28 June 2021 to 2 July 2021
Europe/Vienna timezone

BazNet: A Deep Neural Network for Confident Three-component Backazimuth Prediction

1 Jul 2021, 09:00


e-Poster T3.6 - Artificial Intelligence and Machine Learning T3.6 e-poster session


Mr Joshua Dickey (Air Force Technical Applications Center (AFTAC), FL, USA)


Three-component stations traditionally rely on polarization analysis to estimate the backazimuth of each arriving wave. Unfortunately, these polarization estimates suffer from both high error and low confidence, and contribute very little to the downstream association algorithms at the IDC. Here, we present BazNet, a deep neural-network-based backazimuth predictor for three-component stations. For existing stations with ample historical training data, the technique achieves an overall median absolute error of around 14◦, a modest improvement over polarization. More importantly, each estimate is accompanied by a robust certainty measure, which is highly covariant with the error. By integrating the BazNet predictions and certainties into NETVISA, we demonstrate the potential of this algorithm to enhance global association at the IDC.

Promotional text

This work explores the use of a temporal convolutional neural network architecture for improved three-component backazimuth estimation, potentially enhancing the seismic signal processing pipeline used at the IDC for nuclear test monitoring and verification.

Primary authors

Mr Joshua Dickey (Air Force Technical Applications Center (AFTAC), FL, USA) Ms Geeta Arora (Bayesian Logic, Inc., CA, USA) Mr Nimar Arora (Bayesian Logic, Inc., CA, USA) Ms Megan Slinkard (CTBTO Preparatory Commission, Vienna, Austria) Mr Noriyuki Kushida (CTBTO Preparatory Commission, Vienna, Austria) Mr Ronan Le Bras (CTBTO Preparatory Commission, Vienna, Austria)

Presentation materials