Date of Graduation

12-2011

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Engineering (PhD)

Degree Level

Graduate

Department

Biological and Agricultural Engineering

Advisor

Dharmendra Saraswat

Committee Member

Sreekala G. Bajwa

Second Committee Member

Jackson Cothren

Third Committee Member

Thad Scott

Abstract

In this study, geospatial tools and techniques were developed to support specific aspects of watershed pollution management, such as quantifying land-use change (LUC) impacts, prioritizing subwatersheds, and communicating field-scale impacts, using the soil and water assessment tool 2009 (SWAT 2009) model.

For the land-use change objective, a geospatial tool titled SWAT2009_LUC was developed that enables SWAT modelers to prepare specific input files for simulating concurrent land-use changes during the SWAT 2009 model simulations. Testing of the tool for the Illinois River Drainage Area in Arkansas (IRDAA) watershed showed that the tool accurately represented temporal land-uses within the model. Model simulations with and without the activation of the LUC module showed that groundwater was under predicted by up to 15%, while surface runoff was over predicted by up to 13% at the subwatershed scale when a single land use layer was used. Overall, the results showed that activating LUC module using the SWAT2009_LUC tool exhibits hydrological simulations that are different from those resulting from a single land use layer.

For the subwatershed prioritization objective, a modeling approach was developed for prioritizing the 12-digit hydrologic unit code subwatersheds of the IRDAA watershed using the SWAT 2009 model output for sediment, total phosphorus (TP), and nitrate-nitrogen (NO3-N). The model was calibrated and validated at seven locations for total flow, base flow and surface runoff, and at three locations for water quality outputs. A multi-objective function consisting of percent relative error (RE), Nash-Sutcliffe efficiency (NSE), percent bias (PBIAS), coefficient of

determination (R2), and ratio of the root mean square error to the standard deviation of measured data (RSR) was used to guide model evaluations. The resulting priority subwatersheds comprised only 24% of the total area of the watershed but contributed 49% of sediment, 33% of TP, and 27% of NO3-N simulated loadings. Statistical relationships between priority subwatersheds and their various characteristics assisted with supporting the prioritization results. For the IRDAA watershed, this approach produced results that could assist watershed management agencies in optimizing allocation of limited resources in addressing water quality issues.

For the third objective, Field_SWAT, a simple graphical user interface (GUI) driven tool, was developed to map SWAT simulations from hydrological response units (HRUs) layer to a user-defined field boundaries layer. The SWAT model divides a watershed into HRUs based on unique land cover, soil type and slope. HRUs are a set of discontinuous land masses that are spatially located in the watershed but their responses are not tied to any particular field. The Field_SWAT tool ingests SWAT outputs and helps in visualizing them at field-scale using four different spatial aggregation methods. The tool was applied for mapping SWAT model's annual runoff and sediment outputs from 218 HRUs to 89 individual field

boundaries in an agriculturally dominated watershed in Northeast Arkansas. Area-weighted spatial aggregation method

resulted in most suitable mapping between HRU and field outputs. This research demonstrates that Field_SWAT could potentially be a useful tool for field-scale targeting of conservation practices and communicating model outputs to watershed managers and interested stakeholders.

Share

COinS