Effects of golf course construction and operation on nutrient runoff

Kansas State University has built an 18-hole championship golf course near Manhattan. Kansas, on Little Kitten Creek (LKC) watershed. The LKC watershed, previously native grassland, therefore, undertook dramatic changes in land-uses since the commencement of golf course construction in July 1998. A five-year research project starting from early 1998 enabled us to establish background surface water quality in terms of nutrients (total N and total P) and sediment concentrations, and evaluate changes of water quality in periods of golf course construction and operation. Water quality and quantity monitoring systems were set up. Water quality data were divided into three sets namely: pre-construction, during-construction, and operation. The impacts of construction and operation of the golf course on surface water quality were evaluated in terms of pollutant concentrations and pollutant transport rates. Construction of the golf course greatly deteriorated the surface water quality by increasing the nutrient concentration two to three times and sediment content about four times. The mass transport rates of nitrogen and phosphorus were increased three times and the rate of total suspended solids was about ten times. During the operation period, concentrations of total nitrogen and phosphorus were still about twice as high as native level even though the concentration of total suspended solids was about the same. There was no or limited impact of the golf course operation on surface water quality from the mass transport point of view. During the early operation period, the overall losses of fertilizer through surface runoff were less than 5%. However, losses from individual runoff event could be very high if fertilizer were applied shortly before a significant rainfall.; Efforts were made to simulate the sediment yields of the watershed before and during the golf course construction, using Agriculture Non-point Source Pollution model (AGNPS98) together with ArcView GIS. Comparison of measured data with the results predicted by AGNPS98 indicated that the model did a good job in simulating sediment yield, especially on a long-term base.; An artificial neural network model (ANN) was developed to predict nutrient concentrations in stream. Inputs of the model were concentration of total suspended solids, land-use index, and fertilizer application index. The model did a good job for all the three developing periods.