Simulation of Spatial Distribution of PM_2.5 Concentration in Shaoyang City Coupled With Multi-Source Environmental Factors

Based on an artificial neural network (ANN) spatial simulation model coupled with multi-source environmental factors, a kilometer-resolution simulation mapping of PM_2.5 concentration in Shaoyang City was achieved, and the spatial distribution characteristics of annual average values from 2016 to 2023 were obtained. The simulation results indicate that the ANN model can effectively capture the spatio-temporal variation characteristics of PM_2.5 concentration. The model validation demonstrated a fitting accuracy of 0.82 and a root mean square error not exceeding 6 μg/m³. This model can accurately reflect the spatial heterogeneity of pollutant concentrations, providing a basis for the scientific formulation of pollution control measures.