← Back to Sessions

Predicting Atmospheric Structure for Infrasound Propagation Using Machine Learning

Session: Infrasound and the Seismo-Acoustic Wavefield II

Type: Oral

Date: 4/21/2021

Presentation Time: 10:00 AM Pacific

Description: 

Infrasound signals travel through the atmosphere, a dynamic medium. Therefore, it is important to have accurate atmospheric specification information for modeling infrasound propagation. The National Centers for Environmental Prediction (NCEP) and the European Centre for Medium-Range Weather Forecasts’ (ECMWF) weather forecast models and reanalysis products are often used to determine atmospheric profiles for infrasound propagation modeling before or after an event of interest. These models are based on physical principles but contain uncertainties that lead to inaccuracies in infrasound propagation modeling, especially for low signal-to-noise ratio (SNR), quickly-attenuating, waveforms generated by small events. Machine learning approaches have proven useful in predicting tropospheric weather. Using a Long Short-Term Memory (LSTM) network we predict the temperature, wind speed, and wind direction of the lower and middle atmosphere at radiosonde stations locations throughout the conterminous United States. We train the model using 10 years of historical radiosonde weather observations from 2009-2019 and predict temperature, wind speed, and wind direction up to an altitude of ~50 km. Predictions are made for 24 and 48 hours in the future and outperform the ECMWF weather reanalysis products for the same time and location.

This research was funded by the National Nuclear Security Administration, Defense Nuclear Nonproliferation Research and Development (NNSA DNN R&D). The authors acknowledge important interdisciplinary collaboration with scientists and engineers from LANL, LLNL, MSTS, PNNL, and SNL. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Presenting Author: Sarah A. Albert

Student Presenter: No

Authors