Arnau Folch1, Antonio Costa2 [631828]

FALL3D-6.2
User Guide
Arnau Folch1, Antonio Costa2
1 Earth Sciences Division
Barcelona Supercomputing Center – Centro Nacional de
Supercomputaci´ on
Edifici Nexus II – c/ Jordi Girona 29
08034 Barcelona, Spain
2Istituto Nazionale di Geofisica e Vulcanologia
Via Diocleziano 328 – 80124 Napoli, Italy
March 2010

FALL3D 6.2 USER’S MANUAL 2
Contents
1 Introduction 3
2 Ash transport model 3
2.1 Governing equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Eddy Diffusivity Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Settling velocity models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Meteorological variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.5 Source term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.6 Particle aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3 Overview of the program FALL3D-6.2 6
4 TheFALL3D-6.2 Input files 7
4.1 The control file filename.inp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
4.1.1 BLOCK TIME UTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.1.2 BLOCK GRID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.1.3 BLOCK FALL3D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.1.4 BLOCK GRANULOMETRY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.1.5 BLOCK SOURCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.1.6 BLOCK OUTPUT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.2 The database file filename.dbs.nc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.3 The granulometry file filename.grn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.4 The source file filename.src . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.5 The points file filename.pts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5 Program Setup 13
5.1 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.1.1 Unix/Linux/Mac X OS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.1.2 Windows OS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.2 Execution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.2.1 Unix/Linux/Mac X OS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.2.2 Windows OS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Appendices 14
Appendix A: Format of the meteo profile file ( filename.profile ) . . . . . . . . . . . . . . . . 14
Appendix B: The GRD format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Appendix C: The NetCDF format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

FALL3D 6.2 USER’S MANUAL 3
1 Introduction
FALL3D-6.2 is a 3-D time-dependent Eulerian model for the transport and deposition of volcanic ash.
The model solves the advection-diffusion-sedimentation (A DS) equation on a structured terrain-following
grid using a second-order Finite Differences (FD) explicit s cheme. Different parameterizations for the
eddy diffusivity tensor and for the particle terminal settli ng velocity can be chosen. The code, written
inFORTRAN-90 , is available for Unix/Linux/Mac X Operating Systems (OS). A set of pre- and post-
process utility programs and OS-dependent scripts to launc h them are also included in the FALL3D-6.2
distribution package. Although the model has been designed to forecast volcanic ash concentration in the
atmosphere and ash loading at the ground, it can also be used t o model the transport of other kinds of
airborne solid particles. The model inputs are meteorologi cal data, topography, grain-size distribution,
shape and density of particles, and mass rate of particle inj ected into the atmosphere. The FALL3D-6.2
model can be used as a tool for short-term ash deposition fore casting and for volcanic fallout hazard
assessment. FALL3D-6.2 is available in two versions called PUB (Public) version and PROF (Professional)
version.
More information on http://www.bsc.es/projects/earthsc ience/fall3d/
and http://datasim.ov.ingv.it/Fall3d.html
2 Ash transport model
In this section we briefly describe the governing equations a nd the main assumptions used in FALL3D-6.2 .
For further details see Costa et al. (2006); Folch et al. (200 9).
2.1 Governing equations
The main factors controlling atmospheric transport of ash a re wind advection, turbulent diffusion, and
gravitational settling of particles. Neglecting particle -particle interaction effects (collisions, aggregation,
etc.), the Eulerian form of the continuity equation written in a generalized coordinate system ( X,Y,Z )
is (Byun and Schere, 2006; Costa et al., 2006):
∂C
∂t+VX∂C
∂X+VY∂C
∂Y+ (VZ−Vsj)∂C
∂Z=−C∇ ·V+C∂Vsj
∂Z
+∂
∂X/parenleftbigg
ρ∗KX∂C/ρ ∗
∂X/parenrightbigg
+∂
∂Y/parenleftbigg
ρ∗KY∂C/ρ ∗
∂Y/parenrightbigg
+∂
∂Z/parenleftbigg
ρ∗KZ∂C/ρ ∗
∂Z/parenrightbigg
+S∗(1)
whereCis the transformed concentration, V= (VX,VY,VZ) is the transformed wind speed, KX,KYand
KZare the diagonal terms of the transformed eddy diffusivity te nsor,ρ∗is the transformed atmospheric
density, and S∗is the transformed source term. FALL3D-6.2 solves Eq. (1) for each particle class jusing
a curvilinear terrain-following coordinate system ( X=mx, Y =my, z →Z), wheremis the map
scale factor and Z=z−h(x,y), withh(x,y) denoting the topographic elevation, and ( x,y,z ) are the
Cartesian coordinates. The scaling factors for this partic ular transformation are given in Table 1 (Byun
and Schere, 2006). The generic particle class jis defined by a triplet of values characterizing each particl e
(dp,ρp,Fp), that are, respectively, diameter, density, and a shape fa ctor. Fordpwe use the equivalent
diameterd, which is the diameter of a sphere of equivalent volume. For t he shape factor Fpwe choose the
sphericityψ, which is the ratio of the surface area of a sphere with diamet erdto the surface area of the
particle. In our approximation, each triplet ( d,ρp,ψ) is sufficient to define the settling velocity. Effect of
Earth’s curvature are considered when the lat-lon coordina te system is used through the Jacobian of the
transformation.
2.2 Eddy Diffusivity Tensor
InFALL3D-6.2 only the diagonal components of the Eddy Diffusivity Tensor, i.e.the vertical Kzand
the horizontal Kh=Kx=Kycomponents, are considered.
The available choices for describing the vertical componen tKzare:
1. Option CONSTANT ,i.e.Kz= constant , where the constant value is assigned by the user;

FALL3D 6.2 USER’S MANUAL 4
Parameter Scaling
Coordinates X=mx;Y=my;Z=z−h(x,y)
Horizontal Velocities VX=mvx;VY=mvy
Vertical velocity (VZ−VSj) =J−1/bracketleftBig
(vz−vsj)−m/parenleftBig
vx∂h
∂x+vy∂h
∂y/parenrightBig/bracketrightBig
Diffusion Coefficients KX=Kx;KY=Ky;KZ≃KzJ−2
Concentration C=cJ/m2
Density ρ∗=ρJ/m2
Source Term S∗=SJ/m2
Table 1: Scaling factors for a terrain-following coordinat e system (x=mX, y =mY, z →Z). (x,y,z )
are the Cartesian coordinates, mthe map scale factor (for the UTM coordinate system m= 1) andJis
the determinant of the Jacobian of the coordinate system tra nsformation.
2. Option SIMILARITY . In this case, inside the Atmospheric Boundary Layer (ABL), FALL3D-6.2
evaluatesKzas:
Kz=

κu∗z/parenleftBig
1−z
h/parenrightBig/parenleftbigg
1 + 9.2h
Lz
h/parenrightbigg−1
h/L≥0 stable
κu∗z/parenleftBig
1−z
h/parenrightBig/parenleftbigg
1−13h
Lz
h/parenrightbigg1/2
h/L≤0 unstable(2)
whereκis the von Karman constant ( κ= 0.4),u∗is the friction velocity, his the ABL height,
andLis the Monin-Obukhov length (see Costa et al., 2006). The exp ression above comes from an
extension of the Monin-Obukhov similarity theory to the ent ire ABL (Ulke, 2000). On the other
hand, above the ABL ( z/h> 1),Kzis considered a function of the local vertical wind gradient , a
characteristic length scale lc, and a stability function Fcwhich depends on the Richardson number
Ri:
Kz=l2
c/vextendsingle/vextendsingle/vextendsingle/vextendsingle∂U
∂z/vextendsingle/vextendsingle/vextendsingle/vextendsingleFc(Ri) (3)
whereU=/radicalBig
y2x+u2y. ForlcandFc,FALL3D-6.2 adopts the relationship used by the CAM-3.0
model (Collins et al., 2004):
lc=/parenleftbigg1
κz+1
λc/parenrightbigg−1
(4)
Fc(Ri) =

1
1 + 10Ri(1 + 8Ri)stable (Ri>0)
√
1−18Ri unstable ( Ri<0)(5)
whereλcis the so-called asymptotic length scale ( λc≈30m).
The available choices for describing the horizontal compon entKh=Kx=Kyare:
1. Option CONSTANT ,i.e.Kh= constant , where the constant value is assigned by the user;
2. Option RAMS. In this case, a large eddy parameterization as that used by t he RAMS model (Pielke
et al., 1992) can be used for evaluating Kh:
Kh=Prtmax
km; (CS∆)2/radicaltp/radicalvertex/radicalvertex/radicalbt/parenleftbigg∂vx
∂y+∂vy
∂x/parenrightbigg2
+ 2/bracketleftBigg/parenleftbigg∂vx
∂x/parenrightbigg2
+/parenleftbigg∂vy
∂y/parenrightbigg2/bracketrightBigg
 (6)
wherePrtis the turbulent Prandtl number (typically Prt≈1),km= 0.075∆4/3, ∆ =√∆x∆y,
∆xand ∆yare the horizontal grid spacings, and CSis a constant ranging from 0.135 to 0.32.

FALL3D 6.2 USER’S MANUAL 5
3. Option CMAQ. In this case, the horizontal diffusion is evaluated as in the CMAQ model (Byun and
Schere, 2006):
1
Kh=1
Kht+1
Khn(7)
where:
Kht=α2∆x∆y/radicalBigg/parenleftbigg∂vx
∂x−∂vy
∂y/parenrightbigg2
+/parenleftbigg∂vy
∂x+∂vx
∂y/parenrightbigg2
(8)
Khn=Khf/parenleftbigg∆xf∆yf
∆x∆y/parenrightbigg
(9)
whereα= 0.28 and the values of Khfand ∆xf= ∆yfdepend on the algorithm. Using this
parameterization, for a large grid size the effect of the tran sportive dispersion is minimized, whereas
for a small grid size the numerical diffusion term is reduced ( Byun and Schere, 2006). Thanks
to the heuristic relationship (7), the smaller between KhtandKhndominates. In our case we
setKhf= 8000 m2s−1for ∆xf= ∆yf= 4 km and also a minimum value for Khequal to
km= 0.075∆4/3was imposed.
2.3 Settling velocity models
There are several semi-empirical parameterizations for th e particle settling velocity vsif one assumes that
particles settle down at their terminal velocity:
vs=/radicalBigg
4g(ρp−ρa)d
3Cdρa(10)
whereρaandρpdenote air and particle density, respectively, dis the particle equivalent diameter, and Cd
is the drag coefficient. Cddepends on the Reynolds number, Re=dvs/νa(νa=µa/ρais the kinematic
viscosity of air, µathe dynamic viscosity). In FALL3D-6.2 different options are possible for estimating
settling velocity, such as:
1.ARASTOOPOUR model (Arastoopour et al., 1982):
Cd=

24
Re(1 + 0.15Re0.687)Re≤988.947
0.44 Re> 988.947(11)
valid for spherical particles only.
2.GANSER model (Ganser, 1993):
Cd=24
ReK 1/braceleftBig
1 + 0.1118 (ReK 1K2)0.6567/bracerightBig
+0.4305K2
1 +3305
ReK 1K2(12)
whereK1= 3/[(dn/d)+2ψ−0.5],K2= 101.8148(−Logψ)0.5743are two shape factors ( dnis the average
between the minimum and the maximum axis, dis the equal volume sphere), and ψis the particle
sphericity ( ψ= 1 for spheres). For calculating the sphericity is practica l to use the concepts of
operational and working sphericity, ψworkintroduced by Wadell (1933); Aschenbrenner (1956),
which are based on the determination of the volume and of the t hree dimensions of a particle
respectively:
ψwork= 12.8(P2Q)1/3
1 +P(1 +Q) + 6/radicalbig
1 +P2(1 +Q2)(13)
withP=S/I,Q=I/L, whereLis the longest particle dimension, Iis the longest dimension
perpendicular to L, andSis the dimension perpendicular to both LandI.

FALL3D 6.2 USER’S MANUAL 6
3.WILSON model (Walker et al., 1971; Wilson and Huang, 1979) using the interpolation suggested by
Pfeiffer et al. (2005):
Cd=

24
Reϕ−0.828+ 2/radicalbig
1−ϕ Re ≤102
1−1−Cd|Re=102
900(103−Re) 102≤Re≤103
1 Re≥103(14)
whereϕ= (b+c)/2ais the particle aspect ratio ( a≥b≥cdenote the particle semi-axes).
4.DELLINO model (Dellino et al., 2005):
vs= 1.2605νa
d/parenleftbig
Arξ1.6/parenrightbig0.5206(15)
whereAr=gd3(ρp−ρa)ρa/µ2
ais the Archimedes number, gthe gravity acceleration, and ξis a
particle shape factor (sphericity to circularity ratio). I t is recommended to not extrapolate this
option for particle diameter beyond the range used in the exp eriments by Dellino et al. (2005).
Since for FALL3D-6.2 the primary particle shape factor is the sphericity ψ, for sake of simplicity, ϕin
(14) andξin (15) are calculated approximating particles as prolate e llipsoids (the same approximation
is used for estimating dn).
2.4 Meteorological variables
FALL3D-6.2 uses an off-line strategy, i.e.the meteorological variables are calculated independentl y by
a different meteorological model or information, and interp olated to the spatial and temporal grid of
FALL3D-6.2 .
FALL3D-6.2 reads time-dependent meteorological data (wind field, air t emperature, Monin-Obukhov
lengthL, friction velocity u∗, and ABL height h) and topography from a database file in NetCDF format.
This file can be created by an external utility program ( SETDBS ). This strategy allows FALL3D-6.2 to be
used from micro- to meso-scale. In the PUB distribution vers ion there are a few options to generate this
database depending on the scale of application. For the poss ible choices see Section 4.2.
2.5 Source term
FALL3D-6.2 reads the time-dependent source term (mass released per uni t time at each grid point) from
an external file. This file can be generated by the SETSRC utility program, as i) a point source, ii)
imposing a mushroom-like shape (Suzuki option) or iii) by using a model based on the Buoyant Plume
Theory (BPT; see Section 4.4).
2.6 Particle aggregation
A subroutine describing particle aggregation in presence o f water is being tested. Aggregation model and
validation tests are presented in Costa et al. (2010); Folch et al. (2010).
3 Overview of the program FALL3D-6.2
FALL3D-6.2 needs the following input files:
•An input file where control parameters and options are specifi ed (filename.inp ). This file is read
byFALL3D-6.2 and the utility programs.
•A database file in NetCDF format (see Appendix C) containing a ll meteorological data and the
topography ( filename.dbs.nc ).
•A granulometry file specifying the characteristics of the pa rticles emitted into the atmosphere
(filename.grn ).

FALL3D 6.2 USER’S MANUAL 7
•A source file specifying the discharge rates at the source poi nts, typically along the eruptive column
(filename.src ).
•Optionally, a file specifying a list of points ( filename.pts ) where tracking of variables is performed
(e.g.to compute ash arrival times, accumulation rates, etc).
The formats of the input files are described in Section 4. The FALL3D-6.2 package comes with a set of
utility programs that can be used to generate the input files:
•The utility SETDBS can be used to generate the database file filename.dbs.nc created in accord to
the parameters specified in the blocks TIMEUTCandGIRDof the input file filename.inp . This util-
ity program reads meteo data from different sources and inter polates variables onto the FALL3D-6.2
computational grid. The time slice of the database must be eq ual or larger than the simulation
time slice.
•The utility SETGRN can be used to generate the granulometry file filename.grn in accord to the
parameters specified in the block GRANULOMETRY of the input file filename.inp . This program
generates only Gaussian and Bi-Gaussian distributions. Fo r other distributions the user must
provide the granulometry file.
•The utility SETSRC can be used to generate the source file filename.src in accord to the parameters
specified in the blocks TIMEUTCandSOURCE of the input file filename.inp .
The use of these utilities, although recommended, is not nec essary if the user provides some of the
necessary files directly.
Once a simulation is concluded, FALL3D-6.2 outputs the following files:
•The results file ( filename.res.nc ) in netCDF format. This file can be processed using several
open-source programs ( e.g.ncview, Panoply, ncl, etc.) to generate plots and animation s.
•The log file ( filename.log ) contains information about the run, including summary of i nput data,
error and warning messages, etc.
•The tracking points files ( filename.pts.* ) contain information about evolution of variables at the
tracked points. There exist an output file for each point spec ified in the input file filename.pts .
4 The FALL3D-6.2 Input files
4.1 The control file filename.inp
The control input file is an ASCII file made up with a set of block s that define all the computational and
physical parameters needed by FALL3D-6.2 and the rest of utility programs. Each program reads only
the necessary blocks of the file. Parameters within a block ar e listed one per record, in arbitrary order,
and can optionally be followed by one or more blank spaces and a comment. Maximum allowed lenght is
256 characters including comments. A detailed description of each record is given below. Real numbers
can be expressed following the FORTRAN notation ( e.g., 12e7 = 12 ×107).
•BLOCK TIME UTC: Defines variables related to time.
•BLOCK GRID : Defines the characteristics of the FALL3D-6.2 computational mesh.
•BLOCK FALL3D : Defines the variables needed by FALL3D-6.2 program.
•BLOCK GRANULOMETRY : Defines the variables needed by SETGRN program.
•BLOCK SOURCE : Defines the variables needed by SETSRC program.
•BLOCK OUTPUT : Defines the FALL3D-6.2 strategy for dumping of results.

FALL3D 6.2 USER’S MANUAL 8
4.1.1 BLOCK TIME UTC
This block of data defines variables related to time and is use d byFALL3D-6.2 and by the utility programs
SETDBS andSETSRC .
•YEAR: Database starting year.
•MONTH : Database starting month (1-12).
•DAY: Database starting day (1-31).
•BEGINMETEODATA(HOURSAFTER00): Time (in hafter 0000UTC of the starting day) at which
meteorological data starts in the database file.
•TIMESTEPMETEODATA(MIN) : Time step (in min) of the meteo data in the database file.
•ENDMETEODATA(HOURSAFTER00): Time (in hafter 0000UTC of the starting day) at which me-
teorological data ends in the database file. This time slice h as to be larger than time slices defined
by the records ERUPTION START(HOURSAFTER00)andRUNEND(HOURSAFTER00). If not, the
program will stop.
•ERUPTION START(HOURSAFTER00): Eruption start hours (after 0000UTC of the day). These are
ntvalues (nt≥1) that indicate the starting times of the different eruptive phases. Any type
of transient behavior can be contemplated by adding a sufficie nt number of intervals. Eruptive
conditions (plume height, MFR, etc.) are assumed constant d uring each phase ( i.e.a quasi-steady
approximation is used). The first value must be equal or large r than the value of the record
BEGINMETEODATA(HOURSAFTER00).
•ERUPTION END(HOURSAFTER00): Eruption end hour (after 0000UTC of the starting day). If th e
SETSRC program is used to generate the source term, this is the time s lice at which source term is
switched off ( i.e.the time at which the last eruptive phase ends).
•RUNEND(HOURSAFTER00): Run end hour (after 0000UTC of the starting day). Must be equ al or
lower than the value of the record ENDMETEODATA(HOURSAFTER00). Note that, in general, a run
should continue even when the source term is switched off ( i.e.when the eruption has stopped) in
order to allow the remaining airborne particles to sediment completely.
4.1.2 BLOCK GRID
This block of data defines the variables needed by SETDBS program to generate the FALL3D-6.2 grid.
Note that time and spatial coverage of the database must incl ude theFALL3D-6.2 simulation interval.
•COORDINATES : Defines the map projection. Possibilities are UTM or LON-LA T. Note that the
UTM option can only be used if the domain is within a unique UTM zone. The use of the UTM
coordinate system in large domains covering more than one UT M zone is not allowed (in this case,
the LON-LAT option accounting for Earth’s curvature must be used instead). The sub-blocks UTM
orLONLATare read in each case respectively.
•LONMIN : Minimum longitude (in decimal degrees) of the domain ( i.e.longitude corresponding to
the bottom left corner). Only used in the LON-LAT option.
•LONMAX : Maximin longitude (in decimal degrees) of the domain ( i.e.longitude corresponding to top
right corner). Only used in the LON-LAT option.
•LATMIN : Minimum latitude (in decimal degrees) of the domain ( i.e.latitude corresponding to
bottom left corner). Only used in the LON-LAT option.
•LATMAX : Maximin latitude (in decimal degrees) of the domain ( i.e.latitude corresponding to top
right corner). Only used in the LON-LAT option.
•LONVENT: Vent longitude. Only used in the LON-LAT option.
•LATVENT: Vent latitude. Only used in the LON-LAT option.

FALL3D 6.2 USER’S MANUAL 9
•UTMZONE : UTM zone code in format nnL ( e.g.33S). Only used in the UTM option.
•XMIN: minimum x-coordinate of the domain (bottom left corner). UTM coordin ates must be given
in m. Only used in the UTM option.
•XMAX: maximum x-coordinate of the domain (top right corner). UTM coordinat es must be given in
m. Only used in the UTM option.
•YMIN: minimum y-coordinate of the domain (bottom left corner). UTM coordin ates must be given
in m. Only used in the UTM option.
•YMAX: maximum y-coordinate of the domain (top right corner). UTM coordinat es must be given in
m. Only used in the UTM option.
•XVENT:x-coordinate of the vent. UTM coordinates must be given in m. O nly used in the UTM
option.
•YVENT:y-coordinate of the vent. UTM coordinates must be given in m. O nly used in the UTM
option.
•NX: Number of grid nodes in the x-direction.
•NY: Number of grid nodes in the y-direction.
•ZLAYER(M): Array of heights (in m) of the vertical z-layers in terrain following coordinates. It is not
necessary to specify the number of vertical layers since it i s automatically calculated by the program.
The vertical layers can be specified manually (as an array of v alues) or, for equally spaced vertical
discretization, simply indicating the limits and the incre ment ( e.g.FROM 0 TO 10000 INCREMENT
1000).
4.1.3 BLOCK FALL3D
This block of data defines the variables needed by FALL3D-6.2 program.
•TERMINAL VELOCITY MODEL : Type of terminal settling velocity model. Possibilities a reARASTOOPOUR ,
GANSER ,WILSON , andDELLINO .
•VERTICAL TURBULENCE MODEL : Type of model for vertical diffusion. Possibilities are CONSTANT or
SIMILARITY .
•VERTICAL DIFFUSION COEFFICIENT (M2/S) : Value of the diffusion coefficient (in m2/s). Only used
ifVERTICAL TURBULENCE MODEL = CONSTANT
•HORIZONTAL TURBULENCE MODEL : Type of model for horizontal diffusion. Possibilities are CONSTANT ,
RAMS, orCMAQ.
•HORIZONTAL DIFFUSION COEFFICIENT (M2/S) : Value of the diffusion coefficient (in m2/s). Only
used ifHORIZONTAL TURBULENCE MODEL = CONSTANT .
•RAMSCS: Value ofCSin the RAMS model (see eq. 6). Only used if HORIZONTAL TURBULENCE MODEL
= RAMS .
4.1.4 BLOCK GRANULOMETRY
This block of data defines the variables needed by SETGRN program.
•DISTRIBUTION : Type of distribution. Possibilities are GAUSSIAN orBIGAUSSIAN .
•NUMBEROFCLASSES : Number of granulometric classes.
•FIMEAN: Mean value of Φ (Gaussian distribution). For Bi-Gaussian d istributions two values must
be provided.

FALL3D 6.2 USER’S MANUAL 10
•FIDISP: Value ofσ(Gaussian distribution). For Bi-Gaussian distributions t wo values must be
provided.
•FIRANGE : Minimum and maximum values of Φ (Φ minand Φmaxrespectively).
•DENSITY RANGE : Values of densities ρminandρmax(inkg/m3) associated to Φ minand Φmaxpar-
ticles. Lineal interpolation is assumed. In particular, if ρmin=ρmax, density is constant for all
classes.
•SPHERICITY RANGE : Values of sphericity ψminandψmaxassociated to Φ minand Φmaxparticles.
Lineal interpolation is assumed. In particular, if ψmin=ψmax, sphericity is constant for all classes.
4.1.5 BLOCK SOURCE
This block of data defines the variables needed by the SETSRC program. This program generates the
source term (eruptive column) for each of the nt≥1 eruptive phases.
•VENTHEIGHT : Height of the vent a.s.l. (in m).
•SOURCETYPE: Type of source distribution. Possibilities are POINT ,SUZUKI orPLUME .
In the case SOURCETYPE = POINT only the sub-block POINTSOURCE is used:
•MASSFLOWRATE(KGS) : Array of values of the mass flow rate (in kg/s) for the nteruptive phases.
Alternatively, the user can use the word estimate andSETSRC automatically computes the MFR
from the column heights based on empirical fits (Mastin et al. , 2009). This is the typical situation
during an eruption, when column height is likely to be the onl y observable available.
•HEIGHTABOVEVENT(M): Array of heights of the plume (in m above the vent) for the nteruptive
phases. Note that the plume heights must be lower than the top of the computational domain,
specified in the record ZLAYER(M)of theGRIDblock. If not, the program will stop.
In the case SOURCETYPE = SUZUKI only the sub-block SUZUKISOURCE is used:
•MASSFLOWRATE(KGS) : Array of values of the mass flow rate (in kg/s) for the nteruptive phases.
Alternatively, the user can use the word estimate andSETSRC automatically computes the MFR
from the column heights based on empirical fits (Mastin et al. , 2009). This is the typical situation
during an eruption, when column height is likely to be the onl y observable available.
•HEIGHTABOVEVENT(M): Array of heights of the plume (in m above the vent) for the nteruptive
phases. Note that the plume heights must be lower than the top of the computational domain,
specified in the record ZLAYER(M)of theGRIDblock. If not, the program will stop.
•A: Array of values of the parameter Ain the Suzuki distribution for the nteruptive phases (Pfeiffer
et al., 2005).
•L: Array of values of the parameter λin the Suzuki distribution for the nteruptive phases (Pfeiffer
et al., 2005).
In the case SOURCETYPE = PLUME only the sub-block PLUMESOURCE is used:
•SOLVEPLUMEFOR: Possibilities are MFRorHEIGHT . In the first case SETSRC solves for the mass flow
rate given the column height, whereas in the second does the o pposite. Since the plume equations
use the mass flow rate as an input, the first option requires an i terative procedure.
•MFRSEARCHRANGE : Two values nandmsuch that 10nand 10mspecify the range of MFR values
admitted in the iterative solving procedure ( i.e.it is assumed that 10n≤MFR ≤10m). Only
used ifSOLVEPLUMEFOR=MFR .
•MASSFLOWRATE(KGS) : Values of the mass flow rate (in kg/s) for the nteruptive phases. Only
used ifSOLVEPLUMEFOR=HEIGHT .

FALL3D 6.2 USER’S MANUAL 11
•HEIGHTABOVEVENT(M): Heights of the plume (in m above the vent) for the nteruptive phases.
Note that the plume heights must be lower than the top of the co mputational domain, specified in
the record ZLAYER(M)of theGRIDblock. Only used if SOLVEPLUMEFOR=MFR .
•EXITVELOCIY (MS): Values of the magma exit velocity (in m/s) at the vent for the nteruptive
phases.
•EXITTEMPERATURE (K): Values of the magma exit temperature (in K) at the vent for th enteruptive
phases.
•EXITVOLATILE FRACTION (IN%) : Values of the magma volatile at the vent for the nteruptive
phases in weight percent.
4.1.6 BLOCK OUTPUT
This block of data defines the output strategy of the FALL3D-6.2 program.
•POSTPROCESS TIMEINTERVAL (HOURS) : Postprocess time interval in hours.
•POSTPROCESS 3DVARIABLES : Possibilities are YESorNO. IfYES,FALL3D-6.2 writes 3D concentration
in the output file filename.res.nc . IfNO, only 2D variables are written to the output file.
•POSTPROCESS CLASSES : Possibilities are YESorNO. IfYES,FALL3D-6.2 writes results for all the
classes. If NO, only total results are written.
•TRACKPOINTS : Possibilities are YESorNO. IfYES,FALL3D-6.2 writes the tracking points files.
4.2 The database file filename.dbs.nc
This file written in NetCDF format (see Appendix C) contains t ime-dependent meteorological data (wind
field, air temperature and density, humidity, etc) needed by FALL3D-6.2 . The file can be created by the
external utility program SETDBS , which reads meteo data and interpolates to the FALL3D-6.2 space-time
domain. There are a several options to generate this databas e depending on the scale of application. This
strategy allows FALL3D-6.2 to be used from micro- to meso-scale. The possible choices ar e described
below.
•The simplest option consists of using a horizontally unifor m wind derived from a vertical profile,
typically obtained from sounding measurements or from indi rect reconstructions. The vertical
profile needs to be specified in the file filename.profile in the format described in the Appendix
A. In this case, in addition to the profile filename.profile it is also necessary to furnish a
topography file filename.top in GRD format (see Appendix B).
•The second choice ( CALMET option) uses data derived from the output of the meteorologi cal diag-
nostic model CALMET (Scire et al., 2000). This option is used for assimilating an d interpolating
short-term forecasts (or re-analysis) from Mesoscale Mete orological Prognostic Models (MMPM)
to a finer scale. In this case only the UTM coordinate system ca n be used. Note that the output
ofCALMET is a binary file that depends on the architecture of the machin e were it was generated.
Moreover note that this option is compatible only with a CALMET output time step equal to an hour
(i.e.,nsecdt=3600 ).
•The third choice ( NCEP-1 option) uses data from NCEP re-analysis 1
(see: http://www.esrl.noaa.gov/psd/data/gridded/data .ncep.reanalysis.html).
•Other choices (available only in the PROF version) contempl ate several global and mesoscale me-
teorological models such as ARPA-SIM, ETA, GFS, or NMMb.

FALL3D 6.2 USER’S MANUAL 12
nc
diam(1) rho(1) sphe(1) fc(1)
…
diam(nc) rho(nc) sphe(1) fc(nc)
Table 2: Format of the granulometry file filename.grn .
4.3 The granulometry file filename.grn
The granulometry file is an ASCII file containing the definitio n of the particle classes (a class is character-
ized by particle size, density and sphericity). This file can be created by the utility program SETGRN . Note
thatSETGRN only generates distributions which are Gaussian or bi-Gaus sian inψ(log-normal in d) and
linear inρandψ.FALL3D-6.2 can handle more general distributions but, in this case, the granulometry
filefilename.grn has to be defined directly by the user. The file format is descri bed in Table 2 and the
meaning of the used symbols is the following:
•nc: Number of particle classes.
•diam: Class diameter (in mm).
•rho: Class density (in kg/m3).
•sphe: Class sphericity.
•fc: Class mass fraction (0-1). It must verify that/summationtextfc= 1.
4.4 The source file filename.src
The source file filename.src is an ASCII file containing the definition of the source term. T he source
can be defined for different time phases during which source va lues are kept constant. The number,
position and values ( i.e.Mass Flow Rate) of the source points can vary from one time sli ce to another
and cannot overlap. There is no restriction on the number and duration of the time slices. It allows, in
practise, to discretize any kind of source term. This file can be defined directly by the user, in the format
described in Table 3, or created by using the utility program SETSRC . In the last case, the filename.src
is created in accord to the parameter specified in the SOURCE block of the filename.inp . The options to
be chosen in the filename.inp arei) a point source, ii) a mushroom-like shape (Suzuki option) or iii) an
eruption column model based on the Buoyant Plume Theory (BPT ). The format of the file filename.src
is described in Table 3 and the meaning of the used symbols is t he following:
•itime1 : Starting time of the time slice (in sec after 00UTC of the sta rting day).
•itime2 : End time of the time slice (in sec after 00UTC of the eruption starting day).
•nsrc: Number of source points (can vary from one interval to anoth er depending on the column
height).
•nc: Number of particle classes.
•MFR: Mass flow rate (in kg/s).
•x: Longitude or x-coordinate of the source isrc.
•y: Latitude or y-coordinate of the source isrc.
•z:z-coordinate of the source isrcabove ground level (a.g.l.) (in m).
•src: Mass flow rate (in kg/s) of each granulometric class for this point source. It must be verified
that/summationtext/summationtextsrc(isrc,ic ) =MFR .

FALL3D 6.2 USER’S MANUAL 13
itime1 itime2
nsrc nc
MFR
x y z src(1,1) … src(1,nc)
…
x y z src(nsrc,1) … src(nsrc,nc)
Table 3: Format of the source file filename.src . Repeat this block for each time slice
.
4.5 The points file filename.pts
This file contains the names (identifiers) and coordinates of the points to be tracked. It is used only
when the record TRACKPOINTS in the input file filename.inp is set to YES. The format of the file
filename.pts consists of lines (one line per point) with three columns spe cifying the point name, the
point longitude (or x-coordinate if UTM coordinates are used), and the point lati tude (ory-coordinate
if UTM coordinates are used). There is no limit on the number o f points.
5 Program Setup
5.1 Installation
To install FALL3D-6.2 and the utility programs uncompress and untar the file Fall3d-6.2.PUB.tar.gz .
It will create the folder structure shown in Table 4. The pack age contains the source codes, scripts,
documentation, and a run examples.
•For Unix/Linux/Mac X OS it is necessary to have a FORTRAN compiler and the NetCDF library
(http://www.unidata.ucar.edu/software/netcdf/) alrea dy installed (version 3.6 or later). It is manda-
tory to compile FALL3D-6.2 using the same FORTRAN compiler that has been used to compile the
NetCDF library.
5.1.1 Unix/Linux/Mac X OS
For Unix/Linux/Mac X OS the package comes with an automatic i nstallation script. Proceed as follows:
1. Enter the directory Install and edit the Install script file. Set up the variables HOMEFALL3D
(FALL3D-6.2 home directory path), Libnetcdf (path of the NetCDF library in your computer),
andCOMPILER (name of the FORTRAN compiler).
NOTE: Automatic installation is possible for the following standard compilers: gfortran, ifort,
f90, xlf90 . If you want to compile using a different compiler it is necess ary to modify the Makefiles
and the Scripts manually.
2. Run the Install script. This will compile FALL3D-6.2 and the utility programs, modify the scripts
introducing your FALL3D-6.2 path and check the installation process.
3. Optionally, create an alias to the Script-manager file located in the folder Scripts . This script
allows for launching FALL3D-6.2 and the utility programs directly from the command line.
5.1.2 Windows OS
Not yet available for the current versions.
5.2 Execution
To create a new run named problemname simply create a new directory problemname in the folder
Runs, copy the control input file from the example run (rename it as problemname.inp ), and modify it
depending on your needs.

FALL3D 6.2 USER’S MANUAL 14
Level 1 Level 2 Level 3 Description
Fall3d Documents Contains the manual.
Install Contains installation scripts.
Runs Run-name Contains the examples, one folder each.
Scripts Contains the script files.
Sources ser FALL3D-6.2 sources (PUB version).
Utilities LibMaster Master library.
SetDbs SETDBS utility program.
SetGrn SETGRN utility program.
SetSrc SETSRC utility program.
Table 4: Default structure of FALL3D-6.2 folders.
5.2.1 Unix/Linux/Mac X OS
FALL3D-6.2 and the utility programs can be launched using the Script-manager with a series of argu-
ments. It is recommended to use an alias for this script that c an be called directly from any location (in
the following it is assumed that the alias is Launch ). From any location:
•TypeLaunch SetGrn problemname to run the SETGRN utility program for problemname .
•TypeLaunch SetDbs problemname meteo to run the SETDBS utility program for problemname .
Heremeteo is one of the following: profile/calmet62/ncep1 .
•TypeLaunch SetSrc problemname to run the SETSRC utility program for problemname .
•TypeLaunch Pub problemname to run the PUB version of FALL3D-6.2 forproblemname .
5.2.2 Windows OS
Not yet available for the current versions.
Appendices
Appendix A: Format of the meteo profile file ( filename.profile )
For the profile option, the utility SetDbs needs an ASCII file containing the definition of the vertical
wind profile and a topography file of the domain in GRD format (s ee Appendix B). In this case wind
velocities are assumed constant on all the domain in a terrai n-following coordinate system. The remain-
ing variables are assumed with the values of the Standard Atm osphere. The format of the profile file
(filename.profile ) is described in Table 5 and the meaning of the used symbols is the following:
•pcoord : Coordinates where the profile was measured; either as UTM or lon-lat coordinates.
•pdate : Starting time when the profile was measured; the format of th e date is yyyymmdd ,i.e.year,
month, day.
•itime1 : Initial time in sec after the starting time pdate of validity of the meteo data contained in
the following nzlayers.
•itime2 : Final time in sec after the starting time pdate of validity of the meteo data contained in
the following nzlayers.
•nz: Number of the database vertical layers.
•z: Vertical coordinate of the layer (in m a.s.l.).
•ux: windx-velocity (in m/s).
•uy: windy-velocity (in m/s).

FALL3D 6.2 USER’S MANUAL 15
pcoord
pdate
itime1 itime2
nz
z(1) ux(1) ux(1) T(1)
…
z(nz) ux(nz) ux(nz) T(nz)
itime3 itime4
…
Table 5: Format of the meteo data file filename.profile.dat for thePROFILE case. Repeat this block
for each meteo time increment.
Appendix B: The GRD format
The structure of a GRD format file is described in Table 6 and th e meaning of the used symbols is the
following:
•NX: Number of grid points along x-direction.
•NY: Number of grid points along y-direction.
•XO:x-coordinate (UTM in m) of the grid bottom left corner.
•XF:x-coordinate (UTM in m) of the grid top right corner point.
•YO:y-coordinate (UTM in m) of the grid bottom left corner point.
•YF:y-coordinate (UTM in m) of the grid top right corner point.
•VAL: Value at each grid point. It consists of an array of NX×NYvalues stored starting from the
bottom-left corner and moving towards right then up towards the top-right corner.
NX NY
XO XF
YO YF
MAX(v) MIN(v)
VAL(i,1) … … i=1:NX
… … …
VAL(i,j) … … i=1:NX
… … …
VAL(i,NY) … … i=1:NX
Table 6: Format of a GRD file filename.grd .
Appendix C: The NetCDF format
NetCDF (network Common Data Form) is a set of software librar ies and machine-independent data
formats that support the creation, access, and sharing of ar ray-oriented scientific data (available at:
http://www.unidata.ucar.edu/software/netcdf/). FALL3D-6.2 uses the standard NetCDF format for both
database input file ( filename.dbs.nc ) and results output file ( filename.res.nc ). Only the PROF
version comes with the utility to view, manipulate or transf orm NetCDF files. However, there is a good
number of open-source codes to do so. For example:
•ncview and ncdump (http://opendap.org/download/nc clients.html).
•Panoply (http://www.giss.nasa.gov/tools/panoply/).

FALL3D 6.2 USER’S MANUAL 16
•GrADS (http://www.iges.org/grads/).
•NCL, the NCAR Command Language (http://www.ncl.ucar.edu/ ).
References
Arastoopour, H., Wang, C., Weil, S., 1982. Particle-partic le interaction force in a diluite gas-solid system.
Chem. Eng. Sci. 37 (9), 1379–1386.
Aschenbrenner, B., 1956. A new method of expressing particl e sphericity. J. Sediment. Petrol. 26, 15–31.
Byun, D., Schere, K., 2006. Review of the governing equation s, computational algorithms, and other
components of the Models-3 Community Multiscale Air Qualit y (CMAQ) modeling system. Applied
Mechanics Reviews 59, 51.
Collins, W., Rasch, P., Boville, B., Hack, J., McCaa, J., Wil liamson, D., Kiehl, J., Briegleb, B., 2004.
Description of the NCAR Community Atmosphere Model (CAM 3.0 ). Technical Report NCAR/TN-
464+STR, National Center for Atmospheric Research, Boulde r, Colorado.
Costa, A., Folch, A., Macedonio, G., 2010. A model for wet agg regation of ash particles in volcanic plumes
and clouds: I. Theoretical formulation. J. Geophys. Res. (i n press).
Costa, A., Macedonio, G., Folch, A., 2006. A three-dimensio nal Eulerian model for transport and depo-
sition of volcanic ashes. Earth Planet. Sci. Lett. 241, 634– 647.
Dellino, P., Mele, D., Bonasia, R., Braia, L., La Volpe, R., 2 005. The analysis of the influence of pumice
shape on its terminal velocity. Geophys. Res. Lett. 32 (L213 06).
Folch, A., Costa, A., Durant, A., Macedonio, G., 2010. A mode l for wet aggregation of ash particles in
volcanic plumes and clouds: II. Model application. J. Geoph ys. Res. (in press).
Folch, A., Costa, A., Macedonio, G., 2009. FALL3D: A computa tional model for transport and deposition
of volcanic ash. Comput. Geosci. 35 (6), 1334–1342.
Ganser, G., 1993. A rational approach to drag prediction of s pherical and nonspherical particles. Powder
Technol. 77, 143–152.
Mastin, L., Guffanti, M., Servranckx, R., Webley, P., Barsot ti, S., Dean, K., Durant, A., Ewert, J.,
Neri, A., Rose, W., Schneider, D., Siebert, L., Stunder, B., Swanson, G., Tupper, A., Volentik, A.,
Waythomas, C., 2009. A multidisciplinary effort to assign re alistic source parameters to models of
volcanic ash-cloud transport and dispersion during erupti ons. J. Volcanol. Geotherm. Res. 186, 10–21.
Pfeiffer, T., Costa, A., Macedonio, G., 2005. A model for the n umerical simulation of tephra fall deposits.
J. Volcanol. Geotherm. Res. 140, 273–294.
Pielke, R., Cotton, W., Walko, R., Tremback, C., Nicholls, M ., Moran, M., Wesley, D., Lee, T., Copeland,
J., 1992. A comprehensive meteorological modeling system- RAMS. Meteor. Atmos. Phys. 49, 69–91.
Scire, J., Robe, F., Fernau, M., Yamartino, R., 2000. A User’ s Guide for the CALMET Meteorological
Model. Tech. Rep. Version 5, Earth Tech, Inc., 196 Baker Aven ue, Concord, MA 01742.
Ulke, A., 2000. New turbulent parameterization for a disper sion model in atmospheric boundary layer.
Atmos. Environ. 34, 1029–1042.
Wadell, H., 1933. Sphericity and roundness of rock particle s. J. Geol. 41, 310–331.
Walker, G., Wilson, L., Bowell, E., 1971. Explosive volcani c eruptions I. Rate of fall of pyroclasts.
Geophys. J. Roy. Astron. Soc. 22, 377–383.
Wilson, L., Huang, T., 1979. The influence of shape on the atmo spheric settling velocity of volcanic ash
particles. Earth Planet. Sci. Lett. 44, 311–324.