If you are not continuing from the previous section, first retrieve
conc_case_control.txt and conc_case_setup.txt.
We found that both the computation of the stability and the estimate of the resulting turbulence was a function of the mixed layer or boundary layer depth (used interchangeably in these discussions). The default approach is to use the mixed layer depth that is output from the meteorological model. This field is not always available and if missing, the depth is computed from the temperature profile as the height at which the potential temperature exceeds the surface temperature by two degrees.
 To compute the plume using a mixed layer depth computed from the temperature profile, open menu #7, press Reset, and check the radiobutton marked From temperature profile, save the change, run the model, display the results, and then compare the resulting plume with the default calculation using the WRF estimated depths. The temperature profile result is very similar to the default.
 To determine the sensitivity of the calculation to the mixed layer depth, we can force the model to use a constant user set value. A review of the last profile in the MESSAGE file for this last calculation shows the tracer mixed up to about 2500 m. To double the depth, open menu #7 and check the Set as Constant radiobutton and enter the value 5000. The button highlight will disappear when the value is changed from its default, but the calculation will use this value and the resulting plume looks very much like the calculation using the temperature profile except that all the concentrations are about half of their previous values.
 The fixed 5000 m calculation results in a warning message about puff splitting being turned off. The forced larger vertical mixing depth, requires additional puffs because they can no longer be merged if their heights are too different. A larger value for the maximum particle number (menu #4) would eliminate this message.
 To contrast the previous calculation, reduce the mixed layer depth to half its original value. Change the Set as Constant field to 1250, rerun the model, and now the resulting plume is again similar in structure, but now concentrations are higher by a about a factor of two.
In summary, the results show that the magnitude of the air concentration prediction is directly related to the mixed layer depth. However, this conclusion only applies in situations of strong vertical mixing, when the particles interact with this mixing boundary.
