T212 Machine Learning in Hydrogeology

Thursday, 29 October 2020
1:30 PM - 5:30 PM
Meeting Rooms

DOWNSCALING OF GRACE TWSA IN HIGHLY STRESSED AQUIFER SYSTEMS, THE CENTRAL VALLEY AQUIFER, USING MACHINE LEARNING

Paper number: T212-5
Program: Technical Sessions
Time: 2:20 PM – 2:35 PM

Authors

Abstract

This study will present a machine learning (ML) - based downscaling algorithm that produces a higher spatial resolution groundwater level anomaly (GWLA) from the GRACE data by utilizing the relationship between Terrestrial Water Storage Anomaly (TWSA) from GRACE and other land surface and hydro-climatic variables (e.g. precipitation) in aquifers where there is significant amount of abstraction of groundwater. Different studies have been undertaken to downscale GRACE in various aquifer systems. But, the studies did not consider groundwater withdrawal information. This potentially reduce the accuracy of the downscaled product, specifically in highly anthropogenic stressed aquifer systems. This study will attempt to address the above limitation in existing downscaling models, specifically, the lack of consideration of groundwater withdrawal (human impact). Therefore, the objective is to map high-resolution Groundwater Level Anomaly (GLWA) from GRACE TWSA in anthropogenically stressed aquifer system, the Central Valley Aquifer (CVA) by considering anthropogenic stress indicator variable in the downscaling model using machine learning (ML). Withdrawal of groundwater from clay dominated aquifer systems could result in land subsidence. Ground deformation (subsidence) can be measured using various approaches including satellite-based remote sensing, in-situ/satellite-based Global positioning system (GPS) measurements, and in-situ extensometers. This study will utilize combination of CGPS, InSAR, and Extensometer data to identify groundwater withdrawal induced land deformation and use the result to represent human impact in the ML model. Other land surface and hydroclimate variables including Geology, Land cover and land use, precipitation, stream flow, and soil moisture will be integrated in the model. The downscaled GWLA data will be used to determine the spatial distribution of groundwater storage anomalies relative to pumping centers and depleted regions. It is expected that uncertainty in the downscaled product will be reduced as the result of incorporating subsidence as human impact proxy during the downscaling process.

Geological Society of America Abstracts with Programs. Vol. 52, No. 6

doi: 10.1130/abs/2020AM-358858

© Copyright 2020 The Geological Society of America (GSA), all rights reserved. Permission is hereby granted to the author(s) of this abstract to reproduce and distribute it freely, for noncommercial purposes. Permission is hereby granted to any individual scientist to download a single copy of this electronic file and reproduce up to 20 paper copies for noncommercial purposes advancing science and education, including classroom use, providing all reproductions include the complete content shown here, including the author information. All other forms of reproduction and/or transmittal are prohibited without written permission from GSA Copyright Permissions.