HIGHLIGHTING water balance were studied. The model was setup

HIGHLIGHTING THE IMPACTS OF BRUSH MANAGEMENT ON WATER BALANCE FOR TEXAS ROLLING PLAINS USING SWATAbhinav Kandpal1, Clyde Munster1, Srinivasulu Ale21 Department of Biological and Agricultural Engineering, Texas A&M University, College Station, TX, USA, 2 Texas A&M AgriLife Research, Texas A&M University System, Vernon, TX, USAABSTRACT.The stream dynamics can be predicted very efficiently using hydrologic models. Texas, a state so diverse in topography has an interesting hydrology. In the current study, we have tried to access the hydrologic cycle of the Texas Rolling Plains (TRP) with the help of the Soil Water Assessment Tool. In the TRP region range brush is the predominant land use type covering more than 60% of the area. For the purpose of study the area under range brush was changed to pasture in the land use tab of SWAT and its impact on water balance were studied. The model was setup for the period of 1996-2016 with 1996 -2007 being the calibration and the rest being the validation period. The model was manually calibrated for streamflow using some of the sensitive parameters that were obtained from the literature review of the study area. To check the validity of the model three statistical parameters namely, Nash-Sutcliffe Efficiency (NSE), Percentage Bias (PBIAS) and RSR were used. Due to unavailability of weather data for year 2008 it was removed from the calculations as it was yielding erroneous values. For the calibration period NSE of 0.822, PBIAS of -6.61% and RSR of 0.421 was observed. For the validation period NSE of 0.79, PBIAS of 0.02% and RSE of 0.319 were observed which show that the model performed very well for the region. The calibrated model was used for simulating brush management. The model was re-run with the new land use and changes in water yields were studied for the sub-basins comprising of pasture. The results showed a slight increase of 2.5% in the average yearly water yield and considerable increase of 82.4% in groundwater flow. Furthermore, a slight decrease of 7.4% in surface runoff and 1.4% in evapotranspiration was observed.Keywords.Soil Water Assessment Tool, land use, calibration, range brush, pastureINTRODUCTIONBrazos river basin is one of the major river basin of Texas. It starts from the confluence of the Double Mountain Fork and the Clear Fork in the Rolling Plains of Texas. It has a mean annual discharge of 237.5 CMS (Wikipedia Contributions “Brazos River”) which serves the needs of many counties of Texas. Through the entire Brazos basin a variety of crops are grown, with cottonand winter wheat being the predominant one in the Rolling Plains. Around 35% of the area is irrigated using the water from various sources and the rest if rain-fed. Apart from the Brazos River a significant portion of water is extracted by the Seymour aquifer underlying the region. With the undergoing changes in climate there is need to conserve water so that the needs of the future could be addressed. In Texas, a new brush management conservation practice has been initiated by the USDA and NRCS which involves selective removal of woody trees to increase water yields in the downstream regions (Harwell et al., 2016). In addition, government organizations are funding new projects to look through the impacts of changing land use to streamflowSince the development of SWAT there have been many studies highlighting the use of SWAT to predict the effect of change on land use on streamflow. Yet apart from that many other computer and mathematical models are used to predict the same effects. Bui (2011) used HEC-HMS to access the impacts on streamflow based on GCMs for the coming decades. In addition, (Hurkmans et al., 2009) used the VIC model to process the hydrological changes by using different land use data for different years. U.S. Geological Survey’s Precipitation Runoff Modeling System was used by (S. Qui et al., 2009) to demonstrate the potential change in water yields. They used two GCMs and different land use scenarios for Trent River basin in North Carolina which predicted a 14% -20% increase in water yield for the region.Kundu(2002) used SWAT with a combination of Markov Chain Model to evaluate the impacts of changing land use on Evapotranspiration. They also tried to highlight changes in curve numbers of soil for upcoming decades. USGS conducted an independent study for the Double Mountain Fork Watershed which simulated the effects of brush management on water yields (Harwell et al., 2016). Which concluded that changing land use from shrub land to grassland increased the water yield by 1700 gallons per acre,MATERIALS AND METHODS Description of Study AreaThe Seymour Watershed is a part of the Brazos river basin which runs through the entire Texas cutting the state into two halves. The total drainage area of the watershed is 1567.89 sq. miles. The major soil types in the area are interbedded sand, gravel and silt. The average yearly precipitation for the region is 23.4 in. with peak being between May to September as observedfrom the data of 1996- 2016. The annual temperature of the region ranges between 53.8°F – 77.6 °F.SWAT Model DescriptionSoil and Water Assessment Tool (SWAT) model is a semi–distributed physical-based basin–scale, continuous hydrologic simulation model developed by the United States Department of Agriculture (Arnold et al., 1998). The model is based on the regional water budget equation. Soil Water Assessment Tool (SWAT) model works as an agglomeration of crop growth model (EPIC), daily rainfall and hydrology (CREAMS), pesticide component (GLEAMS). These models are run together in a Simulator for Water Resources in Rural Basins (SWRRB) model (Arnold etFig. 1: Description of Study Areaal., 2012). In SWAT the watershed is divided into sub-basins which are further sub-divided regions of homogenous land use, soil and slope that are known as Hydrologic Response Unit (HRUs). Using all these models SWAT helps us evaluate the streamflow taking hydrologic data as inputs.Input DatasetSWAT model is a combination of many hydrologic and water quality models that uses a diverse range of input data. A 30m X 30m Digital Elevation Model was downloaded from the Earth Explorer of the USGS for the elevation input to the model. The DEM was projected to NAD 1983 Lambert Conformal Conic projection so that it would overlay completely with other data. The area is has a total of 5 weather station around it. For the purpose of study daily rainfall data for the stations in Seymour, Aspermont and Guthrie were extracted from the NCDC NOAA website for the period of 1996 -2016. The missing data for these stations filled up by averaging out the values from adjacent stations. All the weather stations were lacking data for the year 2008 thus it was removed from the period of study. Land use raster layer for the year 2014 was downloaded from the National Agricultural Statistics Service(https://nassgeodata.gmu.edu/CropScape/) which underlined the fact that range brush was the predominant soil type in the region.For the purpose of proper calibration of the model streamflow data for two USGS gages 08082000 and 08080500 was downloaded for the period of study and entered into the model as point sources. These gages are the most upstream points in the watershed which are a part of Double Mountain Fork and Clear Fork of the Brazos respectively. In addition, the outlet of the watershed was selected as the other USGS gage at Seymour 08082500. Furthermore, for Soil data input to the model, a raster layer of Soil Survey Geographic Database (SSURGO) was downloaded from the NRCS website. This data is interpolated by model using a lookup table understood by SWAT. As