16.4 Particle Size Distributions

In this example we will rerun the original volcanic ash simulation with a particle size distribution rather than just four particle sizes. If you have not already run that simulation, go back to that section and review the calculation procedure. You can also just retrieve control_volc.txt and setup_volc.txt files into the GUI from the volcano directory.

1. Before running the base volcanic ash calculation again, delete any PARINIT files in the working directory, then open the concentration grid menu and add the deposition level and then save and run. After the simulation completes, open the display menu and select all for the pollutant and force the contours incrementally (500000 down to 1000) to optimize the deposition plot. This will be the reference calculation. Naturally the other results are the same as before.


2. One of the namelist options NBPTYP defines the number of bins assigned to each particle size in the pollutant section of the CONTROL file. The default value is 1 so that each particle size is associated with only one bin. A value larger than one will create that number of particle size bins centered about each value in the CONTROL file. The program creates the mass distribution for a range of particle sizes given just a few points within the distribution. We assume that dV/d(log R) is linear between the defined points for an increasing cumulative mass distribution with respect to particle diameter. The input points in the CONTROL file should be sorted by increasing particle size.


3. The NBPTYP is one of the few variables that cannot be set in the GUI. You have just completed the basic volcanic ash simulation with four different particle sizes, then go to the /hysplit/working/ directory and edit SETUP.CFG and add the line NBPTYP=5 to define five particle sizes for each pollutant (size) defined in the CONTROL file. Save and then go ahead and run the model.


4. The standard output will show that there has been a particle size redistribution and when the run completes the MESSAGE shows the size and emission rate for the 5 particles assigned to each of the four original particle size bins.


5. Now display the deposition results for the last time period and although the differences between this simulation using 20 particle sizes versus the original one using 4 should be much more evident than it would be for air concentrations, or the maps of the particle positions, the results are very comparable, suggesting the original four particle size bin distribution was adequate for these simulations.

A more refined simulation providing for the automatic redistribution of the particle sizes into multiple bins results in a less discontinuous output, especially for the deposition, as the larger particles fall out of the calculation more rapidly than the smaller ones. This may be more evident for calculations where the particle size distribution changes with height.