Machine Learning Applications for Evaporation Predictions From Small Reservoirs: Potential Water Savings for Lower Rio Grande Valley, Texas

Author

Syed Muhammad Fahad Abdullah, The University of Texas Rio Grande Valley

Date of Award

8-1-2025

Document Type

Thesis

Degree Name

Master of Science in Engineering (MSE)

Department

Civil Engineering

First Advisor

Chu-Lin Cheng

Second Advisor

Jude Benavides

Third Advisor

Rafael Almeida

Abstract

Local-scale reservoirs are important to regional water balance, but these are often overlooked. This study presents a robust machine learning (ML) approach leveraging reanalysis datasets to estimate daily evaporation for local-scale reservoirs in semi-arid South Texas. Selected models were trained with daily lake evaporation model (DLEM) estimates and used climatic and reservoir-specific properties as feature input variables. The multi-reservoirs training approach ensured applicable model generalization. Results show promising predictive performance with R² values ranging from 0.55–0.67 (testing) and 0.64–0.78 (validation), NSE values ranged from 0.54 0.67 (testing) and 0.64–0.78 (validation), and RMSE values ranged between 1.52–1.80 mm/day (testing) and 1.22–1.58 mm/day (validation). The findings highlight potential water savings of up to 2.1×105 ac-ft per year, which is equivalent to ~8% of the capacity of one major regional reservoir, if floating solar photovoltaic (PV) is deployed to cover 30% of its surface.

Comments

Copyright 2025 Syed Muhammad Fahad Abdullah. All Rights Reserved. https://proquest.com/docview/3275325910

Recommended Citation

Abdullah, S. M. (2025). Machine Learning Applications for Evaporation Predictions From Small Reservoirs: Potential Water Savings for Lower Rio Grande Valley, Texas [Master's thesis, The University of Texas Rio Grande Valley]. ScholarWorks @ UTRGV. https://scholarworks.utrgv.edu/etd/1787