Element: Software: Data-Driven Auto-Adaptive Classification of Cryospheric Signatures as Informants for Ice-Dynamic Models

University Of Colorado At Boulder, Boulder CO

Investigators

Abstract

The objective of this project is to develop and automatize a connection between Earth observation data and numerical models of Earth system processes. Both collection of Earth observation data from satellites and modeling of physical processes have seen unprecedented advances in recent years. However, data-derived information is not used to inform modeling in a systematic and automated fashion. This creates a bottleneck that is growing with the data revolution. The award supports the development of a software cyberinfrastructure aimed at reducing this bottleneck by automating classification and parameterization. The proposed cyberinfrastructure will be implemented in a general and transportable way, but its functionality will be demonstrated by addressing a concrete open problem in glaciology: the acceleration during a glacier surge, which is characterized by an increase to 100-200 times the flow normal velocity. Glacial accelerations are important, because they constitute the largest uncertainty in sea-level-rise assessment. The team, from University of Colorado will combine their expertise of field work and data collection with their background in software development to generate a high-quality application that will be made available under an open source license to the broader scientific community. The project will engage graduate and undergraduate students in the software development, thus contributing to the development of future generations of scientists and cyberinfrastructure professionals. The results will also be used to inform activities in K-12 schools and other outreach efforts. The proposed data-driven auto-adaptive classification system is expected to provide a tool to the Earth Sciences community that allows it to employ the unprecedented detail in satellite image and SAR data (WordView, Sentinel-1 and Sentinel-2) necessary to extract information on surface properties and processes that were previously indiscernible. In that the classification will automatically adapt to changing conditions in time and space, it will provide a consistent parameterization of spatial processes that can be used to drive numerical simulations of Earth system processes. Thus, a direct connection will be established between data analysis and modeling. In a pilot study, the data-modeling connection will be demonstrated through classification of crevasse patterns, which result from deformation, and optimization of the basal sliding parameter in a three-dimensional model of a glacier surge. Hence the pilot study will advance understanding of ice dynamics. A second application is a sea-ice classification system, aimed to aid in mapping and understanding the changing Arctic sea-ice cover. The planned automated connection between data-driven automated classification and optimization of model parameters is expected to lay the foundation for a transformation of the data-modeling world in Earth sciences, atmospheric and polar sciences. The project will also advance machine learning and spatial statistics through realization of a science-driven approach to computer science and cyberinfrastructure. This award by the NSF Office of Advanced Cyberinfrastructure (OAC) is jointly supported by the Cross-Cutting Program within the NSF Directorate for Geosciences, The OAC Cyberinfrastructure for Emerging Science and Engineering Research (CESER) program and the EarthCube Program jointly sponsored by the NSF Directorate for Geosciences and the OAC. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.