Development of a Runoff Pollution Empirical Model and Pollution Machine Learning Models of the Paddy Field in the Taihu Lake Basin Based on the Paddy In Situ Observation Method

Agricultural non-point source (NPS) pollution has become a prominent problem for China’s water quality. Paddy fields pose high risk of to surrounding bodies. The paddy in situ observation method (PIOM) can calculate the runoff load without changing original characteristics and agricultural management measures. In this study, we carried out multisite field experiments during rice growing period Taihu Lake basin calculated loads. Then, developed empirical model (RPEM) machine learning models fields. Based on PIOM, average volume was 342.1 mm, loads mainly occurred early-stage seedling tillering stages. mean TN, NH4+-N, TP CODMn were 10.28, 3.35, 1.17 23.49 kg·ha−1, respectively. N P fertilizer loss rates 4.09 1.95%, RPEM included surface concentration fields, performance which validated based PIOM. irrigation volumes Zhoutie (ZT) Heqiao (HQ) analyzed by PIOM had an absolute difference 1.2~3.1%. With exception ZT, differences ZT HQ measured two methods less than 20%. This result illustrates accuracy feasibility analysis balance 114 datasets, RF provided best NH4+-N TP, SVM CODMn. training set R2 values above 0.8, testing 0.7. support calculation quantitative load. study provides theoretical basis technical quantification formulation control measures basin.