Model for prediction of non-point source pollution load based on self-memory theory

The self-memory theory was applied into a prediction model estimating annual non-point source(NPS) pollution load. The common self-memory model was modified to improve its accuracy. Data used for model establishing include non-point source pollution load of total nitrogen, flow, sediment at the Hua county hydrologic station on the Weihe River and precipitation of upstream basin from 1976 to 1999. The first 21 years’ data were used for training and the last 3 years’ data for testing. The results indicated that the modified self-memory method performed better, and the method could be used to predict NPS pollution load.