Highlights 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001

GeoSCIFramework: Real-time Analytics and Machine Learning for Geoscience and Hazards Research

  • Principal Investigator(s): Charles Meertens, Dave Mencin, and Scott Baker, UNAVCO; Christy Tiampo, University of Colorado - Boulder; Diego Melgar, University of Oregon; and Ivan Rodero, Rutgers University – New Brunswick
  • UNAVCO Staff: Charles Meertens, David Mencin, Kathleen Hodgkinson, Scott Baker
  • Dates: January 1, 2019 - December 31, 2022 (estimated)
  • Location(s): Cascadia and Yellowstone
  • Funding Source: NSF Award 1835791
  • Proposal Name: Collaborative Research:Framework:Data:NSCI:
    HDR:GeoSCIFramework: Scalable Real-time Streaming Analytics and Machine Learning for Geoscience and Hazards Research

Written by the GeoSciFramework team
12 November 2019

Project Overview

The GeoSCIFramework project enables geohazards research and ways to improve earthquake, tsunami and volcano early warning systems through real-time analysis using machine learning. The innovative approach provides real-time streaming analytics on continuous data streams from thousands of high-rate geophysical sensors throughout much of North America. These sensors comprise geodetic and seismic networks currently managed by the NSF GAGE and SAGE facilities, operated by UNAVCO and IRIS, respectively, as well as the NSF Ocean Observatories Initiative (OOI) cabled array data managed by Rutgers University. When combined with satellite radar time series derived from NASA and foreign space agencies observations, this new computational framework gives a coherent high-resolution global-scale view of the motions of the earth.

Existing Tools, New Applications

This project is a collaboration of computer scientists and geoscientists from UNAVCO, the University of Colorado, University of Oregon, and Rutgers University. The planned framework draws from widely adopted free and open source software packages developed to support Internet searches, social media, and intelligence gathering, but not commonly used in the geosciences. Using these technologies to access and analyze very large and growing geoscience datasets, geoscientists can test and enhance algorithms for event early warning while improving their understanding and modeling of natural hazards.

How It Works

Machine learning algorithms and spatio-temporal analyses will be trained using past events and informed by physics-based models. This method supports not only the automatic detection and characterization of rapid events such as earthquakes and tsunamis but also slow-slip events or magmatic intrusions that evolve over a longer period of time, expanding the potential for new scientific discoveries.

The initial use cases are the Cascadia subduction zone and the Yellowstone caldera where the science team has expertise and where there are concentrations of NSF-managed and OOI geophysical instruments and satellite radar acquisitions.

Integral to the project plan is development, documentation and training using collaborative online resources such as GitLab and Jupyter Notebooks, and utilizing NSF XSEDE resources to make larger (petabyte-scale) datasets and computational resources more widely available to students and researchers outside of the project.

Cross-Cutting Funding

This award by the NSF Office of Advanced Cyberinfrastructure is jointly supported by the Cross-Cutting Program within the NSF Directorate for Geosciences, the Big Data Science and Engineering Program within the Directorate for Computer and Information Science and Engineering, and the EarthCube Program jointly sponsored by the NSF Directorate for Geosciences and the Office of Advanced Cyberinfrastructure.

Related Links

Map Center:
Boulder, Colorado


Last modified: 2020-01-28  22:54:06  America/Denver