This menu calls a utility program to convert the HYSPLIT binary concentration file to the DATEM data format and compares the model predicted results with the actual measured data. The Data Archive of Tracer Experiments and Meteorology (DATEM) web page contains additional details about each of the experiments, the data formats, and more complete descriptions of each of the programs used to convert the data and generate the statistical results. The on-line repository has links to several different tracer experiments and their associated meterological data. All experimental data have been converted to a common format. The meteorological data are compatible for direct use by HYSPLIT.
To facilitate application of the analysis software, all the experimental data from the second CAPTEX release are provide in the ./hysplit/datem directory. Included are the control, namelist, measured data, and meteorological data files. The meteorological data is taken from the North American Regional Reanalysis. The computational sequence is as follows:
1) Configure and Run HYSPLIT
In the first section of the menu, a measured data file already in DATEM format should be selected. The input data file represents the HYSPLIT binary concentration output file which is then converted to a DATEM formatted text file using the units conversion factor. The conversion program will match each measurement with a model calculated concentration corresponding to the location and time duration of the measurement. Note that the model temporal output frequency needs to be the same or of finer resolution than the measurement periods, otherwise calculations cannot be matched with the measurements.
The rotation angle entry field (default = 0.0) can be used to test the sensitivity of the model prediction to small changes in wind direction. Setting a non-zero angle will result in the rotation of the calculated plume with respect to the sampling network in the direction specified. The rotation is computed using the first source location specified in the setup menu. The effect of this test is comparable to rotating the transport wind direction by the same amount. However, this sensitivity test only makes sense in the context of plumes that have a relatively homogeneous and stationary structure.
Once the model output file in DATEM format has been created, the statistical evaluation program should be run to compare the model calculations with the measurements in terms of various performance statistics. Two data processing options are available. For verification statistics, all values can be compared, only plume values (both measured and calculated values are greater than zero), or all values but excluding pairs where both measured and calculated values are zero. In terms of averaging, the measured and calculated values can be used directly, temporally averaged resulting in one value for each location, or spatially averaged resulting in one value for each time period. The contingency level is the concentration value required to be exceeded to compute the false alarm rate, probability of detection, and threat-score. The contingency level is also used as the plotting threshold for the measured-calculated scatter diagram which is created by the Scatter Plot menu button.
Note that simulations with multiple species and levels are not supported. The concentration data files follow no specific naming convention and any input or output file name may be selected in the menu. Arbitrary input file names can only be set through the Setup / Grids menu. The statistical output files are always written to stat{A|T|S}.txt where the character A|T|S represents All data, Temporal averaging, or Spatial averaging. An additional output file, data{A|T|S}.txt is created for the scatter diagram plot, which is always named scatter.html. The Rename output field in the GUI adds those text characters between the above prefix and suffix fields: {stat|data}{A|T|S}_{rename}.txt