EUVIMEDEUVIMED entdecken
Crossref

Machine Learning Forecasts of Groundwater Recharge Stress in a Semi-Arid Kovilpatti Region, Southern India

Abinaya R., Ravindran A. Antony, Promilton A. Antony Alosanai, Pitchaimani V. Stephen

Disaster Advances · 31.10.2026 · 83

DOI: 10.25303/1912da083094 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Disaster Advances, online erschienen am 31.10.2026. Verfasst von Abinaya R. und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Groundwater recharge in hard-rock semi-arid terrains depends more on how rainfall is distributed through the year than on its total amount, yet most drought and recharge indices rely primarily on rainfall totals. This study addresses that gap for Kovilpatti region, Tamil Nadu, a crystalline-aquifer, rain-shadow region utilising 125 years (1901-2025) of India Meteorological Department gridded daily rainfall data. A dynamic Recharge-Stress Index (dRRSI) was built from four indicators, SPI-12 deficit, maximum dry-spell length, precipitation concentration index and rainy-day deficit, combined through entropy weighting, which assigned 84% of total weight to rainfallstructure variables. Trend and change-point analysis identified a significant increasing rainfall trend and an abrupt regime shift in 1987 (+25.5% mean annual rainfall), corroborated by an independent rainfall typology. Eight forecasting models were benchmarked at 1, 3, 6 and 12-month lead times. The skill was strong at short horizons (Nash-Sutcliffe efficiency above 0.75 up to three months) but collapsed below the climatological baseline by 12 months. SHAP attribution showed forecasts shift from short-term memory toward rainfall-structure variables as horizon lengthens. Three uncertainty quantification methods all under-covered their nominal 90% target. These findings suggest that recharge stress in such systems is genuinely forecastable only over the short-to-medium term.

Zitieren

APA

R., A., Antony, R. A., Alosanai, P. A. A., & Stephen, P. V. (2026). Machine Learning Forecasts of Groundwater Recharge Stress in a Semi-Arid Kovilpatti Region, Southern India. Disaster Advances, 83. https://doi.org/10.25303/1912da083094

Vancouver

R. A, Antony RA, Alosanai PAA, Stephen PV. Machine Learning Forecasts of Groundwater Recharge Stress in a Semi-Arid Kovilpatti Region, Southern India. Disaster Advances. 2026:83. doi:10.25303/1912da083094

BibTeX
@article{R2026,
  author = {R., Abinaya and Antony, Ravindran A. and Alosanai, Promilton A. Antony and Stephen, Pitchaimani V.},
  title = {Machine Learning Forecasts of Groundwater Recharge Stress in a Semi-Arid Kovilpatti Region, Southern India},
  journal = {Disaster Advances},
  year = {2026},
  pages = {83},
  doi = {10.25303/1912da083094}
}
Kostenlos starten →