Finding the needle in the haystack
Machine learning approach in the search for arsenic hotspots
ME Donselaar (TU Delft - Applied Geology)
L. de Aguiar Paniago de Sousa (Katholieke Universiteit Leuven)
S. Kumar (TU Delft - Applied Geology)
S. Khanam (Environment and Population Research Centre (EPRC))
A.K. Ghosh (Mahavir Cancer Sansthan and Research Centre)
Cynthia Corroto (Universidad de Buenos Aires)
D. Ghosh (TU Delft - Sanitary Engineering)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Groundwater contamination with naturally-occurring arsenic (As) poses a serious health threat of global proportions. Implementation of focused and sustainable As-mitigation measures is hampered by the inability to pinpoint enigmatic As-hotspots in the vast area of continental alluvial basins with elevated toxicity risk. The catalyser role of invasive vegetation in anoxic fluvial oxbow lakes in combination with microbial metabolism processes are key in mobilizing As from its solid state into the shallow aquifer domain and accumulation in porous and permeable fluvial point-bar sediment. This insight opens up the opportunity for a cross-disciplinary approach to construct predictive machine learning-based object-based geospatial models to locate As-hotspots by the analyses of (1) satellite imagery of alluvial geomorphology, (2) oxbow-lake vegetation density, and (3) ground-truth databases of oxbow-lake/aquifer biogeochemistry and fluvial sedimentology.