Abstract: Based on tracer test data in a nuclear power plant site and combined with the MADS optimization method, we compare the differences of diffusion parameters obtained with data processing methods of single sampling average, single release average, single sampling one time optimization, and single release one time optimization. It turns out that, the calculated diffusion concertrations are close to the measured in general, the data is credible and can reflect the environmental characteristics of atmospheric diffusion parameters of the site. The results of data processing based on single sampling are not as credible as those from single release. Both processing methods of single release provide similar and relatively credible results, which can be selected for use according to other environmental characteristics of the site or from the conservative aspects of dispersion results.

Key words: MADS, Optimization, Tracer test, Atmospheric diffusion parameters