CALCULATING FLOWS USING REVERSE ROUTING AND ANALYZING DRIVERS OF HYDROLOGIC AND ECOSYTEM SERVICE FUNCTION IN STORMWATER DETENTION PONDS

Abstract

Stormwater detention ponds are an integral component of modern stormwater management systems. Designed to temporarily store stormwater runoff, these ponds regulate the release of water to mitigate downstream flooding and improve water quality by allowing sediments and pollutants to settle. Beyond these primary functions, detention ponds provide additional ecosystem services that are less frequently examined, including enhancing biodiversity, maintaining environmental flow regimes critical for downstream ecosystems, and contributing to aesthetic and recreational value as one of the major sources of nature in urban areas. Characteristics of the watersheds, engineering design elements, and the hydrologic regime in ponds all influence how these ecosystem services function and introduce synergies and tradeoffs among and between services both in pond and downstream. Reverse level pool routing is a method used to reconstruct an inflow hydrograph based on a known outflow hydrograph and the reservoir's stage-storage characteristics. Traditionally employed in larger river systems or storage reservoirs, this study adapts a second-order trapezoidal approximation of the reverse routing equation for use in stormwater detention pond systems. The approach relies on water level measurements, estimated stage-storage relationships, and standard equations for outflows through weirs and culverts. To evaluate its performance, synthetic and observed water level data were subjected to various levels of noise and bias, assessing the sensitivity of computed inflows and outflows to these errors. Smoothing techniques were then applied to the water level data to reduce error impacts. The results demonstrate that higher storm attenuation (i.e., the percentage reduction in peak inflow) generally corresponds to lower error propagation between noisy and bias-altered scenarios compared to a base scenario with no added error. This attenuation is influenced by outlet design and storm size, with more

restrictive outlets and medium-intensity storms achieving the highest attenuation rates. Overall, reverse routing in stormwater detention ponds proves to be a cost-effective and reliable approach for estimating inflows and outflows, particularly under conditions involving restrictive outlets and high-resolution pressure transducer measurements with minimal noise. Partial Least Squares Regression (PLSR) was then employed to predict and identify relationships, first between engineering design and watershed predictor variables and hydrologic response variables, and then between engineering design, watershed, and hydrologic predictor variables and ecosystem service indicator response variables. From the PLSR models, engineering design variables emerged as stronger predictors of hydrologic variables compared to watershed variables. Similarly, hydrologic and engineering design variables were found to have greater predictive power for ecosystem service indicators than watershed characteristics. Key engineering design variables frequently identified in the models included outlet restrictiveness, pond depth, pond volume/watershed area, and surrounding vegetation, and key hydrologic variables included bounce, percentage of days water level below outlet, and temperature variance. While the relationships between these predictors and ecosystem service indicators revealed both synergies and tradeoffs, engineering design features can be strategically modified during the design process to alter hydrologic regimes and enhance both in-pond and downstream ecosystem services of interest.