16.3 Assimilation of Satellite Data




In the previous section, the last valid time of the particle position file was 12 UTC on the 15th of April. We can compare the particle positions from the calculations in the last section with the aerosol index from the NASA TOMS satellite archives available on-line. These are sun synchronous orbits which means that the pass over Iceland-England would occur around local solar noon or about 1200 UTC.

  1. Start by pressing the Advanced / Satellite Data / FTP NASA TOMS menu tab to open the download menu. Set your email address for the anonymous FTP password field and enter the download date to 10 04 15 and press Get Data File. A pop-up message will appear indicating the name of the file accessed. All files get renamed to a standard convention. This particular file will be renamed to ga100415.txt. In the event the FTP fails, copy the file from \Tutorial\volcano to the \Hysplit4\working directory.

  2. Before we open the satellite data viewer, go to the working directory and rename the concentration output file cdump to another name, such as cdump_100415. This is not necessary, but the viewer will display the satellite data, particle file, and concentration file. It is easier to view only two fields rather than three.

  3. Open the Advanced / Satellite Data / TOMS viewer. No changes are required from the default settings. It works in a similar manner to the Quick Start interface. Initially the screen only shows a map background. Press and hold the left mouse button while dragging the pointer from the lower-left to upper-right corners of the desired map zoom area. In this case the region between Iceland and Europe. Then right-click the mouse to redraw the map in the zoomed box. Subsequent left-clicks will advance the display to the next time period. The particle-satellite display should show the bulk of the particles on the northern edge of the higher Aerosol Index grid cells.

  4. Actually the overlap between the model calculation and observations is quite good. The small difference for the first 1000 km may be just due to variations in release height. There is a procedure we can use to adjust the particle positions before continuing the calculation. Press the Concentration / Utilities / Particle Adjustment menu tab to open the shift menu.

  5. There are two adjustment options, defining a window and translating particles, or rotating them about a point. We will choose the rotation option. Select the rotation radio-button. The rotation point defaults to the source location and enter a rotation angle of 10 degrees (positive = clockwise) to a distance of 1000 km. All particles within this distance will be rotated. The menu already defaults to writing the adjusted particles to PARINIT which would then default to the input file for the next simulation cycle. Process the points and the script will automatically open the particle display where there is an obvious discontinuity at 1000 km. This could be eliminated by selecting the blend option.

  6. Go back and run the particle-satellite viewer, but this time enter the PARINIT file in the particle file text entry field to see how the adjusted particle positions compare with the observations. It appears that perhaps 10 degrees was too much and 5 degrees would have been a better choice.

On your own try a few of the other adjustment options and then restart the calculation. The fit with observations for this case is somewhat fortuitous as no particular effort was made to get detailed source amount or release height information. Note that the TOMS viewer is only available for Windows while the particle adjustment procedure can be run on all computing platforms. Although the GUI is linked to the TOMS data files, there are other public sites available with OMI and MODIS image data.