HYSPLIT Forum: hysplitbbs.arl.noaa.gov

We have simulated the Chornobyl Cs-137 accidental releases (happened in April-May of 1986) on the regional scale (mainly on the territory of Ukraine). We have used different sets of input meteorology and parameterizations of the source term. Some of the results (calculated cumulative/total deposition patterns) are attached below. The pattern obtained with the particle mode of HYSPLIT (10000 particles are released every emission cycle) is shown on Fig1. Fig2 and Fig3 contain similar results obtained with the pure puff (Gaussian) mode, with 1 and 100 puffs released every emission cycle respectively. Any other input information (meteorology, source, CONTROL file etс. ) was the same for these three simulations. However, as it can be seen from the figures the results calculated with the pure puff mode show underestimation.

Might the underestimation be related to small number of puffs? What is an optimal number of puffs to obtain the same results as by using the particle mode of HYSPLIT. How can we speed up HYSPLIT calculations in the pure puff mode?

Please check your MESSAGE file to see if you run out of puff after splitting. If that occurred you should get a Warning message in the MESSAGE file.