报告详情
Detecting Groundwater Over-Exploitation Using Multi-Sensor Satellite Data and a Multi-Model Framework
编号：23
访问权限：仅限参会人
更新：2026-11-10 16:12:45 浏览：9次
Online
报告开始：2026年11月20日 17:15（Asia/Kolkata）
报告时间：15min
所在会场：[S3] Satellite UAV and Non-Terrestrial Networks/Networking, Security and Blockchain Systems [S3-1] Satellite UAV and Non-Terrestrial Networks/Networking, Security and Blockchain Systems
暂无文件
摘要
Groundwater over-exploitation across Karnataka hard-rock aquifers is officially assessed at the taluk level by the Central Ground Water Board, leaving sub-taluk depletion patterns unresolved. We investigate whether multi-sensor satellite observations can recover these classifications while guarding against spatial autocorrelation and data artifacts. Using Google Earth Engine, we compile 10-year pre-monsoon sequences from 2016 to 2025 across 13 spectral, radar, and climate channels. On a strict cohort of 50 taluks, comprising 44 over-exploited and 6 safe, uniform 30-patch sampling removed missingness shortcuts, and negative controls scored near or below chance, between 0.36 and 0.49. Under spatially buffered leave-one-talukout validation, the satellite-augmented model reached an area under the ROC curve (AUC) of 0.705 at 25 km, compared to 0.485 for coordinates plus topography. The paired AUC difference of 0.220 has a 95 percent confidence interval from minus 0.289 to 0.591, so this remains a pilot study. On a broader 54-taluk cohort, regularized logistic regression had the highest AUC (0.843) among classical machine learning, bio-inspired, and deep learning models, though confidence intervals overlap. A prototype map of model-estimated likelihood per sub-taluk block is illustrative only, pending validation against observation wells.
关键词
Groundwater over-exploitation, multi-sensor remote sensing, Google Earth Engine, spatial cross-validation, machine learning, hard-rock aquifers
报告人
Varshitha D
Student BGS College of Engineering and Technology稿件作者
全部评论
重要日期
-
会议日期
11月20日
2026
至11月22日
2026
-
10月31日 2026
初稿截稿日期
-
11月07日 2026
提前注册日期
-
11月22日 2026
注册截止日期
主办单位
United Societies of Science
承办单位
Pondicherry University
协办单位
IEEE Section
IEEE Madras Section
IEEE Madras Section
联系方式
- ICon INDIA 2026
- ic******@usssociety.org
发表评论