Abstract
This dissertation research has three primary objectives. Firstly, this work examines the use of a statistical flow methodology for characterizing ecologically important stream flows, the Indicators of Hydrologic Alteration (IHA), as a means to evaluate hydrologic model performance. Typically, IHA has been used to identify the extent of human impacts on a stream’s hydrology and to set management goals to restore the stream ecology. In this work, the use of the seven “extreme low flow” statistics of IHA is extended to the evaluation of the performance of a hydrologic simulation model under low flow conditions. Specifically, this work uses the IHA framework to evaluate the accuracy of the Chesapeake Bay Program Phase 5 (CBP5) watershed model during low flow events on a regional scale that is relevant to many water supply planners and managers. Because the CBP5 model's primary focus is predicting the Bay’s water quality, the measures used to calibrate the CBP5 model focused primarily on the calibration of the entire hydrological record and had only secondary emphasis on specific flow regimes, such as low flows and very low flows, although these flows are important for both stream ecologies and water supply planners. To provide a comparative performance benchmark, the performance of the simple Drainage Area Ratio (DAR) method relative to the IHA low flow statistics is also determined. This work demonstrates the use of IHA statistics for model evaluation in a case study, the Rivanna River watershed, a central Virginia subcatchment within the Chesapeake Bay drainage. For rivers with a large proportion of unregulated flow contributions, it is concluded that the computationally simple DAR model with appropriate surrogate watershed generally characterizes the extreme low flow conditions slightly more accurately than the CBP5 model. However, unlike the CBP5 model, the DAR model predicts future flows based solely on historical data, and thus the DAR model cannot predict flow impacts caused by hydrological alterations, thus limiting its use in water supply management. Nevertheless, this analysis suggests that incorporation of a low-flow-specific metric into the CBP5 calibration could improve its utility for water supply management and planning at a regional scale.
Secondly, this work develops and demonstrates a methodology to specifically assess the inter-relationships between estimated precipitation, observed stream flow, and hydrologic model performance. To satisfy this objective, this work introduces a new concept called ‘precipitation fidelity,’ which is the correspondence of stream outflow to the estimated precipitation used as input into a hydrologic model. Simple annual and daily precipitation fidelity indices are defined. The use of the precipitation fidelity indices is then demonstrated for the Rivanna Watershed as modeled using the CBP5 model and the associated precipitation input data set. The precipitation fidelity results are used in conjunction with model output to identify the effect of precipitation estimation accuracy on model performance at both long time scale and short time scales. Based on the daily precipitation fidelity measure, in the headwater watersheds, about a quarter of the days lack fidelity between the precipitation input and the observed stream flows. Days when the estimated input precipitation has runoff-generating rainfall, but the observed stream discharge does not increase, have the highest average relative daily modeling errors and high area-weighted daily modeling errors. These results indicate that precipitation needs to be better represented in the headwater subwatersheds. Regression analysis using the Analysis of Covariance method was used to determine statistical similarity between annual estimated precipitation and observed and modeled stream flows. Regression results suggested that direct hydrology calibration of the subwatershed of interest leads to both a higher level of correspondence between estimated precipitation and modeled flows and an acceptable ‘goodness of fit’ between the modeled and observed data.
Lastly, this work employs a novel simulation-optimization modeling approach to modify the design of detention ponds to preserve the natural ecological flows, while satisfying the requisite regulatory flow requirements. This work utilizes an innovative ecological flow paradigm: the eco-flow statistics. The eco-flow statistics consist of nine hydrological flow statistics that have been shown to be particularly relevant to ecological quality. The statistics include annual and seasonal ecodeficits and ecosurplus, calculated using median annual and seasonal functional duration curves, and the total seasonal ecochange. A new metric called the ‘ecodifference’ is defined as the weighted sum of the nine eco-flow statistics and represents the hydrologic alteration in the stream. The ecodifference in a receiving stream can be calculated using the outflow hydrograph from a detention pond hydrologic simulator. First, a design approach- using a hydrologic model, detention pond model, and the ecodifference metric- is used to design a series of flow controls in a detention pond outlet control structure that reduces the ecological impact to the stream caused by development, while meeting current design regulations. Then, a simulation-optimization strategy that incorporates a genetic algorithm with the design approach is introduced to design an outlet control structure that best minimizes the ecological impact to the stream. For a case study site, optimized designs have demonstrated that improvements in ecological flows can be achieved while meeting design regulations. By introducing this approach for eco-detention ponds, and then demonstrating its performance, this work has potential to impact stormwater management design practice.
