This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
 Author's personal copy

Journal of Volcanology and Geothermal Research 178 (2008) 366–377

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j vo l g e o r e s

Ash fallout scenarios at Vesuvius: Numerical simulations and implications for

hazard assessment
G. Macedonio a,⁎, A. Costa b,1, A. Folch a,2
a
Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano, Napoli, Italy
b
Department of Earth Sciences - University of Bristol, UK

a r t i c l e i n f o a b s t r a c t

Article history: Volcanic ash fallout subsequent to a possible renewal of the Vesuvius activity represents a serious threat to
Received 18 February 2008 the highly urbanized area around the volcano. In order to assess the relative hazard we consider three
Accepted 22 August 2008 different possible scenarios such as those following Plinian, Sub-Plinian, and violent Strombolian eruptions.
Available online 14 September 2008
Reference eruptions for each scenario are similar to the 79 AD (Pompeii), the 1631 AD (or 472 AD) and the
1944 AD Vesuvius events, respectively. Fallout deposits for the first two scenarios are modeled using
Keywords:
volcanic ash fallout
HAZMAP, a model based on a semi-analytical solution of the 2D advection–diffusion–sedimentation
volcanic hazard equation. In contrast, fallout following a violent Strombolian event is modeled by means of FALL3D, a
computer model numerical model based on the solution of the full 3D advection–diffusion–sedimentation equation which is
Vesuvius valid also within the atmospheric boundary layer. Inputs for models are total erupted mass, eruption column
height, bulk grain-size, bulk component distribution, and a statistical set of wind profiles obtained by the
NCEP/NCAR re-analysis. We computed ground load probability maps for different ash loadings. In the case of
a Sub-Plinian scenario, the most representative tephra loading maps in 16 cardinal directions were also
calculated. The probability maps obtained for the different scenarios are aimed to give support to the risk
mitigation strategies.
© 2008 Elsevier B.V. All rights reserved.

1. Introduction
Volcanic ash fallout constitutes a serious hazards to communities
settled around active explosive volcanoes. The assessment of such
hazards are, consequently, a matter of importance for public safety in
volcanic regions like the highly urbanized area around Vesuvius,
inhabited by millions of people. Hazards associated with a possible
renewal of explosive activity at Vesuvius have been a subject of several
studies (e.g., Barberi et al., 1990; Macedonio et al., 1990; Cioni et al.,
2003). However, all previous studies assessed hazards for a Sub-Plinian
scenario only and did not consider other higher and lower magnitude
events. The most severe ash fallout hazard associated with Plinian and
Sub-Plinian eruptions, further subdivided in Sub-Plinian I and Sub-
Plinian II by Cioni et al. (2008-this issue), is due to the collapse of roofs
because of ash loading that potentially involves areas up to few
hundreds of km2. Effects and damage associated with a violent
⁎ Corresponding author.
E-mail addresses: macedon@ov.ingv.it (G. Macedonio), costa@ov.ingv.it (A. Costa),
afolch@ov.ingv.it (A. Folch).
1
Now at Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano,
Napoli, Italy.
2
Now at Earth Science Department, Barcelona Supercomputing Center, Centro
Nacional de Supercomputación, Spain.
Strombolian eruption are less severe from those of a Plinian or a Sub-
Plinian scenario. However, several eruptive episodes at Etna and other
volcanoes (e.g., 1944 AD Vesuvius) have shown that eruptions of lower
magnitude can also create serious problems to local communities,
especially in very urbanized areas (such as the Neapolitan area). Beside
possible roof collapses by ash loading in limited areas close to the
volcano, volcanic ash can also cause, even for minor events, severe
disruption of transportation systems due to loss of visibility, disruption
to communications due to interference to radio waves or direct damage
to communications, temporary shut down of airports and aerial
corridors, partial or total destruction of agricultural crops and damage
to forestry, irritation of eyes and skin and potential respiratory
symptoms produced by ash inhalation. Our goal here is to investigate
ash fallout in the Vesuvius area for three different scenarios representa-
tive of a Plinian, a Sub-Plinian I, and a violent Strombolian eruptions
respectively using as a reference well-documented eruptions such as
those occurred in 79 AD, 1631 AD, 472 AD, and 1944 AD.
The Sub-Plinian II category, characterized by a lower intensity and
Magnitude than Sub-Plinian I (Cioni et al., 2008), was not investigated
in the present study.
Among the Plininan eruptions, the 79 AD Vesuvius eruption is
surely one of the most extensively studied eruptions ever. It is famous
for the destruction of the Roman towns of Pompeii and Herculaneum
and for the detailed chronicles written by the Roman lawyer Pliny the

0377-0273/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jvolgeores.2008.08.014
 Author's personal copy
G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377
Younger who described the events in which his uncle, Pliny the Elder,
died. The volcanological aspects of the 79 AD eruption have been
described by numerous authors (e.g. Lirer et al., 1973; Santacroce,
1983; Sigurdsson et al., 1985; Carey and Sigurdsson, 1987; Santacroce,
1987; Lirer et al., 1993; Scandone et al., 1993). The eruption comprised
several phases. The graded Plinian fallout deposit associated with the
purely magmatic phase, with a high sustained eruption column
between 15 and 26 km (Carey and Sigurdsson, 1987), shows a lower
layer of white phonolitic pumice (“White pumice”) overlaid by an
upper layer of tephritic-phonolitic pumice (“Gray pumice”). This
eruptive phase ended with the collapse of the eruptive column and
the subsequent emplacement of pyroclastic flows.
Two Sub-Plinian I eruptions are considered here (those occurred in
1631 and 472 AD). The eruption occurred in 1631 AD started on 16
December 1631 and produced a sustained column for about 8 h with
an average height of 16 km, whereas the eruption occurred in 472 AD,
also known as the “Pollena” eruption (Santacroce, 1983; Rosi and
Santacroce, 1983; Mastrolorenzo et al., 2002; Principe et al., 2004;
Rolandi et al., 2004; Cioni et al., 2008-this issue), produces a sustained
column between 12 and 20 km.
Finally, the violent Strombolian eruption occurred in 1944 AD and
was the last eruptive event that occurred at Vesuvius. The eruptive
phase started with a sequence of lava flows followed by fire-
fountaining episodes which evolved into a convecting eruptive
column sustained for few days ending up with a phase of Vulcanian
explosions (Imbò, 1949) with maximum column height of 5 km.
For each scenario we computed ground ash load probability maps
which can provide potential decision-support to the hazard assess-
ment for future eruptions at Vesuvius.
2. Models for volcanic ash transport and fallout
The ash fallout models we apply consider only the transport and
deposition of lapilli and coarse ash, i.e., tephra particles from few mms to
several microns in diameter (from Φ≃−5 to Φ≃4). Beyond a certain
distance from the eruption column, the dispersion and sedimentation of
tephra is governed mainly by wind transport, turbulent diffusion, and
settling of particles by gravity (Armienti et al., 1988; Macedonio et al.,
1988). Particles with a similar dynamic regime are commonly grouped
into families or classes which, in general, can have interactions among
them. Aggregation of finer particles during their fall (e.g. Cornell et al.,
1983) is the most obvious evidence for particle interaction. However, if
the concentration of particles is sufficiently dilute one can neglect
interaction among particles belonging to different classes. Under this
hypothesis the concentration of particles of class j is described through
the mass conservation equation as (Costa et al., 2006):
ACj ACj ACj
 ACj AVsj
þ UX þ UY þ UZ −Vsj ¼ −Cj j  U þ Cj
At AX AY AZ AZ
     
A ACj =ρ A ACj =ρ A ACj =ρ
ρKX þ ρKY þ ρKZ þ Sj
AX AX AY AY AZ AZ
where Cj denotes concentration of particle class j, t is time, (UX,UY,UZ)
are the components of the wind velocity vector, KX,KY and KZ are the
turbulent diffusion coefficients (diagonal terms of the turbulent
diffusion tensor), ρ is the atmospheric density, and Vsj and Sj stand,
respectively, for the terminal settling velocity and source term for
class j. The advection–diffusion–sedimentation (ADS) Eq. (1), is of
general applicability and must be solved numerically. However, under
certain additional hypothesis, it can be simplified to derive analytical
or semi-analytical solutions.
2.1. A model for high eruptive column: Plinian and Sub-Plinian scenarios
For large eruptive columns most of the ash transport process
occurs outside the Atmospheric Boundary Layer (ABL). Consequently,
367
in a first-order approach, it can be assumed that the vertical diffusion
component is negligible with respect to the horizontal ones and that
the terrain effects and the vertical wind component become a second-
order effect. In addition, if one assumes also that the horizontal wind
components are constant in time and uniform along the horizontal
domain, Eq. (1) simplifies to:
ACj ACj ACj AVsj Cj A2 Cj A2 Cj
þ Ux þ Uy − ¼ Kx 2 þ Ky 2 þ Sj : ð2Þ
At Ax Ay Az Ax Ay
The model HAZMAP (Macedonio et al., 2005) we used for the
simulations is based on a semi-analytical solution of the simplified
Eq. (2). Dividing the vertical computational domain into Nlayer layers in
which settling and wind velocity are assumed constant, the total mass
on the ground MG can be computed as the sum of the contributions
from each of the point sources distributed above the vent and from
each particle settling velocity class:
" #
Nvs Nsources
Mi fj ðx−xGi Þ2 þð y−yGi Þ2
MG ðx; yÞ ¼ ∑ ∑ exp − ð3Þ
j¼1 i¼1 2πσ 2Gi 2σ 2Gi
→
where xYGi ¼ Yx 0i þ ∑k Uk Δtk and σ2Gi = 2KΣkΔtk are the center and the
variance of the Gaussian respectively (Δtk = (zk − zk − 1)/vs,k is the time
for a particle to cross the layer k), Nsources indicates the number of
source points, Nvs is the total number of settling velocity classes, Mi is
the total mass emitted from the point source in the layer i (ΣiMi = Mtot,
with Mtot total mass injected into the system), and fj is the fraction of
that mass belonging to the settling velocity class j (Σjfj = 1). In accord to
Macedonio et al. (2005) and Pfeiffer et al. (2005), the source term for
the Plinian and Sub-Plinian I eruptions was described using a modified
parameterization proposed by Suzuki (1983).
2.2. A model for low eruptive column: the violent Strombolian scenario
For low eruptive columns (such the ones produced, for instance,
during violent Strombolian events) most of the simplifying assump-
tions made to reduce Eqs. (1) to (2) are not longer valid because
significant transport of ash occurs inside the ABL. In fact, wind fields
and turbulent tensor components inside the ABL are much complex
and terrain effects are, in general, not negligible. Moreover effects of
bent-over on the plume due to the wind are not negligible. On the
other hand, since winds in the lower part of the atmosphere vary
rapidly in time and in space, there is a strong need to forecast the
temporal evolution of both airborne ash concentration and ash
deposit on the ground. It follows that simplified steady semi-
analytical models like HAZMAP are not fully adequate in this context.
For this reason we make use of the model FALL3D (Costa et al., 2006)
to simulate ash fallout for the case of the violent Strombolian scenario.
ð1Þ
This model adopts realistic near surface wind fields that account for
terrain effects, uses a realistic evaluation of the turbulent atmospheric
diffusion based on the Monin–Obukhov similarity theory and a Large
Eddy approach, and describe the source term using a model based on
the Buoyant Plume Theory (BPT) as in Bursik (2001). For further
details about the model see Costa et al. (2006). Obviously, Eq. (1)
needs to be solved numerically and, in consequence, the model
becomes less suitable to perform statistical studies and/or to solve
inverse problems because the computational requirements become
greater.
3. Ash fallout simulations and hazard maps
In this Section we describe the volcanological and meteorological
data used for the analysis of ash fallout hazard from explosive
eruptions at Vesuvius and we show the obtained hazard maps for the
different scenarios.
 Author's personal copy

368 G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377

3.1. Volcanological input and reference eruptions Table 1

Parameters used to obtain ground load probability maps for the three different scenarios

According to Cioni et al. (2008-this issue), explosive eruptions are Plinian Sub-Plinian Violent Strombolian
classified into five different eruption types: Plinian, Sub-Plinian Model HAZMAP HAZMAP FALL3D
(further subdivided into Sub-Plinian I and Sub-Plinian II), violent Ground domain (km) 70 × 70 70 × 70 70 × 70
Strombolian, Ash emission events and phreatic explosions. In this Vent coordinates (UTM) (451737;4519302) (451737;4519302) (451737;4519302)
Vent elevation (m) –a –a 1250
paper we focus on the three major types: Plinian, Sub-Plinian
Number of velocity 10 10 5
(restricting the study to Sub-Plinian I type) and violent Strombolian. classes
In the following, we summarize the different input parameters used in Bulk grain-size PFb c
µ = −1; σ = 1c
the simulations and the strategy adopted for the computation of the population
Vesuvius ash loading probability maps, which include: Diffusion coefficients 5000 5000 LE + STd
(m2/s)
Average column height 27e 18f 4g,h
(1) Definition of three reference eruptions for Plinian, Sub-Plinian I (km)
and violent Strombolian scenarios based on the study of the Column shape A = 4, λ = 1.5 A = 3, λ = 1 BPTh
past behavior of Vesuvius. parameters
Average mass flow rate 8 × 107 3 × 107 5 × 105
(2) Both computational models (HAZMAP and FALL3D) were
(kg/s)i
validated by simulating the fallout deposit of three historical Duration (h)j 7 4.5 110
events such as the 79 AD, the 1631 AD and the 1944 AD Total mass (kg)k 2 × 1012 5 × 1011 2 × 1011
eruption. Results allowed also for estimating the typical ranges Meteorological setl NCEP/NCAR NCEP/NCAR NCEP/NCAR
of the input parameters (i.e. total mass associated with the a
Model not accounting for topography.
fallout phase, eruption column height, bulk particle settling b
Pumice Flow (PF) (Macedonio et al., 1988).
c
velocity distribution, column shape parameters, the diffusion Bulk grain size distributions reported in Fig. 2 obtained in accord to Cioni et al. (data
presented at the Exploris Final Workshop, Naples, May 10–13, 2006).
coefficient) associated with the Plinian, Sub-Plinian I and d
Not an input parameter. Computed by the model using large Eddy (LE) approach for
Violent Strombolian scenario respectively. horizontal component and Similarity Theory (ST) for vertical component (Costa et al.,
(3) Input parameters for the Plinian scenario were chosen as those 2006).
e
derived through best-fitting the 79 AD eruption simulated by Average between “White” and “Grey” phase values from RUN V5 in Pfeiffer et al.
using HAZMAP by (Pfeiffer et al., 2005). Input parameters for (2005).
f
Value obtained from best fit field deposit data of Cioni et al. (2003).
the Sub-Plinian I scenario were chosen similar to those g
Observations from Imbò (1949).
obtained through a best-fit of the AD 1631 eruption deposits, h
Calculated using the BPT (Bursik, 2001) coupled with the recorded wind.
although for the mass and the column height we cautiously i
Estimated using BPT (Sparks, 1986; Costa et al., 2006) to reproduce column height.
j
considered larger values similar to those estimated for the 472 For the Plinian scenario an effective duration of 7 h was considered in order to generate
an erupted mass of 2 × 1012 kg (relative to the coarser ashes only), whereas a duration of
AD eruption (Neri et al., 2008-this issue; Cioni et al., 2008-this
about 14 h is obtained considering a total mass equal to 4 × 1012 kg (see text for details).
issue). For the violent Strombolian scenario, simulated by using k
From the solution of the inverse problem reproducing deposits.
FALL3D, the controlling parameters given by the best fit of the l
Wind profiles from NCEP/NCAR re-analysis for the period 1968–2003 (Kalnay et al.,
1944 AD eruption were considered as reference values. In fact a 1996).
semi-quantitative reconstruction of the 1944 AD event was
possible thanks to the volcanological and meteorological data associated with the proximal deposit only (within a radius smaller than
reported by Imbò (1949), wind intensity measured at the 50 km). Solving the inverse problem for the ash deposits Pfeiffer et al.
Osservatorio Vesuviano, and information derived from dated (2005) estimated this mass to be around 2 × 1012 kg. This mass is
photos taken during and shortly after the eruption. A summary attributed to coarser ashes mainly deposited in the proximal zone where
of the main features of the fallout phases of these scenarios and the contribution of fine ashes is almost negligible (Pfeiffer et al., 2005).
other important parameters are reported in Table 1. However, in the bulk grain size distribution of Plinian and Sub-Plinian
(4) Computation of the ash loading probability maps running the eruption column, up to 50% of the mass can be attributed to fine ashes
models with the parameters obtained from the reference (Φ N 3), as it is shown in Fig. 2 (data presented by Cioni et al. at the
eruptions and a 36-year winds profiles data set (Fig. 1). Exploris Final Workshop, Naples, May 10–13, 2006) and indicated by the
solution of the inverse problem we solved in reconstructing the deposit
3.1.1. The Plinian eruption scenario of Vesuvius 1631 AD eruption (see Section 2). Accounting for that, a total
Simulations for the Plinian scenario are based on the best-fit mass of about 4 × 1012 kg can be associated to the Plinian phase of the 79
parameters obtained by Pfeiffer et al. (2005) (RUN V5). The total mass AD eruption. Here, we use the values obtained by Pfeiffer et al. (2005)
was assumed equal to the sum of those relating to the “White Phase” and where a coarser granulometry was considered. However fine ash mass is
“Gray Phase” deposits of the 79 AD eruption. According to Carey and not very relevant in terms of ground loading because most of it was
Sigurdsson (1987), the column increased from about 15 to 26 km during deposited outside of the computational domain we used (N35 km from
the White Pumice phase, then it reached a maximum of 32 km and the vent). The estimates column height obtained by Pfeiffer et al. (2005)
finally decreased to about 27 km during the Gray Pumice phase. The range between ca. 21 km for the White Phase and ca. 33 km for the Gray
grain size distribution of pumice deposited at the end of the Gray pumice Phase, with a duration of the fallout phases of about 7 and 11 h
phase reported in Macedonio et al. (1988) was taken as input for the respectively. Here, for the entire Plinian phase (about 18 h) we used a
eruption column grain size distribution. The remaining parameters were weighted average column height of 27 km (Pfeiffer et al., 2005). Using the
set equal to the weighted average of the best parameters of each single Buoyant Plume Theory (BPT, e.g., Sparks, 1986; Costa et al., 2006) with
phase (Pfeiffer et al., 2005) considering a weight proportional to the mass the parameters of Table 1, we can reproduce a column height of 27 km
itself. The mass used here for assessing hazard from ash loading is that using a Mass Flow Rate (MFR) of ~8 × 107 kg/s. For a duration of 18 h this

Fig. 1. Wind distribution diagrams corresponding to the point of the NCEP/NCAR global mesh nearest to Naples (40°N, 15°E), for 36 years in the period 1968–2003 (after Costa et al.,
2008). Horizontal wind vectors (intensity and direction) are reported at the different altitudes (30, 20, 10, and 5 km respectively). Blue corresponds to winter (January, February,
March) winds, green to spring (April, May, June) winds, red to summer (July, August, September) winds, yellow to autumn (October, November, December) winds. Magenta
corresponds to whole year winds, i.e., those used to obtain the probability maps. Black arrows denote the average values. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)
 Author's personal copy

G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377 369
 Author's personal copy
370 G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377
Fig. 2. Bulk grain size distribution considered as representative for a Plinian/Sub-Plinian
eruption column (on the top) and for a violent Strombolian eruption column (on the
bottom) respectively. Bulk values of 80% pumices and 20% lithics/crystal were considered.
means a total mass of about 5 × 1012 kg which is comparable with the
value we found through best fit considering 50% of fine ash mass.
3.1.2. The Sub-Plinian I eruption scenario
Input parameters for the Sub-Plinian I scenario were found
through best fitting field deposits data from Cioni et al. (2003). In
this case since we do not have direct information on the grain size
distribution inside the column, the bulk grain size distribution chosen
as representative of a Sub-Plinian I eruption is that of Fig. 2 (in accord
to Cioni et al. (2008-this issue) and reported by Cioni et al. at the
Exploris Final Workshop, Naples, May 10–13, 2006). Results of the
fitting procedure are shown in Fig. 3. The total mass of the 1631 AD
eruption was estimated to be about 2 × 1011 kg with an average column
height of about 16 km.
Using the BPT (e.g., Sparks, 1986; Bursik, 2001) with these values, we
can reproduce a column height of 16 km with a MFR of ~2 × 107 kg/s.
The above parameters were obtained through a minimization
procedure which accounted for the best fit of the total final deposits
and, at the same time, of deposits associated with both “Black” and
“White” phases. In particular, the inverse problem we solved for the
single phases indicated that column height associated with the initial
“Black” phase was ≳ 20 km with a total mass of ~ 1.5 × 1011 kg whereas
the height associated with the “White” phase was ≲ 10 km with a total
mass of ~ 0.5 × 1011 kg.
Moreover, using the BPT model we obtained a mass flow rate of
~ 4 × 107 kg/s for the “Black” phase and ~ 3 × 106 kg/s for the “White”
phase. Considering a total mass of ~ 1.5 × 1011 kg for the “Black” phase
and ~0.5 × 1011 kg for the “White” phase, we have a duration of about
1–3 h for the “Black” phase and 5–7 h for the “White” phase. These
values are comparable with the total mass of 1.8 × 1011 kg (0.07 km3
DRE) estimated from isopachs of the deposits by Rosi et al. (1993).
The 472 AD (“Pollena”) eruption, also Sub-Plinian I, produced a
column with a height between 12 and 20 km (Sulpizio et al., 2005).
However, there are different estimations about the volume of the
deposit of the Pollena eruption, such as: 0.32 km3 (Rosi and
Santacroce, 1983), 0.5 km3 (Mastrolorenzo et al., 2002), 0.42 km3
(Cioni et al., 2003), 1.2 km3 (Rolandi et al., 2004), and 1.38 km3
(Sulpizio et al., 2005), corresponding, respectively, to a mass of
2.9 × 1011, 4.5 × 1011, 3.8 × 1011, 1.1 × 1012 and 1.2 × 1012 kg (assuming a
mean deposit density of 900 kg/m3 as reported in Cioni et al. (2003)).
Here, following Cioni et al. (2008-this issue), for the mass of the
Sub-Plinian I scenario (see Table 1), we adopt the average value of
5 × 1011 kg which corresponds to a deposit of about 0.5 km3.
Moreover, a column height of 18 km was adopted for Sub-Plinian I
eruptions. Scenarios corresponding to greater volumes, such as those
suggested by Sulpizio et al. (2005), are implicitly considered in the
Plinian scenario.
3.1.3. The violent Strombolian scenario
The explosive phase of the 1944 AD eruption started during the
night of 21 March. Following a series of fire-fountaining episodes, the
eruption intensity increased to form a convecting eruptive column. A
maximum column height of about 6 km was reached around 1 p.m.
(local time) on 22 March. Eruptive columns from 2 to 4 km were
produced in the following days (from 23 to 26 March). In order to
reconstruct the wind field during this period we used records from a
ground station located at the Osservatorio Vesuviano historical
building (in Herculaneum) as well as a “pseudo-sounding” at the
same location. Upper air winds were constrained by Imbo's descrip-
tion (Imbò, 1949). On March 23 upper wind intensity varied from
about 17 m/s at 5 km to 11 m/s at 1.3 km and had a NW direction.
Lower wind intensity was around 4 m/s at 1 km blowing from W
direction. A sudden increase of wind intensity (of about 10–15 m/s)
and change in direction (to NNE) were observed on March 25. In order
to reconstruct the 1944 event, measured winds were spatially
assimilated on a fine scale (accounting for topographic effects) and
temporally interpolated by using the meteorological processor
CALMET (Scire et al., 2000). The mass of the explosive phase estimated
from the deposits is about 2.5 × 1011 kg (Cioni et al., 2003), in
agreement with the volume of violent Strombolian eruptions of the
order of 108 m3, or less, reported by Cioni et al. (2008-this issue) and
the column height reached a maximum of 5 km (Cioni et al., 2003).
Using a buoyant plume theory which accounts for plume bent over
due to wind (Bursik, 2001), essential for low eruption columns, we
came out with an average mass flow rate of 5 × 105 kg/s (in agreement
with Cioni et al., 2008-this issue), in order to reproduce the observed
average column height of about 4 km. Considering this average mass
flow rate and the duration of the explosive phase (about 120 h) the
erupted mass would be 2.16 × 1011 kg, a value in good agreement with
that reported by Cioni et al. (2003).
Bulk granulometry reported in Fig. 2 is used as representative of
grain-size distribution of a violent Strombolian eruption column.
The rest of the input parameters were evaluated through a semi-
quantitative reconstruction of the event by using FALL3D model.
Simulation results show good agreement with observations (see
Fig. 4). In fact, besides the excellent reproduction of the total mass, this
model was able to reconstruct satisfactorily the observed deposits.
The area enclosed by the 10 cm isopach is about 180 km2, close to the
270 km2 reported in Cioni et al. (2003). Some deposit values observed
were from 30 to 100 cm at Terzigno, 10 cm at Angri and Pagani, and
3 cm at Meta di Sorrento (Imbò, 1949). These values are in good
agreement with the simulation, which gives 74.3 cm at Terzigno,
 Author's personal copy

G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377
Fig. 3. Reconstruction of the 1631 deposit (on the left) using HAZMAP and the input values reported in Table 1. Plot shows deposit thickness contours (in cm). Comparison between the computed and observed values (Cioni et al., 2003) is also
shown on the right.

371
 Author's personal copy
372 G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377
Fig. 4. Reconstruction of the 1944 event using FALL3D and the input values reported in Table 1. Plot shows deposit thickness contours (in cm) after 5 days. The asymmetric contours
pointing S are a consequence of the abrupt change in wind direction registered on March 25, during the end of the explosive phase. For comparison, values of 30–100 cm at Terzigno,
10 cm at Angri and Pagani, and 3 cm at Meta di Sorrento were reported after fallout.
13.5 cm at Angri, 9.6 cm at Pagani, and 2.6 cm at Meta di Sorrento
(Fig. 4).
3.2. Meteorological data
Meteorological data (e.g. wind velocity profiles) are needed to
simulate ash dispersal from volcanic eruptions. When the wind
profiles were not available, such as for the reconstruction of the AD 79
and 1631 eruptions, wind intensity and direction were obtained from
best-fitting the deposit, whereas, for the reconstruction of the 1944
eruption, we used wind profiles from Imbò (1949) and ground wind
measured at the historical building of the Osservatorio Vesuviano
(Herculaneum) during the eruption.
For the statistical analysis of the ash dispersal from possible
eruptions at Vesuvius, the wind dataset was extracted from the daily-
average wind profiles corresponding to the point of the NCEP/NCAR
global mesh (Kalnay et al., 1996) nearest to Naples (node with
latitude = 40°N and longitude = 15°E, about 100 km south-east of
Vesuvius) as given by the NCEP/NCAR re-analysis (http://www.cdc.
noaa.gov/cdc/reanalysis/) for the period 1968 to 2003. It comprises
36 years (in the period between 1968 and 2003) of data with daily
values at 17 pressure levels. These levels correspond approximately to
vertical levels between ground and 32 km in altitude. In this work, we
interpolated the pressure levels every 500 m in altitude. The statistics
of wind profiles are shown in Fig. 1. It can be observed how, high in the
atmosphere (i.e., above 10 km), winds mainly blow in the direction
East-West with a season-dependent intensity, whereas close to the
ABL (i.e., below 10 km) directions are less anisotropic and the seasonal
influence becomes weaker.
3.3. Hazard maps
The hazard associated to ash fallout is related to the scale of the
eruption. Here, we consider three different scenarios, representative
of the Plinian, Sub-Plinian I and violent Strombolian eruptions
occurred at Vesuvius in the last 20 ka (see e.g. Cioni et al., 2008-this
issue). According to Marzocchi et al. (2004), the occurrence of each
scenario has a different probability of occurrence, depending on the
magnitude of the eruption and the repose time preceding the
eruption. In particular, for a repose time between 60 and 200 years,
Mount Vesuvius shows a probability of an eruption of small scale
(VEI = 3 or violent Strombolian) of 72%. This probability decreases to
27% for medium scale eruptions (VEI = 4 or Sub-Plinian I) and to 1% for
large scale eruptions (VEI = 5 or Plinian). When the repose is increased
(repose time N60 years), the probability of larger eruptions increases,
being respectively 65%, 24% and 11% for eruptions respectively of
VEI = 3, VEI = 4 and VEI = 5 (Marzocchi et al., 2004). These probabilities
are similar to those estimated by Neri et al. (2008-this issue), who
estimated “median” probabilities of about 4%, 17%, 28% and 38%, for
Plinian, Sub-Plinian I, Sub-Plinian II and violent Strombolian eruptions
respectively. For each eruption scenario, several ash fallout simula-
tions, corresponding to different wind profiles, were performed. Fig. 5
shows the ground load probability maps obtained for the Plinian
scenario for loadings able to produce significant roof collapses (300
and 400 kg/m2). As is clear from the probability maps presented, a
scenario like that would be highly destructive and involve a very large
urbanized area (Cioni et al., 2008-this issue; Neri et al., 2008-this
issue). However, following Neri et al. (2008-this issue), this scenario
would have a “median” probability of 4% to occur. As mentioned
above, the low probability of occurrence of this event is confirmed by
the independent estimates of Marzocchi et al. (2004) who give a
probability of occurrence of this scenario of 11% for a repose time
greater than 60 years (between 60 years and infinity), and less than 1%
if a time window between 60 years and 200 years is considered.
Probability maps for the Plinian and Sub-Plinian scenarios have
been obtained using 36 years of daily winds extracted from the Kalnay
et al. (1996) NCEP/NCAR Reanalysis I database (i.e., 365 (or 366) days
times 36 years in the period 1968–2003 = 13149 HAZMAP runs). In
 Author's personal copy

G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377 373

Fig. 5. Ash loading probability maps for Plinian scenario. Probability is normalized to 100%. top) 300 kg/m2, bottom) 400 kg/m2.

contrast, meteorological data for the violent Strombolian scenario has
been simplified due to computational limitations. FALL3D simulations
have been done only for a “representative” year period, obtained by
averaging the daily winds (day by day) in the 36-year period. For all
the three scenarios we have computed ash loading probability maps
for several thresholds values from 100 to 1500 kg/m2. Here we present
only maps for loading thresholds of 300 and 400 kg/m2. These values are
considered critical for roof collapse of low and medium quality buildings
respectively at the Neapolitan area (Zuccaro et al., 2008-this issue).
Fig. 6 shows the ground load probability maps obtained for the
Sub-Plinian I scenario for loadings able to produce significant roof
collapses (300 and 400 kg/m2). A scenario like that involves a wide
 Author's personal copy

374 G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377

Fig. 6. Ash loading probability maps for Sub-Plinian I scenario. Probability is normalized to 100%. top) 300 kg/m2, bottom) 400 kg/m2.

area where more than one million of people live nowadays. The
probability of occurrence of this scenario was estimated by Marzocchi
et al. (2004) in the range 25–30%, whereas, the independent estimates
by Neri et al. (2008-this issue) give a probability of occurrence of the
Sub-Plinian I scenario in the range 1–49%, with the average value of
17%. This scenario was chosen as the reference scenario for emergency
planning at Vesuvius.
Fig. 7 shows the ground load probability maps obtained for the
violent Strombolian scenario. As we mentioned before, temporal
variations of the meteorological conditions for this scenario are much
 Author's personal copy

G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377 375

Fig. 7. Ash loading probability maps for violent Strombolian scenario. Probability is normalized to 100%. top) 300 kg/m2, bottom) 400 kg/m2.

more important because of both the lower eruptive column and the
longer eruption duration. In order to capture these features simula-
tions for hazard assessment have been done using an “average”
meteorological year obtained by the 36-years average daily winds and
considering an eruption which starts every 5 days with a duration of
5 days. Note that this approach is not consistent with the previous
methodology for hazard assessment. In fact we have not performed
13149 simulations relative to the 36 years but the probability maps in
this case were made considering only 73 (365/5) final outputs.
As for the previous cases, in Fig. 7 we report the ground load
probability maps obtained for the violent Strombolian scenario for
loadings of 300 and 400 kg/m2. As previously discussed this is the
 Author's personal copy

376 G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377

Fig. 8. Simulations of ash loading at the ground using 16 different wind profiles representative of 16 radial sectors. Winds are extracted from the NCEP/NCAR Reanalysis I database for
the period 1968–2003 in the Neapolitan area. Among all the possible winds in the period, for each sector, the wind that produces a deposit with its center of mass as near as possible
to the center of the sector and average distance from the crater was selected. Inner lines correspond to ash loading of 400 kg/m2, whereas outer lines correspond to 300 kg/m2. The
probability of wind direction that produces a deposit with center of mass inside each sector is also indicated.

most probable scenario. However, these results show that the area
subject to roof collapse in this case is encompassed by the hazard area
identified by the Sub-Plinian I reference scenario.
Beside the probability maps of ash loading, in Fig. 8 we present a
complementary representation of the hazard zone for the Sub-Plinian
I scenario. In Fig. 8, 16 selected ash fall simulated deposits, subject to
different winds, are shown. The winds are extracted for the daily
averages of the NCEP/NCAR Reanalysis I database (Kalnay et al., 1996)
in the period 1968–2003. In particular, for all the daily winds, a
simulation of the ash loading at the ground was performed, and the
position of the center of mass of the deposit was calculated and
associated with one of the 16 sectors (each 22.5° wide) radiating from
the crater. For each sector, the average distance and direction of the
center of mass of the deposits was calculated, and the wind that
reproduced a deposit as near as possible to this average was selected.
Then, the 16 selected winds were used to simulate possible deposits in
different directions. This representation of the hazard zone is
complementary to Fig. 6, but is preferable because it allows the
estimation of the effects of the eruption subject to a single wind.
Moreover, in Fig. 8, the probability that a wind produces a deposit with
the center of mass in each sector is also shown.
4. Summary and conclusions
We have computed ground load probability maps for a Plinian,
Sub-Plinian and violent Strombolian scenario likely to occur in case of
a renewal of activity at Vesuvius. A NCEP/NCAR re-analysis wind
dataset composed by 36 years (between 1968 and 2003) of daily
averaged values was used to model ash fallout with the HAZMAP and
FALL3D codes. Results show that the least likely scenario – a Plinian
eruption similar to the 79 AD event – would be catastrophic in the
densely inhabited Neapolitan area. The most probable scenario
associated with a violent Strombolian eruption similar to the 1944
AD event, involves a wide area around Vesuvius, and it is comfortably
included within the hazard area defined by a Sub-Plinian scenario
similar to the 472 AD event. In addition, for the Sub-Plinian scenario a
complementary representation of the hazard effects describing
average ash loadings along 16 cardinal wind sectors, which will
serve to inform hazard mitigation planning, is reported.
Acknowledgments
This work was supported by the “EU Research Project EXPLORIS”
(EVR1-CT-2002-40026) and by the Italian Civil Protection Depart-
ment. We thank G. Ricciardi from the INGV-OV for furnishing
meteorological ground station records during the 1944 AD eruption.
The authors are grateful to the scientists of the Exploris consortium
and, in particular, to the coordinator of the Project A. Neri for helpful
discussions and suggestions. Finally, we wish to thank Willy Aspinall
and an anonymous reviewer for improving the quality of the
manuscript.
References
Armienti, P., Macedonio, G., Pareschi, M., 1988. A numerical model for the simulation of
tephra transport and deposition: applications to May 18, 1980 Mt. St. Helens
eruption. J. Geophys. Res. 93 (B6), 6463–6476.
Barberi, F., Macedonio, G., Pareschi, M., Santacroce, R., 1990. Mapping the tephra fallout
risk: an example from Vesuvius (Italy). Nature 344, 142–144.
Bursik, M., 2001. Effect of wind on the rise height of volcanic plumes. Geophys. Res. Lett.
18, 3621–3624.
Carey, S., Sigurdsson, H., 1987. Temporal variations in column height and magma
discharge rate during the 79 AD eruption of Vesuvius. Geol. Soc. Am. Bull. 99,
303–314.
 Author's personal copy
G. Macedonio et al. / Journal of Volcanology and Geothermal Research 178 (2008) 366–377
Cioni, R., Longo, A., Macedonio, G., Santacroce, R., Sbrana, A., Sulpizio, D., Andronico, D.,
2003. Assessing pyroclastic fall hazard through field data and numerical
simulations: example from Vesuvius. J. Geophys. Res. 108 (B2), 2063–2073.
Cioni, R., Bertagnini, A., Santacroce, R., Andronico, D., 2008. Explosive a activity and
eruption scenarios at Somma–Vesuvius (Italy): a review. J. Volcanol. Geotherm. Res.
178, 331–346 (this issue).
Cornell, W., Carey, S., Sigurdsson, H., 1983. Computer simulation and transport of the
Campanian Y-5 ash. J. Volcanol. Geotherm. Res. 17, 89–109.
Costa, A., Macedonio, G., Folch, A., 2006. A three-dimensional Eulerian model for
transport and deposition of volcanic ashes. Earth Planet. Sci. Lett. 241, 634–647.
Costa, A., Dell'Erba, F., Di Vitoi, M., Isaiai, R.G.M., Orsi, G., Pfeiffer, T., 2008. Assessment of
the volcanic ash loading hazard from Phlegrean Fields Caldera (Italy). Bull. Volcanol.
66, 703–724. doi:10.1007/s00445-008-0220-3.
Imbò, G., 1949. Annali dell'Osservatorio Vesuviano - Volume unico celebrativo del I
centenario dell'Osservatorio Vesuviano. Stabilimento Tipografico G. Genovese,
Napoli, Italy.
Kalnay, E., Kanamitsu, M., Kister, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha,
S.G.W., Woollen, J., Zhu, Y., Chelliah, M., Ebisuzaki, M., Higgins, W., Janowiak, J., Mo,
K., Ropelewski, C., Wang, J., Leetmaa, A., Reynolds, R., Jenne, R., Joseph, D., 1996.
The NCEP/NCAR 40-years reanalysis project. Bull. Am. Meteorol. Soc. 77, 437–470
data available at http://www.cdc.noaa.gov/cdc/reanalysis/.
Lirer, L., Pescatore, T., Booth, B., Walker, G., 1973. Two plinian pumice-fall deposits from
Somma–Vesuvius, Italy. Geol. Soc. Am. Bull. 84, 759–772.
Lirer, L., Munno, R., Petrosino, P., Vinci, A., 1993. Tephrostratigraphy of teh A.D. 79
pyroclastic deposits in the perivolcanic area of Vesuvius. J. Volcanol. Geotherm. Res.
58, 133–149.
Macedonio, G., Pareschi, M., Santacroce, R., 1990. Renewal of volcanic activity at
Vesuvius: tephra fallout. J. Volcanol. Geotherm. Res. 40, 327–342.
Macedonio, G., Pareschi, M., Santacroce, R., 1988. A numerical simulation of the Plinian
fall phase of the 79 AD eruption of Vesuvius. J. Geophys. Res. 93 (B12), 14817–14827.
Macedonio, G., Costa, A., Longo, A., 2005. A computer model for volcanic ash fallout and
assessment of subsequent hazard. Comput. Geosci. 31, 837–845.
Marzocchi, W., Sandri, L., Gasparini, P., Newhall, C., Boschi, E., 2004. Quantifying
probabilities of volcanic events: the example of volcanic hazard at Mount Vesuvius.
J. Geophys. Res. 109 (B11201). doi:10.1029/2004JB003155 18 pp.
Mastrolorenzo, G., Palladino, D., Vecchio, G., Taddeucci, J., 2002. The 472 AD Pollena
eruption of Somma–Vesuvius (Italy) and its environmental impact at the end of the
Roman Empire. J. Volcanol. Geotherm. Res. 113, 19–36.
377
Neri, A., Aspinall, W., Cioni, R., Bertagnini, A., Baxter, P., Zuccaro, G., Andronico, D.,
Barsotti, S., Cole, P., Esposti Ongaro, T., Hincks, T., Macedonio, G., Papale, P., Rosi, M.,
Santacroce, R., Woo, G., 2008. Developing an event tree for probabilistic hazard and
risk assessment at Vesuvius. J. Volcanol. Geotherm. Res. 178, 397–415 (this issue).
Pfeiffer, T., Costa, A., Macedonio, G., 2005. A model for the numerical simulation of
tephra fall deposits. J. Volcanol. Geotherm. Res. 140, 273–294.
Principe, C., Tanguy, J., Arrighi, S., Paiotti, A., Le Goff, M., Zoppi, U., 2004. Chronology of
Vesuvius' activity from A.D. 79 to 1631 based on archaeomagnetism of lavas and
historical sources. Bull. Volcanol. 66, 703–724. doi:10.1007/s00445-004-0348-8.
Rolandi, G., Munno, R., Postiglione, I., 2004. The AD 472 eruption of the Somma volcano.
J. Volcanol. Geotherm. Res. 129, 291–319.
Rosi, M., Santacroce, R., 1983. The A.D. 472 “Pollena” Eruption: volcanological and
petrological data from this poorly known Plinian-type event at Vesuvius. J. Volcanol.
Geotherm. Res. 17, 249–271.
Rosi, M., Principe, C., Vecci, R., 1993. The 1631 Vesuvian eruption: a reconstruction based
on historical and stratigraphical data,. J. Volcanol. Geotherm. Res. 58, 151–182.
Santacroce, R., 1983. A general model for the behaviour of the Somma Vesuvius volcanic
complex. J. Volcanol. Geotherm. Res. 17, 237–248.
Santacroce, R. (Ed.), 1987. Somma-Vesuvius. Vol. Monografia n. 114(8) of Quaderni de La
Ricerca Scientifica. CNR, Roma. 230 pp.
Scandone, R., Giacomelli, L., Gasparini, P., 1993. Mount Vesuvius: 2000 yrs of
volcanological observations. J. Volcanol. Geotherm. Res. 58, 263–271.
Scire,, J.S., Robe, F.R., Fernau, M.E., Yamartino, R.J., 2000. A User's Guide for the CALMET
Meteorological Model. Tech. Rep. Version 5. Earth Tech, Inc., 196 Baker Avenue,
Concord, MA 01742.
Sigurdsson, H., Carey, S., Cornell, W., Pescatore, T., 1985. The eruption of Vesuvius in A.D.
79. Natl. Geogr. Res. 1, 332–387.
Sparks, R., 1986. The dimensions and dynamics of volcanic eruption columns. Bull.
Volcanol. 48, 3–15.
Sulpizio, R., Mele, D., Dellino, P., La Volpe, L., 2005. A complex, subplinian-type eruption from
low-viscosity tephritic to tephritic-phonolitic magma: the AD 472 (Pollena) eruption of
Somma–Vesuvius, Italy. Bull. Volcanol. 67, 673–767. doi:10.1007/s00445-005-0414-x.
Suzuki, T., 1983. A theoretical model for dispersion of tephra. In: Shimozuru, D.,
Yokoyama, I. (Eds.), Arc Volcanism: Physics and Tectonics. Terra Scientific Publish-
ing Company (TERRAPUB), Tokyo, pp. 93–113.
Zuccaro, G., Cacace, F., Spence, R., Baxter, P., 2008. Impact of explosive eruption
scenarios at vesuvius. J. Volcanol. Geotherm. Res. 178, 416–453 (this issue).
doi:10.1016/j.jvolgeores.2008.01.005.