Integrating Diverse Calibration Products To Improve Seismic Location

The monitoring of nuclear explosions on a global basis requires accurate event locations. As an example, under the Comprehensive Nuclear-Test-Ban Treaty, the size of an on-site inspection search area is 1,000 square kilometers or approximately 17 km accuracy, assuming a circular area. Achieving this level of accuracy is a significant challenge for small events that are recorded using a sparse regional network. In such cases, the travel time of seismic energy is strongly affected by crustal and upper mantle heterogeneity and large biases can result. This can lead to large systematic errors in location and, more importantly, to invalid error bounds associated with location estimates. Corrections can be developed and integrated to correct for these biases. These path corrections take the form of both three-dimensional model corrections along with three-dimensional empirically based travel-time corrections. LLNL is currently working to integrate a diverse set of velocity model and empirically based travel-time products that are provided internally by the DOE laboratories and externally through contracts into one consistent and validated calibration set. To perform this task, we have developed a hybrid approach that uses three-dimensional model corrections for a region and then empirically uses reference events when available to improve these path corrections. This empirical approach starts with the best a priori three-dimensional velocity model that is produced for a local region and uses this as a baseline correction. When multiple competing models are provided from various sources for a local region, uncertainties in the models are compared against each other using our ground truth data and our analysts picks. Based on the results of this comparison an optimal model is selected. We are in the process of combining three dimensional models on a region-by-region basis and integrating the uncertainties to form a global correction set.