Date of Award:

8-2024

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Computer Science

Committee Chair(s)

Soukaina Filali Boubrahimi

Committee

Soukaina Filali Boubrahimi

Committee

Shah Muhammad Hamdi

Committee

Hamid Karimi

Abstract

Understanding and predicting streamflow along river basins is vital for planning future developments and ensuring safety, especially with climate change challenges. Our study focused on forecasting streamflow at Lees Ferry, a key location along the Colorado River in the Upper Colorado River Basin. We employed four machine learning models - Random Forest Regression, Long short-term memory, Gated Recurrent Unit, and Seasonal Auto-Regressive Integrated Moving Average; and combined historical streamflow data with meteorological factors such as snow water equivalent, temperature, and precipitation. Our analysis spanned 30 years of data from 1991 to 2020.

Our findings revealed that the Random Forest Regression model consistently outperformed others, particularly when integrating all meteorological factors with historical streamflow data. Using a 24-month historical data window, our model successfully predicted 12 months of streamflow with a Root Mean Square Error (RMSE) of 2.25 and an R-squared value of 0.80, demonstrating high accuracy. Furthermore, to assess the generalizability of our model, we tested it at other locations within the basin, including Greenwood Springs (Colorado River), Maybell (Yampa River), and Archuleta (San Juan).

By enhancing our understanding of streamflow dynamics and leveraging machine learning techniques, our research aims to provide valuable insights for water resource management and decision-making in the face of climate variability.

Checksum

d0aa79db973381360f753c385a65bff6

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS