STILT Particle Model Prediction of Air Parcel
Locations (click here for
explanation):
*NEW* STILT forecasts for CERES 2007 *NEW*
Forecasts of air parcel
locations using particle simulations from the STILT (Stochastic
Time-Inverted Lagrangian Transport) model
STILT is an off-line Lagrangian particle dispersion model that
simulates the transport of air parcels with ensembles of fictitious
particles advected with mean wind velocities as well as stochastic
velocities parameterized to capture the effects of turbulent transport.
The particle ensemble is released at a receptor location and
transported backward in time, representing the air parcels
arriving (in the forward-time sense) at the receptor. Thus the
particles represent influence, marking out the upstream regions
affecting the receptor. The backward-time formulation yields the
same information with significantly reduced computational effort than
the forward-time alternative, generating the spatial- and
temporal-dependence of the upstream influence with a single particle
simulation rather than numerous forward-time simulations from all
potential upstream regions and all timesteps
Wind fields used by STILT
For the CarboEurope Regional Experiment STILT is driven by wind fields
from the following
models:
MeteoFrance
Aladin (10 km, 72
hrs)
ECMWF (0.35 deg
, 84 hrs)
NCEP GFS
(1-deg, 180 hrs, downloaded from NOAA
Air Resources Laboratory)
Output from STILT
The output from the different wind fields can be accessed through corresponding pull-down menues.
Also each receptor (time and location for the experiment) can be specified through these menues.
The outputs are divided into three kinds of
files, with names:
[modelname]MMDDHH.MMDDHH.png
[modelname]MMDDHH.MMDDHHprofile.png
[modelname]MMDDHH.MMDDHH.txt
(1) MAP OF PARTICLE LOCATIONS
The *.png files show maps of the STILT particle locations, started
backward in time. The first set of MMDDHH in the filename, to the
left of the '.' , shows the starting time of the particles in UT. The
second set of MMDDHH in the filename, to the right of the '.', denotes
the time in UT when the wind fields were generated. For example,
"eta051723.051612.gif" is a file showing the output from a run started
at 5/17 UT23 using the Eta wind fields at 5/16 UT12.
Here is an example:
(2) COLUMN HISTORY
The *profile.png files show 3D views of particles started
backward in time from a column above the receptor (measurement location).
The filename convention is similar to the one used for the map above.
Here is an example:
(3) SUMMARY OF PARTICLE STATISTICS
The *.txt files are text files summarizing the statistics of particle
locations at different times to aid in determining flight paths.
The file naming convention is the same as that of the *.png
files.
A sample file looks like the following:
|
time | xsec[nmi] |
x1.1st | y1.1st | x2.1st |
y2.1st | xsec2[nmi] |
ctr.x | ctr.y | djeffco[nmi] |
|
120 | 32.0204 | -105.0317 |
39.6061 | -104.6403 | 40.0468 |
2.7779 | -104.898 | 39.7567 |
14.5282 |
|
240 | 92.8335 | -103.5044 |
40.7122 | -105.1449 | 39.806 | 19.389 | -104.5067 | 40.1585 |
32.5546 |
|
480 | 127.8046 | -102.5625 |
41.0273 | -105.1231 | 40.1634 |
30.8476 | -104.074 | 40.5173 |
60.7643 |
|
960 | 100.7624 | -100.8918 |
40.8833 | -103.1077 | 41.0225 |
43.5653 | -102.1303 | 40.9611 | 151.0362 |
|
1440 | 364.2301 | -100.2913 |
38.4183 | -99.7011 | 44.4718 |
161.8586 | -99.9529 | 41.889 | 263.2645 |
-"time" is the timestep of the model
-"xsec[nmi]" is the length of the cross-section [nautical miles] in the
direction of the 1st principal component axis that contains
90% of the particles
-"x1.1st" and "y1.1st" refer to LON and LAT of one end of the 1st
principal component axis (5%tile of particle location projected into
this axis)
-"x2.1st" and "y2.1st" refer to LON and LAT of the other end of the 1st
principal component axis (95%tile of particle location projected into
this axis )
-"xsec2[nmi]" is the length of the cross-section [nautical miles] in
the direction of the 2nd principal component axis that contains 90% of
the particles
-"ctr.x" and "ctr.y" are the LON and LAT of the center of mass of the
particle ensemble at each timestep
-"dairportname[nmi]" refers to the distance between the airport and the
center of mass of the particle ensemble at each timestep
Urban Plume Studies
In the case of directories named for urban areas (e.g., "Toulouse"), the
output is somewhat different because the flights would
chase after plumes emitted from these areas instead of sampling
upstream air parcels arriving at these areas. In this case
particles are emitted from the urban area and transported forward
in time, but started at times up to 24 hours before the flight time.
Reference
Lin, J. C., C. Gerbig, S. C. Wofsy, A. E. Andrews, B. C. Daube, K. J.
Davis, and C. A. Grainger, A Near-field Tool for Simulating the
Upstream
Influence of Atmospheric Observations: The Stochastic Time-inverted
Lagrangian Transport (STILT) Model, J. Geophys. Res., 108
(D16), 4493, doi:10.1029/2002JD003161, 2003. [Full
Text (pdf)]
Contact:
Christoph Gerbig cgerbig@bgc-jena.mpg.de
|
