Please join us this Thursday at 3:30 in More Hall 230 to hear our very own Sanchit Minocha talk about how to improve estimates of reservoir storage across the globe using remote sensing and AI
For upcoming seminars, see here: https://depts.washington.edu/watersem/
RECLAIM: A Machine Learning Approach to Estimating Reservoir Storage Loss from Sedimentation
Speaker – Sanchit Minocha, PhD student, CEE
Abstract:
Reservoirs are critical infrastructure for freshwater management, providing essential services such as flood control, water supply, and recreation. While designed to store water, these structures also trap sediment—gradually reducing their storage capacity and undermining their intended functions. As a result, the operational lifespan of many reservoirs is often limited more by sediment accumulation than by structural degradation. Traditional methods to estimate storage loss, such as bathymetric surveys, are costly, labor-intensive, and impractical for the millions of reservoirs worldwide.
In this talk, we will introduce RECLAIM (Reservoir Estimation of Capacity Loss using AI-based Model), a machine learning framework developed to estimate sediment-induced storage loss in reservoirs. RECLAIM is trained on a newly compiled global dataset—GRILSS (Global Reservoir Inventory of Lost Storage by Sedimentation). It combines static features (reservoir and catchment attributes) with dynamic variables such as satellite-derived reservoir operations, inflow, and suspended sediment concentration (SSC) spanning four decades. By integrating AI with remote sensing and in-situ data, RECLAIM offers a scalable and cost-effective approach to assess reservoir storage loss globally. RECLAIM can serve as a screening tool to identify reservoirs most critically impacted by sedimentation, enabling targeted interventions and prioritization of costly sediment management strategies.
Bio: Sanchit is currently a PhD student working with Professor Faisal Hossain in the Civil and Environmental Engineering depratment of UW. He received his MS from CEE in 2023, and prior to that he was a gold medalist at the Indian Institute of Technology, Roorkee. He worked as a data hub geoscientist for Schlumberger, and trained people at Cairn Oil and Gas to use machine learning models. Sanchit was the People’s Choice Graduate Winner in the 2022 AWRA Annual Conference Poster Competition. He was recognized for his work on developing a Reservoir Assessment Tool (RAT) that can monitor reservoirs worldwide through the use of remote sensing and hydrologic modeling.