MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202641093699 A) filed by Dr. Surendra Pratap Singh; and Dr. Swagatika Jena on August 03, 2026, for Ai-Integrated Multi-Source Geospatial System And Method For Climatic Trend Detection..
Inventors include Dr. Surendra Pratap Singh; and Dr. Swagatika Jena.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: Abstract Disclosed is AquaTerra-AI, a tri-modal, AI-integrated decision support system that unifies (1) satellitederived cryosphere and land-surface monitoring, (2) statistical climatic trend/shift detection, and (3) GISreferenced ground-truth land-degradation reporting into a single computational pipeline for basin-scale water management. A Satellite & Remote-Sensing Module acquires multi-decadal MODIS/FLDAS-class snow-cover fraction, land-surface temperature and evapotranspiration data at sub-degree grid resolution for defined river-basin zones. A Climatic Trend Engine ingests ground-station daily maximum/minimum temperature and rainfall series for regions served by the basin's canal network. A Ground-Truth & Farmer Feedback Module collects GPS-tagged field survey responses — waterlogging incidence, salinity extent, drainage condition, coping strategy adoption — through a lightweight mobile data-collection interface and geocodes them onto the same GIS reference frame as the satellite and station data. The three harmonised data streams are passed through an AI Trend & Shift-Detection Engine that applies Mann-Kendall and Modified Mann-Kendall testing, Sen's slope estimation, and Pettitt-Mann-Whitney change-point detection to each variable at annual, seasonal and monthly resolution, while an autocorrelationcorrection sub-routine selects the statistically appropriate test automatically. Outputs from all three streams are then fused, together with historical biopsy-style ground-truth labels of waterlogging/salinity severity, inside a Composite Risk-Scoring Module built on an ensemble deep-learning classifier trained on multi-year, multi-zone labelled data. The system outputs an AquaTerra Advisory Report within an operationally short turnaround window, comprising: a snow-melt-driven runoff forecast for the feeder basin, a temperature-anomaly trend classification for the command area, a spatial land- degradation risk map for individual canal minors/ distributaries, and a specific canal water-release and field-drainage-maintenance recommendation. The report is pushed simultaneously to a cloud dashboard for irrigation/canal authorities and Water User Associations and to a farmer-facing mobile/SMS advisory channel, while anonymised outcome data is uploaded back to the cloud for continuous, adaptive retraining of the risk-scoring model — improving predictive accuracy over successive irrigation seasons. By computationally linking upstream cryosphere change, regional climatic trend and downstream farmerreported land degradation — three domains that existing tools address only in isolation — AquaTerra-AI enables pre-emptive, basin-to-farm-gate water and land management decisions rather than the reactive, survey-after-the-fact approach that characterises current practice.
Disclaimer: Curated by HT Syndication.
