Font Size: a A A

Linking of hydrologic and water-quality models to evaluate reservoir water quality as a response to watershed management

Posted on:2006-02-14Degree:Ph.DType:Dissertation
University:University of ArkansasCandidate:White, Kati LFull Text:PDF
GTID:1451390008470839Subject:Engineering
Abstract/Summary:
Hydrologic/water quality models are available that simulate the three phases of nutrient dynamics in a watershed (landscape, instream, and reservoir), but they have rarely been implemented simultaneously. The overall goal of this research was to develop a watershed management tool by combining computer models to simulate landscape, instream, and reservoir nutrient processes, which was accomplished by completing the following objectives: (a) comparing results for a stream reach modeled with (1) the Soil and Water Assessment Tool (SWAT model) to (2) using a SWAT model in conjunction with the QUAL2E model to determine which modeling scheme (1 or 2) best predicted monthly constituent loading from a watershed into a reservoir; (b) linking model(s) that predicted flow and constituent loads exiting a watershed with the CE-QUAL-W2 reservoir model and evaluating alternative watershed management scenarios to determine their effects on reservoir water quality; and (c) identifying model parameters that contributed the largest prediction uncertainty using stochastic analyses. Results from two variations of the Pearson product-moment correlation indicated that correlation coefficients and regression slopes for the two modeling methods (SWAT and SWAT with QUAL2E) were not significantly different (alpha = 0.05). Therefore, the SWAT model (method 1) may be used to predict TP and NO3-N plus NO2-N yields just as well as SWAT loosely coupled to a QUAL2E model (method 2). SWAT model parameters (CN, ESCO, EVRCH, and SPEXP) identified during sensitivity analysis were input as distributions instead of discrete values. Management scenarios indicated that uncertainty in SWAT parameter inputs for CN, ESCO, EVRCH, and SPEXP resulted in W2 model predictions that were within similar ranges. Evaluation of output distributions for each of the four parameters suggested that CN and SPEXP contributed the greatest uncertainty in model predictions, with ESCO and EVRCH contributing less uncertainty in model output. The loosely linked SWAT and W2 modeling scheme provides a holistic approach to looking at nutrient dynamics in a watershed-reservoir system. The linked models can be used to assess the influence different watershed management schemes may have on reservoir chlorophyll- a concentration at targeted locations within the reservoir.
Keywords/Search Tags:Reservoir, Watershed, Model, SWAT, Quality
Related items