Speaker
Description
Standard atmospheric transport modelling can be of great help for categorizing specific radioxenon observations but will not prevail on average when it comes to screening out nuclear explosion signals that are mixed into the global industrial background. The challenge pertains to both poorly characterized emissions and uncertainties in the modelled dispersion. Improved screening can be attained by applying the FLEXPART-LCM (Linear Chemistry Module). It employs a global domain populated with particles advected throughout the atmosphere. Particle masses can be adjusted using a nudging technique, which computes additional mass tendencies for particles within the observation kernel. We will investigate the following exploitation potentials enabled by FLEXPART-LCM: 1) In case of a significant waveform event nudging could be turned off to avoid assimilating any signals from a nuclear explosion. FLEXPART-LCM would then run freely for a few sampling times. Resulting modelled radioxenon background concentrations at IMS stations can still be expected to be more accurate and nuclear signals should become more evident. 2) A more sophisticated use of FLEXPART-LCM output would be to perform nudging without interruption in case of an event of concern. Rather than suspending the nudging, nudging increments could be continuously analyzed via machine learning to identify anomalies.
| [email protected] |
