Impacts of Rainfall Data on Model Refinement in Greater Pittsburgh
The demand for large-scale watershed and sewershed planning studies in the United States has been increasing steadily over the past ten years. In large part, the demand is driven by major government programs regulating combined sewer overflows, sanitary sewer overflows, and storm water discharges. The implementation of these regulatory programs often results in local or regional public agencies embarking upon large multi-year studies requiring a comprehensive inventory of watershed and sewershed infrastructure, a characterization of the hydrologic and hydraulic function of that infrastructure, and analyses of the water-polluting elements and processes. Significant monetary investments are being made into comprehensive field investigations and surveys, hydrologic and hydraulic models, and regional facilities planning to develop and implement short- and long-term combined sewer overflow (CSO) and sanitary sewer overflow (SSO) control strategies. The amount and rate of rainfall and snowfall are the key driving force in the quantity of extraneous flow that enters public sewers and private service laterals and the frequency and duration of wastewater discharges into receiving water bodies. However, it is not uncommon to observe that disproportionately small investments are being made to improve the precision and accuracy of regional rainfall measurement.
It has been demonstrated and documented in cities throughout the United States and around the world that calibrated radar-rainfall systems can provide accurate and precise rainfall measurement for large geographic areas. However, quantitative evaluations are relatively sparse that document whether enhanced rainfall measurement in fact translates to improved flow simulation performance from hydraulic/hydrologic models. The Saw Mill Run sewershed project at Pittsburgh, Pennsylvania provided a unique opportunity to characterize and quantify the impacts of improved precipitation estimates on the accuracy and reliability of model simulations for storm water runoff, wastewater flow, and CSO/SSO volumes. Four alternative rainfall data sources were input into the model and subsequently evaluated:
1) the regional long-term record gauge, located at the Pittsburgh International Airport, approximately 12 mi (19 km) away from the Saw Mill Run sewershed;
2) the nearest short- term gauge, located in the Saw Mill Run sewershed;
3) a dense network of 5 gauges, located in and around the study sewershed; and
4) a calibrated radar-rainfall system, with a pixel resolution of 1 km2.
This paper is only available in PDF Format:
View full text PDF