Global backtracking of anthropogenic radionuclides by means of a receptor oriented ensemble dispersion modelling system in support of nuclear-test-ban treaty verification

 

Andreas Becker, Gerhard Wotawa, Lars-Erik De Geer, Petra Seibert, Roland R. Draxler, Craig Sloan, Real D’Amours, Matthew Hort, Hubert Glaab, Philippe Heinrich, Yves Grillon, Vyacheslav Shershakov, Keiichi Katayama, Yuetang Zhang, Paul Stewart, Marcus Hirtl, Michel Jean, Peter Chen

 

Atmospheric Environment, 2007, Volume 41, Pages 4520-4534, doi:10.1016/j.atmosenv.2006.12.048

 

Abstract - In this paper, we introduce a methodology for quality assessment of backtracking models. We present results illustrating the level of agreement between the backtracking models, and the accuracy of each model and the ensemble model in resolving the geo-temporal reference of a single point source. Both assessments are based on an ensemble of 12 different Lagrangian partic le d ispersion modelling (LPDM) systems utilized in receptor oriented (adjoint) mode during two joint numerical experiments dedicated to source region estimation. A major result of this effort is confirmation of the findings of Galmarini et al. (2004b) and Delle Monache and Stull (2003), with regard to the superiority of the ensemble dispersion forecast over any single model forecast, do also apply to LPDM when utilized for backtracking purposes if no or only vague a priori knowledge of the source time is available. In this case, a certain temporal tolerance needs to be added to the source hypothesis tested during the inversion modelling. This is exactly the situation likely to be encountered by the global radionculide monitoring system currently being developed by the Provisional Technical Secretariat (PTS) of the Comprehensive Nuc lear-Test-Ban Treaty Organization (CTBTO), where quick but reliable source location identification needs to be performed. Here, we will therefore introduce a simple methodology as a template for an electronic emergency response system in the field of dispersion modelling.

 

 

Email me the full report