Now that we have successfully configured an air concentration calculation and compared the results with the ground-level sampling data from CAPTEX release #2, we will configure the model to cover a more limited time period to correspond with the initial aircraft sampling in order to quickly run some sensitivity studies with different dispersion parameters and turbulence settings. First press the center Reset button on the main GUI to clear all previous changes and then retrieve captex_control.txt into the Setup Run menu and captex_setup.txt into the Concentration Configuration menu. This will be the starting point for all the case study customizations.
- To run through many model combinations as quickly as possible and limit the complexity of the analysis, we will only compare the model predictions to the first complete aircraft sampling pass about 200 km downwind from the tracer release point. See the CAPTEX report for a complete description of all the data. This first pass occurred around 0300 UTC on September 26th. Open the Concentration / Setup Run menu and change the run duration from 68 hours to 13 hours, enough time to take the calculation through 0600 on the 26th.
- The short range focus of this example suggests that the higher resolution (hourly) WRF meteorology would be better for testing the model's turbulence and mixing assumptions. Clear the meteorology files menu and add the captex2_wrf27.bin data file.
- To reduce the number of graphical output frames, open the Concentration Grid #1 menu and change the sampling start time from 83 09 25 18 00 to 83 09 26 03 00. This insures that the binary concentration output file will only contain one 3-hour average air concentration period, covering the last three hours of the simulation. Press Save to close all menus and then press the Run Model tab.
- When the run has completed, press the Display / Concentration / Contours menu tab to open the graphics menu. Make the same changes as in previous examples, that is the concentration multiplier should be 1.0E+12 and the output units will be pg. Three additional changes are suggested.
- First, force the map center further to the northeast because we know the plume is going in that direction, rather than over the tracer release point, by selecting the map center Set checkbox and entering a new position of 41.0 -82.5. In this way the map is centered over the plume.
- Second, because we have a sense of the range of potential concentrations, we will set the contours by checking the UserSet radiobutton and enter the following contours using a plus symbol as a delimiter: 50000+20000+10000+5000+2000+1000+500.
- Third, increase the zoom slider bar to 80 to reduce the area outside of the contoured regions.
- Then Execute Display to open the graphic. Note the large red square on the display. This represents the size of the grid cell at the location with the maximum concentration (2.0E+04 pg/m3) noted in the legend on the right side of the graphic.
- In the trajectory section, we started a backward trajectory from the maximum concentration point defined in the 914 m MSL aircraft sampling flight. This file is in the DATEM format. This aircraft pass managed to cover the entire horizontal extent of the plume. The concentration plotting program we have been using, concplot, has a command line option to plot the values from a DATEM formatted file. Note that each aircraft sample has a duration of 6 min and therefore the start and stop times from the original 914 m MSL data file were edited to match the model output period so that concplot can match up the measurement times with the model output times. This edited file, data_case.txt, is provided in the CAPTEX directory and should be selected from the DATEM plot file browse button in the concentration display menu to show the measurement data and the model contours on the same graphic.
The default model simulation showed that the ground-level air concentrations closely matched the position, extent, and magnitude of the plume observed by the aircraft during the first downwind cross-section flight. In the next section, the model simulation and display parameters will be refined further. In general you will find that the air concentration results are not very sensitive to changing the turbulence or mixing parameterizations.