275-7 Machine learning-based prediction of groundwater arsenic risk for sustainable paddy cultivation in Jorhat, Assam, India

Session: Advancing the Understanding and Management of Groundwater Pollution with Arsenic and Other Geogenic Contaminants Using Geospatial Tools, Machine Learning, and Data Science, Part II

Presenting Author:

Authors:

Abstract:

Arsenic (As) contamination in groundwater poses a serious environmental and public health threat in many parts of India, including the floodplains of the Brahmaputra in Assam. While lower Assam has been traditionally affected, recent evidence points to an alarming rise in arsenic levels in shallow aquifers and agricultural produce, particularly rice, in Upper Assam-especially around Jorhat district. Prolonged ingestion of As-contaminated water and consumption of arsenic-laden agricultural products such as rice may lead to serious health concerns like skin cancer, cardiovascular diseases, and developmental disorders. This study aims to investigate As concentration using Random Forest and XGBoost algorithms and environmental predictors (elevation, slope, NDVI, distance from river, and land use/land cover) to delineate As hazard zones. The predicted risk zones were overlaid with rice cultivation zones from NDVI and land use/land cover classifications, highlighting hotspots where As-prone groundwater intersects with paddy growing areas. This research contributes to building a conceptual contamination model, delineating high-risk zones, and recommending community-level mitigation strategies for sustainable use of groundwater and safe food production in arsenic-affected regions.

Geological Society of America Abstracts with Program. Vol. 57, No. 6, 2025

© Copyright 2025 The Geological Society of America (GSA), all rights reserved.