Date of Award

December 2025

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Environmental Engineering

First Advisor

Yeo Y. Lim

Abstract

Accurate rainfall estimation is essential for effective hydrologic modeling, particularly in urban environments where infrastructure must withstand extreme precipitation. This study evaluates the EPA Storm Water Management Model (EPA-SWMM) using MRMS RadarOnly Quantitative Precipitation Estimation (QPE) 1-hour Accumulation as the primary rainfall input for a watershed in Little Rock, Arkansas. The region was selected due to its urban character and the limited availability of radar-based hydrologic studies in Arkansas. 21 storms were analyzed from 2022-2025: 10 in Winter, 7 in Spring, and 4 in Summer. Storms were selected due to their effect on flow. The model was calibrated using February 15th, 2025 rainfall data due to its stratiform rainfall pattern and demonstration of typical hydrograph behavior and the rest were evaluated based on that calibration. Spring and winter storms yielded strong model performance, with average Nash-Sutcliffe Efficiency (NSE) values of 0.8075 and 0.8346, respectively. Summer storms showed greater variability (NSE = 0.6244), likely due to convective morphology and radar limitations. Percent Bias analysis revealed that winter storms tended to undervalue flow, while spring and summer storms overvalued. When outliers were omitted, winter and spring storms underrepresented peak flows consistent with literature on radar QPE underestimation however summer storms overestimated flow. Summer storms are too variable to be recommended for future stormwater analysis in Little Rock however winter storms can be recommended provided designers are aware of typical undervaluation amounts present in simulations. Future work should incorporate multi-sensor MRMS QPE products, refined subcatchment delineation, and field-based stream measurements to improve model accuracy.

Download

Share

COinS