single random variable with design dominated by quantitative geotechnical data.
3. The geology is ignored* and quantitative subsurface geotechnical data are used as a
surrogate for stratigraphy. (*The geology is neatly tucked away in an appendix to
document that it was done, in the event that a reviewer asks about it.)
Published geologic maps, of course, most likely were made for general-purpose use; in other
words, they were made by geologists for use by other geologists. General-purpose geologic maps
need to be translated for engineering application; engineering geologists with training and
experience in supporting engineering projects tend to be best suited to provide the translation of
published geologic maps, as well as to participate in geotechnical site characterization activities.
Engineering geologists tend to have skills in and knowledge of engineering geology,
hydrogeology, geomorphology, and soil and rock mechanics, as well as in-depth knowledge of
geology and familiarity with civil engineering design and practice (Hencher, 2012).
Precedent exists in the United States for including engineering geology in civil engineering
projects, along with guidance from the four notable agencies discussed in the next section.


Guidance provided by the following four agencies of the United States makes clear the point
that engineering geology is an integral part of geotechnical investigations for civil engineering
projects: Federal Highway Administration (FHWA), US Department of Agriculture Natural
Resources Conservation Service (NRCS), US Army Corps of Engineers (USACE), and US
Department of the Interior Bureau of Reclamation (BOR).
The FHWA guidance is in the form of a manual developed for a National Highway Institute
course entitled, Subsurface Investigations –– Geotechnical Site Characterization (Mayne et al.,
2002, p. 1-2), in which the role of the geotechnical engineer is defined as being ““responsible for
acquiring and interpreting soil, rock, and foundation data for design and construction of various
types of structures. The proper execution of this role requires a thorough understanding of the
principles and practice of geotechnical engineering, subsurface investigation techniques and
principles, design procedures, construction methods and planned facility utilization supplemented
with a working knowledge of geology and hydrology.”” A footnote attributed to ““the geotechnical

Foundation Engineering in the Face of Uncertainty


engineer”” states ““The term geotechnical engineering in this manual also applies to engineering
geologists who are involved in subsurface investigations for civil engineering applications.””
The FHWA guidance notes that geologic maps developed for engineering purposes display
local, detailed geologic data that are used to characterize and document the condition of a rock
mass or outcrop. The FHWA guidance also notes that geologic maps should be prepared by or
under the supervision of qualified personnel trained in geology or engineering geology. Those
individuals responsible for geologic mapping must become familiar with the regional and local
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

geology of a site area by studying available reports and maps.

The NRCS is responsible for soil and water conservation by means that include engineered
works, such as flood control dams. NRCS guidance comes in the form of policy (National
Engineering Manual, NRCS, 2010) and technical guidance (National Engineering Handbook, Part
631 Geology, NRCS, 2012). The NRCS (2010, §531.0.A and B) policy states that, ““The State
Conservation Engineer is responsible for ensuring that geologic conditions at NRCS project sites
are sufficiently characterized to support proper conservation planning and sound engineering
design and construction.”” and ““Investigations are conducted by a person holding the appropriate
job approval authority for the class of structure …… and who is trained to recognize geologic
hazards and to provide information adequate for [practical] design. A qualified geologist must
conduct investigations in areas where experience or information is limited, where geologic
conditions are complex or unstable, where the kinds of construction materials to be used are
complex or questionable, or where the potential for significant economic or environmental
damage or loss of life is high if the structure or practice were to fail or perform poorly.””
The NRCS (2012) technical guidance identifies engineering geologic investigations at the
complete spectrum of project-related opportunities: 1) geologic reconnaissance, 2) preliminary
site investigation, 3) detailed site investigation, 4) geologic investigation during project
implementation and construction, 5) investigations for repair, rehabilitation, and
decommissioning of structures, 6) groundwater investigations, 7) investigation of structural
problems caused by erosion or sedimentation, 8) watershed sediment yield studies, and 9) stream
channels and stream corridors. Baynes et al. (2005) used the phrase ““total engineering geology
approach”” to represent the value-added contribution that engineering geologists can make
throughout a project’’s life cycle.
The USACE (2001) guidance is published in Engineer Manual EM-1110-1-1804 entitled,
Geotechnical Investigations. The cover letter transmitting this Manual states, ““This manual
establishes criteria and presents guidance for geotechnical investigations during the various stages
of development for both civil and military projects. …… This manual applies to all USACE
Commands having either military or civil works responsibilities. …… Geotechnical investigations
are made to determine those geologic, seismologic, and soils conditions that affect the safety, cost
effectiveness, design, and execution of a proposed engineering project. Because insufficient
geotechnical investigations, faulty interpretation of results, or failure to portray results in a clearly
understandable manner may contribute to costly construction changes, post-construction remedial
work, and even failure of a structure, geotechnical investigations and subsequent reports are an
essential part of all civil engineering and design projects.””
The USACE (2001, §1-4) elaborates on the geologic content of geotechnical investigations:
““Investigations performed to determine the geologic setting of the project include: the
geologic, seismologic, and soil conditions that influence selection of the project site; the
characteristics of the foundation soils and rocks; geotechnical conditions which influence
project safety, design, and construction; critical geomorphic processes; and sources of
construction materials. A close relationship exists between the geologic sciences and other
physical sciences used in the determination of project environmental impact and mitigation
of that impact. Those individuals performing geotechnical investigations are among the
first to assess the physical setting of a project. Hence, senior-level, experienced personnel
are required to plan and supervise the execution of a geotechnical investigation.

Foundation Engineering in the Face of Uncertainty


Geotechnical investigations are to be carried out by engineering geologists, geological

engineers, geotechnical engineers, and geologists and civil engineers with education and
experience in geotechnical investigations. Geologic conditions at a site are a major
influence on the environmental impact and impact mitigation design, and therefore a
primary portion of geotechnical investigations is to observe and report potential conditions
relating to environmental impact.””
The USACE (2001, Chapter 2) guidance identifies the following levels of detail for civil
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

works projects: 1) Reconnaissance and Feasibility Studies, 2) Preconstruction Engineering and

Design Studies, and 3) Construction Activities. USACE (2001, Chapter 3) identifies regional
geologic and site reconnaissance investigations needed to develop an understanding of the
regional geology of a project area which forms the basis for the scope of data collection and
analysis needed for early site investigations. USACE (2001, Chapter 3) further recognizes the
value of geologic models, calling for geologic background and field data to be used to construct a
geologic model for each site being considered in a site-selection study. The geologic model
should be revised as additional relevant information is obtained and should identify locations and
types of geologic features that could affect performance or cost of construction of project
features. Preliminary studies of the geology, seismology, and hydrology, and their impact on
project economics, are to be used to indicate the most favorable sites before preliminary
subsurface investigations are undertaken.
USACE (2001, §4-2) notes that an areal geologic mapping program can be considered
complete only if existing geologic studies of an area have been combined with current geologic
mapping and appropriate remote sensing techniques. USACE (2001, §4-3) addresses site-scale
geologic mapping and notes that:
““Large-scale and detailed geologic maps should be prepared for specific sites of interest
within the project area and should include proposed structure areas and borrow and quarry
sites. Investigation of the geologic features of overburden and rock materials is essential in
site mapping and subsequent explorations. Determination of the subsurface features should
be derived from a coordinated, cooperative study by geotechnical engineers and geologists.
The geologist should contribute information on origin, distribution, and manner of
deposition of the overburden and rock. The geotechnical engineer or engineering geologist
should determine the engineering properties of the site foundation and potential
construction materials, potential problem materials or conditions, application of geologic
conditions to design, and the adaptation of proposed structures to foundation conditions.””
…… ““A good preliminary geologic map should be prepared prior to making any subsurface
borings to provide an approximate picture of the geologic conditions and hazards at a site.
Such a map permits borings to be strategically located. For each proposed boring, an
estimate should be made of the subsurface conditions that will be encountered, such as
depths to critical contacts and to the water table. This estimate is possible, at least in an
approximate manner, if geologic mapping has been performed to determine the geologic
structure, lithology, and stratigraphy. The process of progressively refining the model of the
geologic structure and stratigraphy by comparison with boring information is the most
efficient and cost-effective means to develop a complete understanding of the geologic site
USACE (2001, §4-4) addresses geologic mapping during construction and notes that:
““Construction maps record in detail geologic conditions encountered during construction.
Traditionally, a foundation map is a geologic map with details on structural [geology],
lithologic, and hydrologic features. It can represent structure [project] foundations, cut
slopes, and geologic features in tunnels or large chambers. The map should be prepared for
soil and rock areas and show any feature installed to improve, modify, or control geologic
conditions.”” …… ““The mapping of foundations is usually performed after the foundation has
been cleaned just prior to the placement of concrete or backfill. The surface cleanup at this

Foundation Engineering in the Face of Uncertainty


time is generally sufficient to permit the observation and recording of all geologic details in
the foundation.”” …… ““The person in charge of foundation mapping should be familiar with
design intent via careful examination of design memoranda and discussion with design
personnel. The actual geology should be compared with the geologic model developed
during the design phase to evaluate whether or not there are any significant differences and
how these differences may affect structural integrity. The person in charge of foundation
[geologic] mapping should be involved in all decisions regarding foundation modifications
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

or additional foundation treatment considered advisable based on conditions observed after

preliminary cleanup. Design personnel should be consulted during excavation work
whenever differences between the actual geology and the design phase geological model
require clarification or change in foundation design. [Geologic] Mapping records should
include details of all foundation modifications and treatment performed.”” …… ““[Geologic]
Mapping of tunnels and other underground openings must be planned differently from
foundation mapping. Design requirements for support of the openings may require
installation of support before an adequate cleanup can be made for [geologic] mapping
purposes. Consequently, mapping should be performed as the heading or opening is
advanced and during the installation of support features. This requires a well trained
geologist, engineering geologist, or geological engineer at the excavation at all times.
Specifications should be included in construction plans for periodic cleaning of exposure
surfaces and to allow a reasonable length of time for [geologic] mapping to be carried out.””
The mission of the Bureau of Reclamation (BOR, 1998) is to ““manage, develop, and protect
water and related resources in an environmentally and economically sound manner ……”” BOR
(1998, p. 129) guidance describes the responsibilities of the engineering geologist to include
defining, evaluating, and documenting ““site-specific geologic conditions relating to the design,
construction, maintenance, and remediation of engineered structures or other sites.”” Geologic
maps and cross sections generated from site-specific data ““(1) serve as a record of the location of
factual data; (2) present a graphic picture of the conceptual model of the study area based on the
available factual data; and (3) serve as tools for solving three-dimensional problems related to the
design, construction, and/or maintenance of engineered structures or site characterization.””
BOR (1998, p. 129-130) states that the engineering geologist performing geologic mapping
is responsible for 1) recognizing geologic conditions that could adversely affect proposed or
existing structures or hazardous waste sites and 2) integrating ““pertinent geologic data …… into a
conceptual [geologic] model of the study area and presenting this conceptual model to design and
construction engineers, other geologists, hydrologists, site managers, and contractors in a form
that can be understood.”” BOR (1998, p. 130) gives the following guidance:
““The engineering geologist needs to realize that geologic mapping for site characterization
is a dynamic process of gathering, evaluating, and revising geologic data and that the
significance of these data, both to the structure and to further exploration, must be
continually assessed. The initial exploration program for a structure is always based on
incomplete data and must be modified continuously as the site geology becomes better
understood. The key to understanding the site geology is through interpretive geologic
drawings such as geologic maps, cross sections, isopachs, and contour maps of surfaces.
These working drawings, periodically revised and re-interpreted as new data become
available, are continuously used to assess the effects of the site geology and to delineate
areas where additional exploration is needed. These drawings are used in designs,
specifications, and modeling and [are] maintained in the technical record of the project.””
It is clear from the discussion in the preceding paragraphs that engineering geology is a
fundamental part of geotechnical investigations for civil engineering projects. The guidance cited
in this section demonstrates that, to these agencies, engineering geology is important and integral
to design and understanding of project performance.

Foundation Engineering in the Face of Uncertainty


Exploration Versus Investigation

It is common for engineers and geologists to use the term ““exploratory”” for geotechnical
borings or subsurface investigations. The FHWA guidance (Mayne et al., 2002) uses ““subsurface
investigation”” in its title, but also uses ““subsurface exploration”” in a first-order heading.
““Exploration”” refers to the process of searching for the purpose of discovery; the term implies
lack of expectation of what might be discovered. In other words, what might be discovered may
well be a surprise. Applying this concept to ““exploratory borings”” suggests that materials and
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

conditions encountered in each boring would be a surprise to the geotechnical team. Collecting
““surprises”” at a number of locations as a series of unordered facts would require inductive
reasoning to synthesize them into an hypothesis of the geotechnical characteristics of the site.
It seems reasonable to consider the references to geologic models in the guidance cited in the
previous section to provide initial hypotheses of the regional or site geology. USACE (2001,
§3.1) specifies that subsurface investigations should begin only after a geologic model is
constructed. Therefore, deductive reasoning would be applied to subsurface data used to test the
hypothesis represented by the geologic model. Consequently, geotechnical borings should be
drilled at locations that are useful for testing and refining the geologic model hypotheses, as well
as at locations needed for geotechnical laboratory testing (e.g., along the centerline of a dam or
within the footprint of a building). This approach is stated explicitly in USACE (2001, §4-3).
Legal liability and litigation associated with consulting geotechnical engineering services in
the United States seems to have tainted some terms, including ““investigation””. It seems that the
term ““exploration”” is preferred for limiting liability exposure because it implies a lower level of
certainty; most consulting reports include disclaimers that stipulate the geotechnical conditions
described are valid only at locations of borings and at the time they were drilled; projections of
geotechnical conditions between borings are approximate and require verification. The deductive
process of developing, testing, refining, and updating geologic models of sites is well documented
and should allow the information from each boring to be used in a systematic investigation, rather
than to consider a series of borings as exploratory surprises. This concept was introduced by
Glossop (1968) in the Eighth Rankine Lecture in which he stated, ““If you do not know what you
should be looking for in a site investigation, you are not likely to find much of value.””

Geologic Models in Geotechnical Engineering

Models are created typically to enhance understanding and facilitate communication;
commonly they are simplified. Models can illustrate component relations, evolutionary
sequences, processes, or conditions using individual sketches, maps and sections, block models,
series of sketches, and flow diagrams. A useful model is Burland’’s Soil Mechanics Triangle
(Burland, 1987). This model has had several iterations of evolution (Morgenstern, 2000; Knill,
2003; Sullivan, 2010) including the one presented in Figure 1 prepared for this paper.
The Geologic Model in Figure 1 is a representation of the site geologic conditions relevant to
the proposed project. The Ground Model in Figure 1 is the Geologic Model expressed in terms of
engineering parameters. The Geotechnical Model in Figure 1 is the Ground Model with design
parameters used to establish predicted performance of the proposed project. Essentially, Figure 1
makes the case that if the Geologic Model is incorrect, then neither the Ground Model nor the
Geotechnical Model can be correct.

Engineering Geology in Reliability-Based Design

Reliability-based design is becoming increasingly important in geotechnical engineering to
provide a theoretical framework to recognize, quantify, and manipulate uncertainty (Kulhawy et
al., 2012). Inherent variability of geotechnical materials is one of the primary sources of
uncertainty that has been subdivided into a smoothly varying trend and a component that
fluctuates about the trend (Phoon et al., 1995), as shown in Figure 2. Geology is not yet
incorporated into geotechnical characterization for the reliability-based design process.

Foundation Engineering in the Face of Uncertainty

Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

Figure 1. Geotechnical site characterization model. Adapted from Burland (1987), Anonymous
(1999), Morgenstern (2000), Knill (2003), Barbour & Krahn (2004), and Sullivan (2010).

Figure 2. Schematic geologic section showing idealized variability of a hypothetical

geotechnical property, making this a Ground Model from Figure 1. Adapted from
Phoon et al. (1995); ““geologic unit”” is used here in lieu of Phoon’’s ““soil layer””.

Akbas & Kulhawy (2010) provide an example of a geotechnical variability analysis in which
the geology of the Ankara Clay in Turkey was considered. The Ankara Clay exhibits high
plasticity and swelling potential, which have been the focus of most published geotechnical
studies about it. The paper by Akbas & Kulhawy (2010) focused on other key index and
performance properties of the Ankara Clay, which is part of a sedimentary sequence of fluvial
origin deposited during the Late Pliocene. Values of Liquid Limit, Plastic Limit, Plasticity Index,
and natural water content for approximately 15 m of the Ankara Clay in a single borehole are
summarized in Figure 3 (part A). Average index property values for the Ankara Clay from many
boreholes were used by Akbas & Kulhawy (2010) in their analysis; the purpose of the discussion
here is to consider the stratigraphy rather than the details of their variability analysis.
Figure 3 (part B) is annotated to show possible clay mineralogy that could explain the
pattern of Liquid Limit and Plasticity Index values. Whereas the values plotted by Akbas &
Kulhawy (2010) were from a single ““soil layer”” of the Ankara Clay, they may actually represent

Foundation Engineering in the Face of Uncertainty

Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

Figure 3. Values of Liquid Limit (WL), Plastic Limit (WP), Plasticity Index (PI), and natural
water content (Wn) from a single borehole in the Ankara Clay; based on Akbas & Kulhawy
(2010). A. Summarized values; B. Annotated diagram speculating on clay mineralogy based on
WL & PI; montmorillonite is a swelling clay mineral; Na denotes sodium; Ca denotes calcium.

two stratigraphic (i.e., mappable) units with different properties. The importance for the analysis
of Akbas & Kulhawy (2010) of ““lumping”” the Ankara Clay into a single unit or ““splitting”” it into
four subunits could be evaluated statistically. However, the point of this discussion is to draw
attention to the geologic thought process of identifying as many details as possible, evaluating
them under different hypotheses, and attempting to develop the most complete (i.e., the best)
geologic explanation. Straight-line interpolation between data points is another geologic issue.


USACE (2001, §1-4) states that ““Insufficient geotechnical investigations, faulty

interpretation of results, or failure to portray results in a clearly understandable manner may
contribute to inappropriate designs, delays in construction schedules, costly construction
modifications, use of substandard borrow material, environmental damage to the site, post-
construction remedial work, and even failure of a structure and subsequent litigation.”” In other
words, a poor job of performing a geotechnical investigation, or doing a poor job of
communicating the results, can be a source of financial risk to a project owner, as well as a
potential source of physical risk to people, equipment, capital works, and the environment.
Baynes (2010) recognized that identifying and managing geotechnical hazards and risks are
fundamental project management activities in all projects that have geotechnical content. The
logical inference by most geotechnical engineers and geologists is that geotechnical hazards are
all part of the ground conditions at a site; however, two of the seven project components
discussed by Baynes (2010) represent hazards that are connected to office management rather
than to specific field conditions. Baynes’’ (2010) geotechnical hazards and risks are summarized
in Table 2.

Foundation Engineering in the Face of Uncertainty


Table 2. Geotechnical hazards and risks identified by Baynes (2010).

Project-component hazard Importance to risk exposure

Poor management of Lack of appreciation of importance of geotechnical conditions
overall project leading to insufficient funding for adequate characterization
Poor management of site Incomplete geotechnical characterization; inaccurate
2 investigation and contract representation of site conditions on construction contract
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

documents documents leading to ““changed condition”” claims

Selection of poor Inadequate or incomplete geotechnical characterization used
analytical model as basis for deciding what factors influence design
Selection of poor design- Inadequate characterization, communication, or understanding
parameter values of site factors leading to selection of improper values
Unforeseeable geologic Inherent quality of geologic complexity leading to residual
details encountered uncertainty even after reasonable level of site investigation
Unforeseen geologic Inadequate or incomplete geotechnical characterization
conditions encountered leading to important site details remaining undiscovered
Ground effects larger than Inadequate appreciation of potential scale of conditions or
anticipated processes recognized during proper site investigation

For this paper seeking to emphasize the value of engineering geology, Baynes (2010)
provides a powerful argument with his comments that risks arise from inadequate understanding
by project managers of the importance of ground conditions, as well as inadequate understanding
of the geologic component of actual ground conditions. A project owner’’s representative seeking
to satisfy a requirement for geotechnical studies at the lowest cost may actually be the largest
single contributor to risk on the project. Geotechnical consultants and their professional liability
insurance companies may be the underwriters of an unwitting risk-mitigation strategy. The risk
realization of a poorly managed project that has inadequate geotechnical characterization most
likely is represented in several of the seven project-component hazards listed in Table 2.


The numerical value of the zone of confidence of stratigraphic or structural contacts on

digital geologic maps according to FGDC (2006) depends on a number of factors, including the
area’’s geologic complexity and the mapping geologist’’s experience. Geologists use the line
pattern (solid, dashed, dotted, queried) to represent scientific confidence in interpretation that is
understood by other geologists, but it is neither quantitative nor understood by non-geologists.
The position accuracy of the contact line or fault can be quantified once the geologist reaches an
interpretation. This geologic ““code”” needs to be improved to express uncertainty and variability.
Two existing approaches may be valuable to consider as a means of expressing uncertainty
and variability of geology for use in general, and on reliability-based design projects in particular:
1) Outcrop confidence and 2) quality levels. NRCS (2002) defines three level of outcrop
confidence as a relative measure of predictability or homogeneity of the lithology and structural
domain of rock units within a site area (Table 3). Examples of outcrop confidence levels are
shown in Figure 4. The confidence level designations for a site depend on the scale of the project.
For example, a project consisting of a single road cut near the crest of a hill probably would focus
only on the geology of the upper part of the hill; whereas a dam project would focus on the
geology of the abutments, alignment across the valley, the slopes above the abutments, and the
slopes above the reservoir, as well as the watershed of the reservoir for sediment yield estimates.
Examples of high outcrop confidence (Figure 4A & 4B) are Stone Mountain granite

Foundation Engineering in the Face of Uncertainty


(Permian quartz monzonite, Whitney et al., 1976) near Atlanta, Georgia, and Navajo sandstone
(Jurassic cross-bedded sandstone, Gregory & Moore, 1931) below Glen Canyon Dam, Arizona.
Examples of intermediate outcrop confidence (Figure 4C & 4D) are East Canyon fault cutting
silty sandstone (Jurassic Preuss Formation, Piety et al., 2010) east of Salt Lake City, Utah, and
sandstone (Permian Schnebly Hill Formation, Lindberg, 2010) at Cathedral Rock, Sedona,
Arizona. Examples of low outcrop confidence (Figure 4E & 4F) are aa lava (1969-74 basaltic
lava erupted from Mauna Ulu, Decker & Decker, 1992) exposed in road cut on Chain of Craters
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

Road, Hawai’’i Volcanoes National Park and volcanic rocks (Tertiary tuff and basaltic andesite
dikes, Mills, 1994) exposed on the Nevada side of the Colorado River at Hoover Dam.
Formally defined quality levels are used by FHWA (2011) for utility location information
needed for subsurface utility engineering (SUE) for transportation projects (Table 4). The quality
levels are hierarchical; Level C would not be undertaken without having the results of Level D;
Level B would not be undertaken without having the results of Level C; Level A would not be
undertaken without having the results of Level B.
The four quality levels summarized in Table 4 correlate reasonably well to stages of
engineering geologic investigation. Available maps, reports, aerial photographs, and remote
sensing data are used in an office-based ““desktop study”” to develop a preliminary geologic model
that frequently becomes the basis for planning the field reconnaissance and possibly locations and
depths of subsurface investigation borings, test pits, and trenches: Level D, preliminary geologic
model based on existing information. Based on the understanding gained from developing a
preliminary geologic model, the field reconnaissance is performed with an objective of validating,
refining, and updating the geologic model: Level C, geologic model with limited verification.
Surface geophysical surveys may be part of field reconnaissance activities, or they may be a
separate task, and sometimes are omitted in geologic model development. If surface geophysical
surveys are performed, then the Level C geologic model is used as the hypothesis to be tested by
appropriate geophysical methods for the site and the geophysical survey results are used to refine
and update the geologic model: Level B, geologic model with limited subsurface data. Invasive
subsurface investigation data from borings, trenches, test pits, adits, shafts, and a variety of field
tests would advance the level of refinement to Level A, geologic model with limited subsurface
data; however, unlike a subsurface utility locating program, developing a comprehensive
geologic model for site characterization is not completed until a site-scale excavation exposes the
subsurface geology in three dimensions. Even then, in many geologic settings, the geologist
might qualify the refined and updated geologic model of the site by adding that it was ““valid only
to the depth of the excavation.””
Therefore, the geologic equivalent of the Level A geologic model has gradations of quality
that, at least to some degree, are based on the same types of considerations that resulted in the
FGDC (2006) guidance in advocating that the numerical value of the zone of confidence be set by
each geoscience organization and mapping geologist. A few well-placed borings can add
significantly to a geologic model; however, a program that includes inclined borings, borehole
acoustic televiewer logging, petrographic analysis, and laboratory testing may be needed to
provide critical data for an improved Level A, geologic model with extensive subsurface data.
Geologic complexity is addressed to at least some degree in the NRCS (2002) outcrop
confidence (Table 3). Morgenstern & Cruden (1977) differentiate between geologic complexity
and geotechnical complexity. Geologic complexity refers to the qualities and details that,
combined, are the focus of geologists seeking to understand and explain the history of geologic
processes that have occurred to produce the formations as they appear in the field today.
Geotechnical complexity relates to variability in strength, stiffness, and hydraulic conductivity of
soil and rock masses as these properties might affect the performance of engineered works. The
contrast between geologic and geotechnical complexity is useful to consider in the context of the
geotechnical site characterization model (Figure 1). Two straightforward examples used by
Morgenstern & Cruden (1977) consist of the spillway rock at Mica Dam on the Columbia River

Foundation Engineering in the Face of Uncertainty


Table 3. Outcrop confidence levels. Adapted from NRCS (2002).

Level Confidence Rock conditions and site characteristics

Rocks are massive and homogeneous, vertically and laterally
1 High
extensive. Site geology has a history of low tectonic activity.
Conditions are generally predictable, with lateral and vertical
2 Intermediate variability. Structural features produced by tectonic activity tend to
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

have systematic orientation and spacing.

Conditions are extremely variable because of complex depositional or
3 Low structural history, mass movements, or landscape evolution.
Significant lateral and vertical changes in characteristics are common.

Figure 4. Examples of high outcrop confidence (A, B), intermediate outcrop confidence (C, D),
and low outcrop confidence (E, F). See text for brief descriptions of geology; photo 4A provided
by AMEC, used with permission; photos 4B-4F were taken by the author.

Foundation Engineering in the Face of Uncertainty


Table 4. Summary of underground utility locating quality levels. Adapted from FHWA (2011).

Quality level Basis for location Utility location quality characteristics

Existing Basic information for utility locations coming from
information existing utility records or verbal recollections.
The most commonly level of information using surveyed
C visible utility features (e.g., manholes, valve boxes, etc.)
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

correlated with records of existing utilities.
Appropriate geophysical methods used to determine
existence and horizontal positions, surveyed to project
B control, of virtually all utilities within the project limits;
addresses problems of inaccurate records, abandoned or
unrecorded facilities, and lost references.
The highest level of accuracy, providing information for
plan-and-profile mapping using nondestructive exposure
A (e.g., vacuum excavation using HDPE tubes) of utilities,
surveyed to project control, to document type, size,
condition, and other characteristics.

in British Columbia and the notorious quick clay formations in Canada and Scandinavia. The
spillway rock mass included biotite schist, quartzite, and marble formations that had been
subjected to one phase of metamorphism and two phases of folding. Despite this geologic
complexity, the rock mass over the scale of the spillway had good rock quality and was
essentially uniform across the spillway. The geologically complex rock mass was relatively
simple in geotechnical terms. In contrast, however, the quick clay formations were deposited in a
geologically recent post-glacial marine environment and are stratigraphically homogeneous.
Local leaching of salt from the clay formations resulted in complex distributions of sensitivities
and associated stability behavior.
Morgenstern & Cruden (1977) identify three types of processes that contribute to geologic
and geotechnical complexity: 1) genetic processes, 2) epigenetic processes, and 3) weathering
processes. Genetic processes are those related to initial deposition or emplacement of the earth
materials. Epigenetic processes are those that act on earth materials after initial deposition of
sediments or emplacement of igneous rocks, and include lithification of sediments into rock
formations; epigenetic also refers to tectonic processes that produce tilting, folding, and faulting
(i.e., rock structure), heat and pressure that produces metamorphic texture (e.g., foliation, slaty
cleavage, and schistosity), and metasomatism that chemically alters rock-forming minerals into
clay minerals, typically significantly below the ground surface by action of hydrothermal and
other fluids commonly at high temperatures. Weathering refers to physical and chemical
processes that act on earth materials at the surface of the earth; weathering essentially includes
low-temperature metasomatism as well as physical processes of burrowing animals, plant-root
growth, freezing-thawing, and heating-cooling. The geologic aspects of genetic, epigenetic, and
weathering processes are implicit in the NRCS (2002) outcrop confidence assessment (Table 3).
Tectonic processes acting on a thick-bedded sandstone (e.g., Figure 4B) could produce
inclined planes of weakness that could control stability of slopes and excavations; in essence, the
planes could control geotechnical behavior. The discontinuities in a rock mass could be
complicated further by clay minerals produced by chemical alteration of rock or coating the
planes of weakness. The overprint of alteration and weathering on rock masses is influenced or
controlled by rock structure and tends to form patterns that are complicated at the scale of
building sites, with obvious associated geotechnical engineering implications.

Foundation Engineering in the Face of Uncertainty


The geotechnical importance of geologic complexity is varied, of course. Morgenstern &

Cruden (1977) discuss the concept of facies as used by geologists to denote characteristics of a
single mappable formation that vary gradually over some vertical interval or lateral distance.
Although the concept of facies is not well known to engineers, Morgenstern & Cruden (1977)
also mention geotechnical facies. The concept of geotechnical facies is consistent with the
geotechnical characterization model (Figure 1). Variability in the content of swelling clay
minerals across a site could result in a single geologic formation having substantially different
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

geotechnical properties (e.g., geologic unit 2a and unit 2b in Figure 2). Anderson (1989) applied
the facies terminology to hydrogeologic models of glacial and glacio-fluvial sediments.


Griffiths et al. (2012) prepared a state-of-the-art paper addressing risk assessment in

geotechnical engineering which discussed stability analysis of highly variable soils. Several
examples of the power of the finite element slope stability approach were given, including those
emphasizing the ability for irregular and multiple failure surfaces to be identified. Griffiths et al.
(2012) note that ““soils and rocks are the most variable of all engineering materials, so when an
engineer chooses ‘‘characteristic values’’ of the soil shear strength for a slope analysis, it is very
likely that some parts of the slope consist of soil that is stronger than the characteristic values, and
other parts that are weaker.”” With an example of a checkerboard pattern of strong and weak soil
elements in a simple two dimensional finite element analysis, Griffiths et al. (2012) held constant
all basic parameters, including the mean value of undrained shear strength, but the strong soil
elements were made stronger and the weak soil elements were made weaker in successive
analyses. The factor of safety of the slope decreased as the ratio of strong to weak soil strength
values increased.
The distribution of soil strength elements in the Griffiths et al. (2012) checkerboard model
was a regular pattern, unlike that found in natural deposits. The growing interest in risk
assessment has resulted in a shift in emphasis from ““factor of safety”” to ““probability of failure””,
and application of First Order Second Moment (FOSM) and First Order Reliability Method
(FORM) to slope stability problems. Limitations of the first order methods are related to their
inability to account for spatial correlation in random materials, and their reliance on simple
shapes of sliding surfaces. The deterministic finite element method has become more versatile
with inclusion of random field theory as incorporated in the Random Finite Element Method
(RFEM). Griffiths et al. (2012) describe the input to RFEM as the mean, standard deviation, and
spatial correlation length of the soil strength parameters; each soil layer in the model has its set of
statistical input parameters. The mean and standard deviation parameters are reasonably well
understood in quantitative data sets; the less well known spatial correlation length parameter is
the distance over which spatially random values will have a tendency to be correlated in a random
field. The random field used by Griffiths et al. (2012) is based on a normal distribution expressed
with a Gaussian function; non-normally distributed parameters may be transformed into normal
distributions as the logarithm or natural logarithm of the measured value.
In the elegant RFEM models produced by Griffiths et al. (2012), the model layers seem to be
defined deterministically as soil layers. Gaussian smoothing is a common effect function on
graphics software for bitmap images. A test of the sensitivity of Gaussian smoothing on a
completely exposed, two-dimensional geologic data set is provided in Figure 5. The initial image
(Figure 5A) is a photograph of decorative stone on a column in the Grand Hall of Union Station
in St. Louis, MO. The rock appears to be gneiss with compositional banding of mafic (dark) and
felsic (light) minerals; local, concentrated shearing truncates the prominent dark band in the
center of the image and two episodes of fracturing and fracture-filling with quartz or feldspar are
marked by the white streaks. The image is rotated counterclockwise from its position on the
column in the Grand Hall. The image represents an area of approximately 1000 mm by 800 mm.

Foundation Engineering in the Face of Uncertainty


It was taken with a 14 megapixel (Mpxl) camera. The portion of the image used for this analysis
was cropped to a size of 2100 x 1501 pixels (3.152 Mpxl) at a resolution of 300 dpi (dots per
inch). This image was converted to 8-bit grayscale and is displayed in Figure 5A; a grayscale map
of the 300 dpi image is presented above the image. A grayscale histogram of Figure 5A is
displayed above Figures 5C and 5D. The statistics of the grayscale values are:

Mean µ = 120 Standard deviation V = 40 Covariance COV = 0.33.

Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

Figure 5. Example of geologic detail becoming lost by smoothing. 5A: 3.152 MPxl image at
a resolution of 300 dpi of decorative gneiss on a column in the Grand Hall, Union station, St.
Louis, MO; grayscale map of image above the image; grayscale histogram of image above
5C. 5A’’: detail of part of Image 5A. 5B: Image 5A converted to 0.008 MPxl image at a
resolution of 72 dpi. 5C: Image 5A with 80-160 threshold applied. 5D: Image 5B with 80-160
threshold applied. 5E: Image 5B with 0.5-pixel Gaussian smoothing and threshold applied.
5F: Image 5B with 1.0 pixel Gaussian smoothing and threshold applied. 5G: Image 5B with
2.0 pixel Gaussian smoothing and threshold applied. 5H. Image 5B with 5.0 pixel Gaussian
smoothing and threshold applied. Grayscale maps above 5E, 5F, 5G, and 5H correspond to
the Gaussian smoothing of the respective images. The T symbol in 5E, 5F, 5G, and 5H
denotes the spatial correlation length parameter. See text for discussion

Foundation Engineering in the Face of Uncertainty


If the image in Figure 5A actually represented a relevant geologic or geotechnical condition,

the grayscale values could be correlated with an appropriate parameter. The example illustrated in
Figure 5 is intended to draw attention to the impact of Gaussian smoothing on geologic details,
stratigraphy and filled fractures, in this case, if the geology were considered to be random.
Some of the details of the rock in Figure 5A are enlarged in Figure 5A’’, which represents an
area of about 450 mm by 360 mm at a resolution of 300 dpi; the location of Figure 5A’’ is
indicated in Figure 5A and 5B by a rectangle with white dashed lines. Figure 5B is the Figure 5A
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

image converted to 100 x 80 pixels (0.008 Mpxl) at a resolution of 72 dpi.

The grayscale histogram of Figure 5A is marked with the mean and standard deviation
values. The mean ± 1 standard deviation values were used as threshold limits for image analysis
using tools available in ImageJ software ( These thresholds are applied
in Figures 5C through 5H; the uniform medium gray tone (grayscale value of 76) replaces all
pixels with grayscale values that are essentially ““average”” (within 1 standard deviation of the
mean). Figure 5C is the 300 dpi image in Figure 5A with the threshold applied, whereas Figure
5D is the 72 dpi image in Figure 5B with the threshold applied.
Figures 5E, 5F, 5G, and 5H are a series of images with Gaussian smoothing applied. The
radius of the smoothing function is set in terms of the image pixel size. It should be noted that the
conversion from 300 dpi to 72 dpi was a form of smoothing, also; some of the fracture filling
details remain in the image, but most of them are no longer discernable. The banding remains
discernable. Figure 5E uses a 0.5-pixel radius for the Gaussian smoothing; the dimensions of the
image area are known approximately, so 0.5 pixel corresponds to a length of 5 mm, which would
be comparable to the spatial correlation length in the RFEM of Griffiths et al. (2012).
Figure 5F uses a 1-pixel radius for the Gaussian smoothing, corresponding to a spatial
correlation length of 10 mm; dark banding is still visible in the image. Figure 5G uses a 2-pixel
radius for the Gaussian smoothing, corresponding to a spatial correlation length of 20 mm; dark
banding is still visible in the image. Figure 5H uses a 5-pixel radius for the Gaussian smoothing,
corresponding to a spatial correlation length of 50 mm; dark banding is no longer discernable in
the image. Griffiths et al. (2012) state that slopes analyzed with higher spatial correlation lengths
result in smooth failure mechanisms, whereas slopes analyzed with lower spatial correlation
lengths display ragged, complex interactions among finite elements in the model. Griffiths et al.
(2012) use a dimensionless representation of the spatial correlation length for the undrained shear
strength (4c = T(ln c)/H, where T(ln c) is the natural log transformation of undrained strength and H
is the height of the slope being analyzed). As H becomes large, 4c Æ 0 and the modeled slope is
essentially homogeneous with a constant strength equal to the median value. As H becomes
small, 4c Æ ’ and the modeled slope is essentially homogeneous, but with different strength
values in each computer simulation.
The Gaussian smoothing example displayed in Figure 5E, 5F, 5G, and 5H shows progressive
loss of geologic detail. The dimensionless spatial correlation values used by Griffiths et al. (2012)
were 4c = 0.5 and 2.0; spatial correlation lengths in Figure 5E, 5F, 5G, and 5H, respectively),
correspond to hypothetical slope heights listed in Table 5. Relatively short spatial correlation
lengths that preserve some of the geologic details in this example correspond to small
hypothetical slope heights. Conversely, larger hypothetical slope heights correspond to relatively
long spatial correlation lengths in this example. Even the relatively long spatial correlation
lengths are small percentages of the example dimensions (e.g., T = 50 mm corresponds to 6.25%
of 800 mm and 5.0% of 1000 mm). This example of Gaussian smoothing suggests that potentially
relevant geologic details could be lost with smoothing distances that are small percentages of
model dimensions.
Terzaghi’’ s (1929) paper entitled ““The Effects of Minor Geologic Details on the Safety of
Dams”” referred to ““features that can be predicted neither from the results of careful investigations
of a dam site nor by means of a reasonable amount of test borings. They include such items as the

Foundation Engineering in the Face of Uncertainty


exact position of and the variations in width of fissures passing through the rock beneath a dam
foundation, the shape and the local variations of the permeability of minor seams of coarse sand
and gravel contained in fine-grained alluvial valley fills, and similar features of minor geologic
importance.”” Although Terzaghi’’ s (1929) paper is missing from Morgenstern & Cruden’’s (1977)
reference list, it is clear that Terzaghi was identifying geologic and geotechnical complexities as
separate, but related, qualities of dam sites. Terzaghi (1929, p. 31) compares two dam sites that
are determined by investigation to be virtually identical ““–– that is, the difference between the two
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

sites consists exclusively of minor geologic details.”” Terzaghi’’ s point related to the experience of
Dam A being successful to serve as suitable precedence for its design to be copied for Dam B at a
virtually identical site.

Table 5. Hypothetical slope height, H, in model units (mm) and in % of model

height (800 mm) for Gaussian smoothing depicted in Figure 5E, 5F, 5G, and 5H.

Dimensionless spatial Spatial correlation length, T (mm)

correlation, 4c 5 10 20 50
0.5 10 mm 20 mm 40 mm 100 mm
2.0 2.5 mm 5 mm 10 mm 25 mm
0.5 1.25% 2.5% 5% 12.5%
2.0 0.31% 0.63% 1.25% 3.13%

Terzaghi (1929, p. 31) states: ““…… we must start with a clear conception of the physical
factors which are likely to endanger a dam. Then we must translate the terms of the geologist into
terms of physics, and finally we must draw practical conclusions.”” The geologist, as a member of
the design team, must 1) understand the nature of the project being planned, 2) fully appreciate
the geologic setting of the site, including the range of possible details that could adversely affect
project performance, and 3) then search for evidence that the geologic details exist at the site. The
geologist should develop an approach for subsurface investigation that can support either
characterizing the site in terms of the geologic details or a conclusion that they are absent from at
least the critical parts of the site. The geologist’’s understanding, appreciation, and
characterization are of little value unless the results are translated into terms that can be used by
the design team to support practical conclusions for the project.
The RFEM (Griffiths et al., 2012) use of random field theory and spatial correlation length
for key geotechnical parameters requires relevant geologic information to be expressed in terms
that facilitate analytical engineering approaches. The geologist must develop geologic models and
make field observations with RFEM input parameters in mind. Much relevant geologic
information is nonrandom (e.g., stratigraphy); geotechnical parameters throughout a stratigraphic
unit, however, have a range of possible values that may be described with a statistical distribution
(mean and standard deviation). The slope depicted in Figure 6 was used by Griffiths et al. (2012)
to demonstrate the value of three-dimensional finite element stability analyses over two-
dimensional limit equilibrium or finite element analyses. The strength properties in these analyses
were treated deterministically; the 3D finite element analysis produced a factor of safety of about
1.5 with deformation confined to the oblique geologic unit. A series of 2D analyses produced
factors of safety ranging from a maximum of about 2.5 to a minimum of about 1.0; a 2D analysis
of a plane within the oblique geologic unit produced a minimum factor of safety of 0.7.
In this example, Griffiths et al. (2012) called the weak unit ““an oblique layer of weak soil.””
From a geologic perspective, the slope shown in Figure 6 has vertical orientation of bedding or a
vertical dike cutting other formations or deposits. The geology depicted in Figure 6 clearly is not
random. The geology depicted in Figure 5 is not random, either. The level of geologic detail in
Figure 6 is sufficiently simple that Griffiths et al. (2012) treated the two geotechnical units
deterministically, whereas the level of geologic detail in Figure 5 is so complex that

Foundation Engineering in the Face of Uncertainty


simplification clearly would be needed to create a reasonable model for just about any kind of
analysis. The simplification resulting from Gaussian smoothing of the resampled image preserved
some level of banding for spatial correlation lengths of 20 mm or less, but the fracture-filling and
local shearing details were lost at the shortest spatial correlation length of 5 mm (Figure 5E). The
challenge for geologists who are developing Geologic Models for sites on which Ground Models
and Geotechnical Models will utilize RFEM analyses is to anticipate how the project-relevant
geologic information will be configured for input. This anticipation is needed from the initial
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

development of a site Geologic Model through field observation, data reduction, geologic
synthesis and analysis, and participation in developing the Ground Model and Geotechnical
Model. The time has come for engineering geologists to think about geologic formations as soil
and rock masses that have a variety of constituents and qualities, including spatial correlation
lengths, as well as means and standard deviations. If geologists do not provide relevant guidance
on geologically sensible subdivision of formations, then the engineer will be forced to treat
geology as a completely random or unnecessary variable and rely on quantitative field and
laboratory test results as a surrogate for geology.

Figure 6. Slope with a vertical geologic unit used by Griffiths et al. (2012) to
demonstrate the value of three-dimensional finite element analysis compared to
two-dimensional analyses with limit equilibrium and finite element methods.


The title of this paper asks if engineering geology is a fundamental input or a random
variable. Based on a variety of references and logical considerations, it seems clear that
engineering geology is a fundamental input that is inherently random below some threshold level
or characteristic length. Geologists need enhanced understanding of methodologies that will
foster improved field observations and geologic interpretations that capture uncertainty and
variability in ways that can be integrated with analytical methods, as well as guidance on
developing Geologic Models that incorporate concepts of outcrop confidence, quality levels, and
complexity in rigorous, quantifiable ways.
Incomplete or inaccurate geotechnical site characterization can lead to selection of incorrect
models, geotechnical properties, or design values. Project managers responsible for only design
and construction will tend to focus on short-term benefits and may view geologic site
characterization as extra and perhaps unwarranted cost if the benefits are perceived to be marginal
for construction, lead to additional issues that must be resolved in order for a conclusion to be
reached, or pertain to savings with superior performance during long-term operation and

Foundation Engineering in the Face of Uncertainty


maintenance. Terzaghi (1929) realized how important it was to ““start with a clear conception of
the physical factors which are likely to endanger”” a project and then ““translate the terms of the
geologist into terms of physics””. The results of these two elements, conception and translation,
essentially comprise a Geologic Model and perhaps a Ground Model, also. The ““practical
conclusions”” that must be drawn can be associated with the Geotechnical Model.
Varnes (1974, p. 42) recognized that ““As computer technology becomes increasingly
employed in geologic science and operated by specialized personnel, we may find that if the
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

practicing field professional fails to define both his [or her] words and the concepts they
represent, then they may, through necessity, be defined by people whose principal business is the
processing of data.”” Similarly, engineering geologists must transform geologic maps to quantify
uncertainty and variability or risk being marginalized by computer scientists and statisticians who
will translate the geology using random field theory without knowledge of geologic principles.
Morgenstern & Cruden (1977) note that ““If a site were more complex than average one
might take more samples to increase the reliability of the investigation. …… [However,] Instead of
simply expanding the number of borings in order to establish reliably the characteristics of a
geotechnically complex site, the geotechnical engineer must return to geological considerations.
The choice of where to sample, of what to sample and of how to sample has to be guided by an
understanding of the nature of the complexity that is being diagnosed. The most important
contribution to increased reliabi1ity of site characterization of complex conditions comes from an
extra effort associated with geological mapping with the interpretation of the nature of the
geotechnical complexity.”” Lumb (1972 quoted in Kulhawy, 2010) stated, ““Ignorance of soil
behaviour is always regrettable but not necessarily reprehensible, provided that ignorance is
recognised and advice sought where necessary, but ignorance of being ignorant can no longer be
condoned.”” The same can be said for ignorance of engineering geology. Similarly, the geologist
pondering the color of the cow in response to the engineer’’s question might also consider whether
the cow might be represented best as a sphere or a cylinder in a 3D finite-element model.


The basis for this paper originated in 2010 with an invitation for the author to prepare a
keynote presentation representing the Geology and Properties of Earth Materials Section of the
Design and Construction Group of the Transportation Research Board. That presentation, with the
same title as this paper, was delivered at the TRB Annual Meeting in 2011. Modified versions of
it were given at the Geological Society of America Annual Meeting in Minneapolis, MN, in 2011,
at a TRB workshop in 2012, and as the State of the Practice Lecture at the 15th George F. Sowers
Symposium at Georgia Tech in Atlanta in 2012. Versions of the presentation were given to a
graduate course in geological or geotechnical engineering at Colorado School of Mines, Golden,
CO, California State Polytechnic University, Pomona, CA, and University of North Florida,
Jacksonville, FL. Discussions following these presentations led to further enhancements. Review
comments provided by William C. Haneberg and an anonymous reviewer are gratefully


Akbas, S.A. & Kulhawy, F.H. (2010). ““Characterization and Estimation of Geotechnical
Variability in Ankara Clay: A Case History””. Geotechnical & Geological Eng. 28, 619––631.
Anderson, M.P. (1989).”” Hydrogeologic facies models to delineate large-scale spatial trends in
glacial and glaciofluvial sediments””. Geol Society of America Bull, 101(4):501––511.
Anonymous. (1999). ““Definition of Geotechnical Engineering””. Ground Engineering Magazine,
32 (11), 39.

Foundation Engineering in the Face of Uncertainty


Barbour, S.L. & Krahn, J. (2004). ““Numerical Modelling: Prediction or Process?””. Geotechnical
News, 44––52.
Baynes, F.J. (2010). ““Sources of geotechnical risk””. Quarterly Journal of Engineering Geology
and Hydrogeology; 43, 321––331
Baynes, F.J., Fookes, P.G. & Kennedy, J.F. (2005). ““The total engineering geology approach
applied to railways in the Pilbara, Western Australia””. Bull Eng Geol Environ, 64, 67––94.
BOR. (1998). Engineering Geology Field Manual. Bureau of Reclamation, US Department of the
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

Interior, US Government Printing Office.

Burland, J.B. (1987). ““Nash Lecture: The Teaching of Soil Mechanics——A Personal View””.
Groundwater Effects in Geotechnical Engineering, 3. Proc. 9th European Conference on
Soil Mechanics and Foundation Engineering. Balkema, Rotterdam/Boston, 1427––1441.
Chamberlain, T.C. (1890). ““The method of multiple working hypotheses””. Science, 15, 92––96.
Decker, R. & Decker, B. (1992). Road guide to Hawaii Volcanoes National Park.
<> (13 Jun 2012).
FGDC. (2006). FGDC Digital Cartographic Standard for Geologic Map Symbolization. US
Geological Survey Geologic Data Subcommittee, Federal Geographic Data Committee
Document Number FGDC-STD-013-2006, 33 (plus 250 pages of appendices).
FHWA. (2011). Subsurface Utility Engineering. Federal Highway Administration, US Dept. of
Transportation, <> (13 Jun 2012).
Fookes, P.G. (1997). ““Geology for Engineers: the Geological Model, Prediction and
Performance””. Quarterly Journal of Engineering Geology, 30, 293––424.
Glossop, R. (1968). ““Eighth Rankine Lecture: The Rise of Geotechnology and its Influence on
Engineering Practice””. Géotechnique. XVIII (2), 107––150.
Gregory, H.E. & Moore, R.C. (1931). The Kaiparowits region, a geographic and geologic
reconnaissance of parts of Utah and Arizona. U.S. Geological Survey Prof. Paper 164.
Griffiths, D.V., Huang, J. & Fenton, G.A. (2012). ““Risk Assessment in Geotechnical
Engineering: Stability Analysis of Highly Variable Soils””. Geotech Engr State of the Art &
Practice: Keynote Lectures GeoCongress 2012 (GSP 226), ASCE, Reston, VA, 78––101.
Hencher, S. (2012). Practical Engineering Geology. Spon Press, London, UK.
Knill, J. (2003). ““Core Values: The First Hans-Cloos Lecture””. Bulletin of the International
Association for Engineering Geology and the Environment. 62, 1––34.
Kulhawy, F.H. (2010). ““Uncertainty, Reliability, and Foundation Engineering: The 5th Peter
Lumb Lecture””. The Hong Kong Institution of Engineers Transactions, 17(3), 19––24.
Kulhawy, F.H., Phoon, K.K. & Wang, Y. (2012). ““Reliability-Based Design of Foundations-A
Modern View””. Geotechnical Engineering State of the Art and Practice: Keynote Lectures
from GeoCongress 2012 (GSP 226), ASCE, Reston, VA, 102––121.
Lindberg, P.A. (2010). Sedona Sinkholes and Groundwater Flow: The Geologic History of Their
Evolution, Coconino and Yavapai Counties, Arizona. AZ Geol Surv Cont Rpt CR-10-C.
Lumb, P. (1972). ““Soil Engineering in Hong Kong””. Asian Building & Construction. Trend (FE)
Publishing, Hong Kong, 23––26.
Mayne, P.M., Christopher, B.R. & DeJong, J. (2002). Subsurface Investigations - Geotechnical
Site Characterization. US Department of Transportation, Federal Highway Administration
Publication No. FHWA NHI-01-031.
Morgenstern, N.R. (2000). ““Common Ground””. GeoEng2000 –– An International Conference on
Geotechnical and Geological Engineering. Technomic, Lancaster, PA, 1, 1––20.
Morgenstern, N.R. & Cruden, D.M. (1977). ““Description and classification of geotechnical
complexities””. International Symposium on the Geotechnics of Structurally Complex
Formations. Italian Geotechnical Society, 2, 195––204.

Foundation Engineering in the Face of Uncertainty


NRCS. (2002). ““Rock Material Field Classification System””. US Department of Agriculture

Natural Resources Conservation Service, National Engineering Handbook Part 631 Geology,
Chapter 12, <> (13 Jun 2012).
NRCS. (2010). National Engineering Manual, Third Edition. US Department of Agriculture
Natural Resources Conservation Service Online Publication Title M_210_NEM
< > (13 Jun 2012).
NRCS. (2012). ““Engineering Geologic Investigations””. US Department of Agriculture Natural
Downloaded from by Universidad Militar Nueva Granada on 02/21/18. Copyright ASCE. For personal use only; all rights reserved.

Resources Conservation Service, National Engineering Handbook Part 631 Geology,

Chapter 2, <> (13 Jun 2012).
Phoon K.K., Kulhawy F.H. & Grigoriu M.D. (1995). Reliability-based design of foundations for
transmission line structures. Report TR-105000. Electric Power Res Inst, Palo Alto, CA.
Piety, L.A., Anderson, L.W. & Ostenaa, D.A. (2010). ““Late Quaternary Faulting in East Canyon
Valley, Northern Utah””. Paleoseismology of Utah 19, UT Geol Surv Misc Pub 10-5.
Sullivan, T.D. (2010). ““The Geological Model””. Geologically Active –– Proceedings of the 11th
Congress of the International Association for Engineering Geology and the Environment,
Auckland, NZ. Taylor & Francis, London. 155––170.
Terzaghi, K. (1929). ““The Effects of Minor Geologic Details on the Safety of Dams””.
Transactions, American Institute of Mining and Metallurgical Engineers. 215, 31––44.
USACE. (2001). Geotechnical Investigations. US Army Corps of Engineers Engineer Manual
EM 1110-1-1804 <
pdf> (13 Jun 2012).
USGS. (1999). Map Accuracy Standards. US Geological Survey Fact Sheet FS-171-99 < http://> (11 Jun 2012).
Varnes, D.J. (1974). The logic of geological maps, with reference to their interpretation and use
for engineering purposes. U.S. Geological Survey Professional Paper 837.
Whitney, J.A., Jones, L.M. & Walker, R.L. (1976). ““Age and origin of the Stone Mountain
Granite, Lithonia District, Georgia””. Geol Soc Am Bull, 87(7), 1067––1077.

Foundation Engineering in the Face of Uncertainty