Abstract:
This dissertation presents an investigation on the use of satellite-based precipitation products (SPPs) in a hydrologic model to estimate water quality indicators in stream simulations. Three SPPs based on different retrieval algorithms are considered: the Tropical Rainfall Measuring Mission Multi-satellite Precipitation Analysis, TMPA 3B42-V7; the Climate Prediction Center’s CMORPH V1.0 product; and the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks Cloud Classification System, PERSIANN-CCS. The three SPPs are compared to rain gauge-based records over a 5-year period across the Occoquan Watershed, a 1500 square kilometer area, located in the suburban Washington D.C. area. The three SPPs are then used as input to the Hydrologic Simulation Program FORTRAN (HSPF) hydrology and water quality model. Each SPP-forced simulation is compared to the reference model simulation forced with the gauge-based observations, in terms of streamflow and several water quality indicators including stream temperature and concentrations of total suspended solids, dissolved oxygen, biochemical oxygen demand, orthophosphate phosphorus, total phosphorus, ammonium-nitrate, and nitrate-nitrogen. First, the skill of each SPP is evaluated on a continuous basis over the 5-year study period. Second, the propagation of errors from input SPPs to simulated streamflow and water quality indicators are evaluated. Third, the model is evaluated during eight extreme hydrometeorological events in terms of simulated streamflow and water quality indicators. Results indicate that the spatiotemporal variability of SPPs, along with their algorithms to estimate precipitation, have a quantifiable impact on both SPP-simulated streamflow and water quality indicators during both continuous and event-based modeling of extreme hydrometeorological events.
