Simulating the transport of dissolved ammonium-nitrogen and phosphate-phosphorous to streams using artificial neural network (ANN)

The export of agricultural fertilizers to streams results in untold harms to downstream water quality and thus ecosystem. Understanding and simulating the export of these chemicals is the first step to control their transport to streams. In this paper, the export of ammonium and phosphate, two major contaminants resulted from agricultural activities, were simulated using artificial neural network (ANN) and daily sediment and runoff data. Results indicated that ANN could simulate the trend of the export of these chemicals to streams successfully; however, it could not simulate the extreme events well. This might be due to the complexity of the transport of ammonium and phosphate, which contains many different processes that could not be captured in daily sediment and runoff data.