Assessing Input Uncertainty and Sensitivity of the Process-Based Wetland Water Quality Model, WetQual
Type of DegreeMaster's Thesis
Forestry and Wildlife Science
MetadataShow full item record
Wetlands are among the most important natural ecosystems having numerous benefits to people and wildlife. One of the key functions of wetlands is water purification. Wetlands are known to improve water quality and act as natural water purifiers. They can trap sediments and other pollutants from waterbodies. However, their functioning can be impacted by various factors. Wetland hydrology and water quality models have been developed to better understand their functioning and represent wetlands in the landscape for improved terrestrial models. WetQual is a recently developed process-based model which simulates hydrology, sediment transport and nutrient and carbon cycles in wetlands. WetQual represents certain processes in detail and therefore require comprehensive input data. In other instances, it uses simpler processes and requires less data. The overarching goal of this study was to explore whether we can reduce the data needs in WetQual or if there is a need to improve certain processes, which may result in more data need. To achieve this goal, I studied (1) the significance of having detailed wetland bathymetry, (2) the efficacy of the water temperature calculation currently employed in WetQual, and (3) the consequences of using sub-daily level input forcings as opposed to the existing daily level inputs such as loading to the wetland, climatic data, etc., on nitrogen, phosphorous, and sediment simulations. A restored wetland in Maryland has been used to answer these questions. Wetland bathymetry is needed for flow routing. It also plays important role for nutrient and sediment transport. Wetland bathymetry data can be best deduced from topographic surveys, but in a landscape abundant of wetlands this poses a daunting task, thus simplifications are needed. Several geometric profiles were assumed, and volume-area-depth relationship were generated assuming that the only known information is maximum depth, surface area and volume. Model performances, parameter sensitivities and predictive uncertainties for each profile were compared to the results obtained with the actual bathymetry. Results from most geometric shapes were comparable to the results from the actual profile. Temperature is known to affect the fate and transport of chemicals and biological activities. The effect of temperature on biochemical reactions rates in WetQual is represented through the Arrhenius equation. Temperature is also key in evapotranspiration (ET) calculation. Water temperature in WetQual is calculated as a simple linear function of air temperature and the coefficients are fixed. As a second experiment, the adequacy of this relationship was explored to see if a more complex model is needed. Results showed that over the whole simulation period, the simple equation is adequate for most constituents, except for nitrate which showed the highest sensitivity to temperature. Further, results revealed that the simple equation can generate some misleading results when seasonality and interannual variability are concerned. Input forcings to WetQual are currently provided at daily time scale. However, since model runs at small time increments for numerical purposes, it internally disaggregates the daily data by simple linear interpolation. In a third experiment, the temporal resolution of temperature and evapotranspiration to the wetland were increased in a more realistic way, and runoff input to wetland was provided at hourly time scale to assess the temporal resolution impacts on model performance, parameter sensitivity and model uncertainty. Increasing the temporal resolution of temperature alone did not have a much impact. On the other hand, model was very sensitive to the temporal resolution of inflow data. Therefore, spending the extra time and money (if available) to obtain sub-daily inflow data is a worthy investment.