Date of Award


Document Type

Open Access Thesis


Earth and Ocean Sciences


College of Arts and Sciences

First Advisor

L. Allan James


Increases in percent impervious area and storm-sewer densities in an urbanized watershed lead to increased flood risk in urban areas. Conventional flood-risk management strategies such as detention ponds and low impact development (LID) can reduce peak flows. Research is needed to resolve questions about which strategy is best-suited for stormwater management under various schemes of sizing, distribution, and cost. Conventional and LID strategies differ in associated costs and benefits in addition to effectiveness and location feasibility. Previous research suggests that conventional strategies require less initial investment for design and construction, though LID is more cost effective in the long-term due to reduced annual maintenance requirements and the potential to distribution costs between centralized programs and public participation. This study used EPA’s Storm Water Management Model (SWMM) for rainfall-runoff simulations to test and compare the effectiveness of conventional and LID management scenarios in reducing runoff depths and peak flows of moderate-magnitude storms in the upper Rocky Branch Watershed (RBW) in Columbia, SC. The SWMM was calibrated and validated with six independent storm events using flow-stage data at a very small, highly urbanized watershed, and Acoustic Doppler Current Profiler (ADCP) discharge data at a larger watershed. Model calibrations and validations were assessed with a Nash Sutcliffe Efficiency (NSE) and each of the six storms yielded NSEs ≥ 0.712. Various configurations and locations of detention ponds and LID were modeled to compare the effectiveness of individual strategies under two levels of initial investment based on unit storage costs ($/m3). Individual application of both strategies was only effective placed upstream in the smaller, highly impervious subcatchment, in which case detention ponds were more effective in reducing peak discharges at both initial investment levels. A localized scenario in which bioretention was clustered in the upper, most-urbanized sub-basin provided a 2.1% greater reduction of peak flow at the primary watershed than a distributed scenario in which bioretention was spread across three different locations.