Although the results were somewhat ambiguous between different meteorological data, the ensemble mean calculations between the two concentration grid resolutions showed about the same performance.

Too broad a conclusion should not be inferred because the physics ensemble calculations were only a subset (26th 06Z to 28th 15Z) of the entire experimental period and too few particles were released in this example.

• Start by creating two new working directories parallel to the existing one
• Perhaps naming them work1 and work2
• Copy the previously saved captex_control.txt and captex_setup.txt into one of those directories
• Renaming the files CONTROL and SETUP.CFG
• Then edit the SETUP.CFG file reducing NUMPAR to 25,000
• Add a line for KRAND=4
• Copy both CONTROL and SETUP.CFG to the other work directory
• Edit each CONTROL file to give the concentration output file a unique name
• Open the command line or terminal, one for each work directory
• Start the HYSPLIT simulation in each by entering ../exec/hycs_std
• When complete, either through the GUI (Utilities / Binary File / Merge) or command line (program = concadd), to combine the two output files

• Distributes particles among processors
• Requires download from mpich.org
• Edit: exec/run_mpi.sh for PATH & MPIPATH
• Edit: conc_mpi.tcl for arglist
• Runs hycm_std, _ens, and matrix