Estimating Groundwater Levels at Multiple Scales in Data-Scarce Regions by Combining Data, Geostatistics, and Physics

In recent decades, the growing reliance on groundwater for irrigation and domestic purposes has underscored the need for precise monitoring of groundwater levels. Conventional methods for mapping groundwater levels, such as manual contouring and calibrated numerical models, face challenges in terms of computational time, complexity, scale, and cost. To overcome these limitations, we propose a novel framework that integrates multiple datasets using physics-informed space-time geostatistical methods to improve the precision of groundwater level estimation. Our approach will combine diverse data sources, including remote sensing, hydrogeological surveys, and historical groundwater measurements, to develop a comprehensive understanding of aquifer systems.

This research focuses on developing and adapting methodologies for physics-guided geostatistics, machine learning, and data fusion. The primary objective is to provide an effective solution to the challenges of groundwater management in regions with limited data, including across Pakistan and globally.