Transcript FLEXPART - Welcome to zardoz.nilu no

The Lagrangian particle dispersion
model FLEXPART
Jimmy LECLAIR DE BELLEVUE
Presentation to UKZN training
session - 09 Nov 2006
3-D visualization of an intrusion of stratospheric
air into the troposphere, resulting from a
FLEXPART simulation.
The domain shown covers Europe, and the
uppermost level is at 13 km
Content









References
Introduction
What is good in FLEXPART ?
System requirements & versions
Summary of the available versions
Setup
Use
Important aspects of the physics
Case studies of stratospheric to tropospheric
exchanges : Examples
 Take-Home message
References

http://zardoz.nilu.no/~andreas/flextra+flexpart.html

Descriptions of FLEXPART in the scientific literature are:

Stohl, A., M. Hittenberger, and G. Wotawa (1998): Validation of the
Lagrangian particle dispersion model FLEXPART against large scale
tracer experiments. Atmos. Environ. 32, 4245-4264.

Stohl, A., and D. J. Thomson (1999): A density correction for Lagrangian
particle dispersion models. Bound.-Layer Met. 90, 155-167.

Stohl, A., C. Forster, A. Frank, P. Seibert, and G. Wotawa (2005):
Technical Note : The Lagrangian particle dispersion model FLEXPART
version 6.2. Atmos. Chem. Phys. 5, 2461-2474.
Introduction

FLEXPART is an atmospheric trajectory and a particle dispersion model,
respectively, that are used by a growing user community.

A recent user survey resulted in 34 groups from 17 countries who have
confirmed to actively use one of the models for a variety of research
purposes.

Applications of the models cover topics like transport of radionuclides
after nuclear accidents, pollution transport, greenhouse gas cycles,
stratosphere-troposphere exchange, water cycle research, and others.

Development supervised by Andreas Stohl and mainly by people from
 Norwegian Institute of Air Research, Kjeller, Norway
 Institute of Meteorology, University of Natural Resources and Applied
Life Sciences, Vienna, Austria
 Preparatory Commission for the Comprehensive Nuclear Test Ban
Treaty Organization, Vienna, Austria
 FLEXPART is being developed continuously !
http://zardoz.nilu.no/~andreas/
Introduction

Lagrangian particle models compute trajectories of a large number of socalled particles (not necessarily representing real particles, but
infinitesimally small air parcels) to describe the transport and diffusion of
tracers in the atmosphere.

The main advantage of Lagrangian models is that, unlike in Eulerian
models, there is no numerical diffusion.

FLEXPART is a Lagrangian particle dispersion model that simulates the
long-range and mesoscale transport, diffusion, dry and wet deposition,
and radioactive decay of tracers released from point, line, area or volume
sources.

FLEXPART can be used forward in time to simulate the dispersion of
tracers from their sources, or backward in time to determine potential
source contributions for given receptors.
What is good in FLEXPART ?

FLEXPART was evaluated using data from three large-scale tracer
experiments, namely the
 Cross-Appalachian Tracer Experiment (CAPTEX),
 Across North America Tracer Experiment (ANATEX)
 European Tracer Experiment (ETEX),
comprising a total of 40 releases.

The results of this validation study are described in Stohl et al. (1998), but
in summary one can say that FLEXPART seems to belong to the better
dispersion models currently available.

This is also supported by the ATMES-II model intercomparison study,
where FLEXPART scored among the best models.

It requires only a short computation time, has a finer spatial resolution and
does not suffer numerical diffusion compared to chemistry transport
models (CTMs).

It is a compromise between simple trajectory calculations and complex
CTMs that makes best use of available computer hardware.

The model is freeware and can be downloaded
System requirements &
versions

FLEXPART, written in FORTRAN 77, is largely platform independent. It
currently runs on SUN, SGI, HP, Compaq Alpha and LINUX workstations.
(also on IBM, done in Reunion Island University)

FLEXPART can be driven with meteorological input data from a variety of
global and regional models, most commonly from the European Centre
for Medium Range Weather Forecasts (ECMWF).

It runs where a Fortran 77 compiler and a GRIB decoding software to
read ECMWF input data.

The memory requirements depend on the spatial domain of your input
fields and the number of particles you want to use.

The ECMWF version of the model is considered as the reference version,
but a new GFS version of FLEXPART is available (porting of the ECMWF
version to use GFS data).
Summary of the available
versions

FLEXPART V6.2 (based on ECMWF input data)



FLEXPART V6.4 for GFS (Contact Caroline Forster : [email protected])
On 24 April 2002 the NCEP AVN model was renamed to the GFS (Global Forecast System)
GFS Products : http://www.nco.ncep.noaa.gov/pmb/products/gfs/



FLEXPART V3.1 for MM5


The Weather Research and Forecasting (WRF) Model is a next-generation mesocale numerical
weather prediction system designed to serve both operational forecasting and atmospheric
research needs.
The effort to develop WRF has been a collaborative partnership, principally among the National
Center for Atmospheric Research (NCAR), the National Oceanic and Atmospheric
Administration (the National Centers for Environmental Prediction (NCEP) and the Forecast
Systems Laboratory (FSL), the Air Force Weather Agency (AFWA), the Naval Research
Laboratory, Oklahoma University, and the Federal Aviation Administration (FAA).
WRF homepage : http://www.wrf-model.org/index.php


Routines for the retrieval of FLEXPART input data from ECMWF
Tools to analyze the output : NCAR Graphics programs and statistical programs available

FLEXPART for WRF (Contact Jerome Fast : [email protected])
Setup of FLEXPART
The pathnames file















A file pathnames must exist in the directory where FLEXPART is started.
It states the pathnames of input and output files:
/home/as/FLEXPART50/options/
/volc/as/contrace/modelresults/forward/
/volc/windcontrace/
/volc/windcontrace/AVAILABLE
/volc/nested/
/volc/nested/AVAILABLE
============================================
Line 1: path where control files "COMMAND" and "RELEASES" are
available
Line 2: name of directory where output files are generated
Line 3: path where meteorological fields are available (mother grid)
Line 4: full filename of "AVAILABLE"-file (mother grid)
Subsequent lines:
Line 2n+3: path where meteorological fields are available (nested grid n)
Line 2n+4: full filename of "AVAILABLE"-file (nested grid n)
The file Includepar

The file includepar contains all relevant FLEXPART parameter settings,
both physical constants and maximum field dimensions.

As the memory required by FLEXPART is determined by the various field
dimensions, it is recommended that they are adjusted to actual needs
before compilation.

To avoid « segmentation fault », change the variables size :
 maxpart(2000)
 maxpoint (1)
 maxrand(2000)
 Compilation : make –f makefile
Use of FLEXPART
The AVAILABLE file

The directory where the meteorological input data are stored, here called
windfields (/volc/windcontrace/ in the above example pathnames file),
contains grib-code files containing the ECMWF data. All meteorological
fields must have the same structure, i.e. the same computational domain
and the same resolution. An example listing of this directory is given
below.

The file names of the grib-code files and their validation dates and times
(in UTC) must be listed in the file AVAILABLE. While it is practical to have
this file reside in the same directory as the wind fields, this is no necessity
and it can also be located elsewhere, as its file name is also given in the
pathnames file.








DATE TIME FILENAME SPECIFICATIONS
YYYYMMDD HHMISS
________ ______ __________ __________
20011028 000000 EN01102800 ON DISC
20011028 030000 EN01102803 ON DISC
20011028 060000 EN01102806 ON DISC
20011028 090000 EN01102809 ON DISC
20011028 120000 EN01102812 ON DISC
Files in directory options

The files in directory options are used to specify the model run.

Very important are :



COMMAND
RELEASES
OUTGRID

File COMMAND :

The most important file is the COMMAND file which specifies (1) the
simulation direction (either forward or backward), (2) the start and (3) the
end time of the simulation, (4) the frequency Tc of the model output, (5)
the averaging time Tc of model output, etc …
Files in directory options

File RELEASES :

RELEASES defines the release specifications :
 The beginning and the ending time of the release,
 Geographical coordinates of the lower left and upper right corners of
the release location,
 type of vertical coordinate (above ground level, or above sea level),
 lower level and upper level of source box,
 the number of particles to be used

The particles are released from random locations within a fourdimensional box extending from the lower to the upper level above a
rectangle (on a lat/lon grid) defined by the geographical coordinates, and
between the release’s start and end.
Files in directory options

The file OUTGRID specifies the output grid, Change if necessary

EXECUTION OF FLEXPART :

./FLEXPART

rather quick, a few dizaines of particles during five days : < 10 minutes
Important aspects of the
physics in FLEXPART
 Mesoscale velocity fluctuations :


Mesoscale motions are neither resolved by the ECMWF data nor covered
by the turbulence
parameterization.

This unresolved spectral interval needs to be taken into account at least
in an approximate way, since mesoscale motions can significantly
accelerate the growth of a dispersing plume (Gupta et al., 1997).

For this, a similar method as Maryon (1998), namely to solve an
independent Langevin equation for the mesoscale wind velocity
fluctuations (“meandering” in Maryon’s terms).

This empirical approach does not describe actual mesoscale phenomena,
but it is similar to the ensemble methods used to assess trajectory
accuracy (Kahl , 1996; Baumann and Stohl, 1997; Stohl, 1998).
Important aspects of the physics in FLEXPART
 Moist Convection :

An important transport mechanism are the updrafts in convective clouds.

They occur in conjunction with downdrafts within the clouds and compensating
subsidence in the cloudfree surroundings.

These convective transports are grid-scale in the vertical, but sub-grid scale in the
horizontal, and are not represented by the ECMWF vertical velocity.

To represent convective transport in a particle dispersion model, it is necessary to
redistribute
particles in the entire vertical column.


For FLEXPART the convective parameterization scheme chosen is from Emanuel
and Zivkovic-Rothman (1999), as it relies on the gridscale temperature and humidity
fields and calculates a displacement matrix providing the necessary mass flux
information for the particle redistribution.

The convective parameterization is switched on using lconvection in file COMMAND.

It’s computation time scales to the square of the number of vertical model levels and
may account for up to 70% of FLEXPART’s computation time using current 60-level
ECMWF data.
Important aspects of the physics in FLEXPART
 Clustered plume trajectories :


In a recent paper, Stohl et al. (2002) proposed a method to condense the complex
and large
FLEXPART output using a cluster analysis (Dorling et al., 1992).

The idea behind this is to cluster, at every output time, the positions of all particles
originating from a release point, and write out only clustered particle positions.

This option can be activated by setting iout to 4 or 5 in file COMMAND.

The number of clusters can be set with the parameter ncluster in file includepar.
 Output :

Condensed particle output using the clustering algorithm is written to the formatted
file trajectories.txt.

Information on the release points (coordinates, release start and end, number of
particles) is written by subroutine openouttraj.f to the beginning of file trajectories.txt.

Subsequently, plumetraj.f writes out a time sequence of the clustering results for
each release point : release point number, time in seconds elapsed since the middle
of the release interval, plume centroid position coordinates, and then for each cluster
the cluster centroid position, the fraction of particles belonging to the cluster, and the
root-mean-square distance of cluster member particles from the cluster centroid.
Case studies of tratospheric
to tropospheric exchanges :
Examples
IRENE
Clustered plume trajectories
REUNION
15/02/02 12TU 9.5-10.5km
Not clustered plume trajectories
19/02/02 12TU 8-9km
Take-home message
 Thanks for your attention !
 You will read the following later :

Access and use of FLEXPART shall impose the following obligations on the user.

The user is granted the right, without any fee or cost, to use, copy, modify, alter,
enhance and distribute FLEXPART, and any derivative works thereof, and its
supporting documentation for any purpose whatsoever, except commercial sales,
provided that this entire notice appears in all copies of the software, derivative works
and supporting documentation.

This software is provided by the University of Munich "as is" and any express or
implied warranties, including but not limited to, the implied warranties of
merchantability and fitness for a particular purpose are disclaimed. In no event shall
the University of Munich be liable for any special, indirect or consequential damages
or any damages whatsoever, including but not limited to claims associated with the
loss of data or profits, which may result from an action in contract, negligence or
other tortious claim that arises out of or in connection with the access, use or
performance of FLEXPART.
Schematic of a lagrangian
model