Sensitivity of hydrology and water quality to variation in land use and land cover data

Abstract

Land use and land cover (LULC) change is critical to hydrologic study as it affects surface runoff, sediment yield, and nutrient load from watersheds. Change in LULC is a continuous process in agricultural watersheds as per the growing and harvesting seasons. In this study, monthly LULC data layers were generated for the Big Sunflower River Watershed (BSRW), Mississippi by classifying cloud free Landsat images from 2014 to 2018 and combining them according to the season in order to obtain dominant LULC data layers for spring, summer, and fall. About 60 % of the total land area in summer was cultivated land, whereas only 20 and 5 % of the total land area were cultivated land during fall and spring, respectively. The rest of the total land area was constantly covered with urban, forest, and water for the three seasons. The overall accuracy and kappa coefficient ranging from of 87 % to 92 % and 0.8 to 0.90 was obtained during accuracy assessment of the seasonal LULC data layers. This suggested that the seasonal data layers can reasonably be used in the Soil and Water Assessment Tool (SWAT) to analyze the effects of LULC data variation on hydrology and water quality of BSRW. SWAT was calibrated and validated for streamflow, sediment yield, and nutrient concentration using the summer LULC data layer and those parameters were applied to the models using the spring and fall LULC data layers. SWAT output for runoff, sediment yield, and nutrient load was found to be very sensitive to the change in LULC data layers. The average amount of runoff and sediment yield was higher during the summer while total nitrogen and total phosphorous yields were higher during the fall and spring, respectively. Agricultural operations are usually conducted during the summer and fields are mostly barren during the spring and fall. The use of seasonal LULC data layers is very beneficial to the scientific community as they can better explain the seasonal variation in hydrology and water quality as compared to the annual cropland data layer that is available for a single season of a year.