T212 Machine Learning in Hydrogeology

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

HIGH-RESOLUTION NATIONAL-SCALE HYDROGEOLOGIC DATASETS IN SUPPORT OF MACHINE LEARNING MODELS

Paper number: T212-3
Program: Technical Sessions
Time: 1:50 PM – 2:05 PM

Authors

  • Kenneth Belitz Presenting Author
    Earth System Processes Division, Carlisle, MA

Abstract

Machine learning (ML) models are empirical. The models are ”trained” so that a response variable can be estimated by a set of independent variables. Consequently, the relevance and accuracy of an ML model depends to a great extent on the relevance and accuracy of the underlying independent variables. And if the goal of modeling is the mapping of groundwater resources, then the independent variables must be available across the entire extent of the area of interest, and, in many cases, mapped as a function of depth. The USGS NAWQA project has been developing high-resolution national scale hydrogeologic datasets to support the mapping of groundwater resources across the conterminous U.S. (CONUS). These datasets are essential because groundwater provides about 45% of the U.S. drinking water supply.

Completed work includes datasets describing the hydrogeologic framework of the CONUS: (1) The mapping of secondary hydrogeologic regions (SHRs) which are a complement to Principal Aquifers (PAs). With the mapping of SHRS, all areas of the CONUS belong to either a PA or an SHR. (2) The mapping of hydrogeologic terrains across the glacial aquifer system. Terrains are areas where the Quaternary sediments are of similar thickness, texture and depositional age, and have been shown to have predictive value for the characterization of groundwater quality. And, (3) the mapping of multi-order hydrologic position (MOHP) at a resolution of 30 meters. There are nine hydrologic orders and two metrics per order (lateral position and distance from stream-to-divide). In the context of ML models, MOHP has been shown to be an indicator of process for lower orders and an indicator of location for higher orders.

Completed and ongoing work includes datasets describing the location of groundwater resources used for drinking water supply across the CONUS: (1) Areal location, at a resolution of 100 meters, of the number of people dependent on domestic wells; (2) areal location, at a resolution of 1 kilometer, of the number of people dependent on public supply wells, and of the PA or SHR that provides water to those people; and (3) maps, at a resolution of 1 kilometer, of the depths to the tops and bottoms of the zone used for public supply wells, and the zone used for domestic supply. These datasets provide context for assessing groundwater resources across the CONUS.

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

doi: 10.1130/abs/2020AM-357568

© 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.