T212 Machine Learning in Hydrogeology

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

FLOODING DRIVES GROUNDWATER ARSENIC CONTAMINATION IN SOUTHEAST ASIA

Paper number: T212-16
Program: Technical Sessions
Time: 4:45 PM – 5:00 PM

Authors

Abstract

Geogenic arsenic (As) contamination in groundwater poses a significant health risk for millions of people living on floodplains in Southeast Asia. An inherent characteristic of alluvial and deltaic floodplain aquifers in Southeast Asia is that dissolved groundwater As concentrations vary markedly over small lateral (10-100m) and vertical (0-10m) spatial scales, making it challenging to identify key variables that effectively capture the principal biogeochemical and hydrologic processes linked to As heterogeneity. Improved understanding of these key variables is needed to predict groundwater As concentrations at the household level with enough accuracy to advise mitigation and preventative measures. We found that the occurrence and frequency of surface flooding serve as master variables, uniting the coupled biogeochemical, hydrologic, and geomorphic features that control As levels in shallow (<75m) aquifers across Southeast Asia. The magnitude of surface water occurrence and inter-annual recurrence (i.e., temporal frequency) effectively captures floodplain hydrologic conditions linked to the availability of organic matter and other important geochemical reactants, which in turn, affect aquifer redox status and drives As mobilization. Remotely-sensed measures of flooding occurrence and frequency were used with a Random Forest model to accurately predict groundwater As concentrations spanning three orders of magnitude (~ 1 to 1000 ug/L) as well as the probability that As levels exceed 10, 50, and 100 ug/L in Cambodia, Vietnam, and Bangladesh. We demonstrate that flooding is directly linked to the environmental conditions that regulate As levels within alluvial and deltaic floodplain aquifers of Southeast Asia, and thus can serve as principal features of statistical models to predict site-specific concentrations of As in groundwater. We anticipate our approach can be reliably applied to similar aquifer systems worldwide to evaluate potential levels of groundwater As contamination and the associated health risks of As exposure with improved accuracy.

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

doi: 10.1130/abs/2020AM-354743

© 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.