15.2 Long-Range I-131 from Fukushima NPP




In this section we will configure a simulation of the I-131 plume from the Fukushima Daiichi accident. There are still many uncertainties regarding the radionuclides that were released, their amount, and the timing of the releases. Consult on-line sources for more recent information. Model results will be compared with gaseous I-131 measurements made at Dutch Harbor, AK, available from the US EPA RADNET sampling network. To avoid unintended options, start by pressing the Reset button on the main menu to clear all previous settings. To save time, the simulation files have been saved in the Japan directory as japan_control.txt and japan_setup.txt. We will review the required changes below.

  1. We will assume that most of the I-131 emissions occured during the venting and explosion at Unit #2 from 1200 UTC on March 14th 2011 through 1200 UTC on the 15th. The total I-131 emitted has been reported to be about 150 PBq, which converts to an hourly rate of about 5 PBq. Open the concentration setup menu and set the model start time to 11 03 14 00 and the starting location to 37.4206 141.0329 100.0 for a run duration of 196 hours (8 days). Set the meteorology file to use the 2.5 degree global reanalysis RP201103.bin in the japan directory. The concentration grid should be set to a resolution of 1.0 degree with a span of 181.0 by 360.0 degrees.

  2. Before proceeding, if this tutorial is being run through the web rather from a CD, it will be necessary to download the meteorological data file RP201103.bin for this example. Any pre-configured CONTROL files used in this example will need to be edited to point to the location of the downloaded data file.

  3. Open the pollutant menu and set the pollutant name to I131 and the release rate to 5.0E+15 for a duration of 24.0 hours. The emission start time should be set to start 12 hours after the simulation start 11 03 14 12 00.

  4. Open the grids menu and set the concentration grid to a horizontal resolution of 1.0 degree with a vertical depth of 500 m with an output averaging time of 24 hours to a file name of fukushima.bin.

  5. The deposition section is the last menu in this group, select the first I-131 radiobutton (for gaseous rather than particulate iodine), which then automatically sets the dry deposition velocity to 0.01, a Henry's constant of 3.0,and a half-life of 8.04 days.

  6. Before starting the simulation, open the advanced configuration menu and set the particle release number to 24000, the maximum particle number to 25000, and force the time step to 30 minutes to speed up the simulation. When all the changes have been made, save, and run the simulation.

  7. When the run completes, open the concentration / utilities / convert to station menu and either load the samplers.txt file from the japan directory, or just enter the Dutch Harbor location 53.903 -166.511 in the menu. Enter the units multiplier of 1000.0 to convert the computational units of Bq/m3 to (milli) mBq/m3, and define the supplemental data file Dutch_Harbor.txt to display the I-131 measurements with the model prediction. Extract the data to fukushima1.txt and create the time series plot. The results show a predicted model peak about five times less than the measurements but arriving about the same time. A sample batch file con2stn.bat is provided to extract the data and plot the results for Dutch Harbor.

  8. If you have the time (this run takes longer), the simulation should be run again using the finer resolution (1/2 deg, 3 hr) global GDAS meteorology (gdas11-22.bin) extracted over the Pacific region for the first ten days (11th-12Z to the 22nd-12Z) of March 2011. Run the model and extract the data to fukushima2.txt. The results now show a predicted model peak about two times larger than the measurements and arriving about a day earlier. Although not discussed above, did you correctly set the units for the ordinate labels on the time series plots?

Contrasting the two identical simulations but with different meteorology, it is clear that the air concentration results are very sensitive to the meteorology and processes that might affect the wet and dry scavenging. The I-131 results presented here are from a very simplified simulation. A better fit with the observations may be achieved with a more complex temporal emission sequence, even finer spatial resolution meteorology and concentration output, and perhaps a greater particle number release rate.