Regional quality analysis of the hydrological environment with an improved random forest model based on the chimpanzee algorithm

Abstract

High-precision evaluations of water environment quality are highly important for improving the accuracy of early warning systems of regional water pollution risk and improving the regional water environment. This paper employs the chimp optimization algorithm (ChOA) to enhance the traditional random forest model, resulting in the chimp optimization algorithm-random forest (ChOA-RF) water quality assessment model for evaluating the Jiansanjiang area in Heilongjiang Province, China. The results show that the overall water environment in Jiansanjiang has the following characteristics: “The water quality of farms in the northwest is poor, and the quality of groundwater is better than that of surface water.” Total nitrogen (TN) and total phosphorus (TP) in surface water and ammonium nitrogen (NH₃-N), ferrum (Fe), and manganese (Mn) in groundwater are the main pollutants. The TP and TN in surface water and the NH₃-N in groundwater exceeded the relevant standards, likely due to the excessive application of chemical fertilizers, especially nitrogen fertilizers. Additionally, Fe and Mn are harmful native substances. According to these findings, targeted improvement strategies, such as reducing nitrogen fertilizer application, plugging well, and increasing the surface water utilization rate, are proposed. Moreover, the ChOA-RF model is compared with the traditional empirical value model and the particle swarm optimization-random forest (PSO-RF) model. The results show that the ChOA-RF model can effectively reduce the root mean square error and mean absolute percentage error and improve the coefficient of determination. The running time and convergence ability are also better than those of the PSO-RF model, which is a more accurate and efficient machine learning model. The model can be used not only for high-precision evaluation of regional water environment quality but also for other machine learning fields.