The matrix approach to source attribution is set
up in a similar manner to the previous discussion but run with
multiple sources. Use three similar, although shifted northwest, starting locations
as with the dust storm simulation to define the source region
(20N-40N and 25W-0W). Using the FNLDUST data, run (500) 3D-particles for 96
hours with concentration output at 100 m every 6 hours (remember to set the output
level back to 100 meters if you still have it set at 0 m from the deposition
case and to turn off deposition), set the maximum number of particles to at least 125000, and set the Grid Center lat/lon to 30N,10W.
Reduce the resolution of the concentration grid to 0.75 to reduce the memory requirements. Prior to executing
the model through the Run Matrix tab of the Concentration / Special Simulations
menu, it is necessary to check the Restructure the concentration grid to the source-receptor
format button in the Advanced
/ Configuration Setup / Concentration / In-line chemical conversion modules menu. This causes the concentration grid to be reconfigured
so that every source location within the matrix (546 in this example) will have its own concentration
grid. This run will take a few minutes to complete.

The final concentration output file can only be
displayed through the Matrix tab of the Display Options menu.
The term “matrix” has two connotations with respect to HYSPLIT. In the earlier application,
a matrix of sources was created. These results are summed to a
single concentration file. In this application, defining a concentration
grid for each source creates a matrix of sources and receptors. This
requires a special preprocessor for the display program. In this
example, selecting the receptor button causes the entered latitude-longitude
value to be treated as the receptor point. The dispersion factors
from each source to that receptor are then interpolated to a grid
and displayed by time period. The last time period (0000 UTC, February 27), coincident
with the TOMS image, is shown below. |
 |

|