Haze Risk Assessment and Trend Evolution in Fenwei Plain Based on PCA-XGBoost

According to the regional disaster system theory, haze risk assessment indicators for the Fenwei Plain were chosen to effectively study the urban haze hazards in terms of the sensitivity of the disaster-causing environment, the hazard of haze components, and the vulnerability of disaster-bearing bodies. The risk level of haze hazards in Fenwei Plain cities was calculated using Principal Component Analysis (PCA). A quantitative assessment of the haze risk in the Fenwei Plain urban agglomeration was carried out using the extreme gradient boosting method (XGBoost) to investigate the spatial and temporal distribution patterns of PM2.5 and its trends. The results show that Xi'an exhibited the highest haze risk index and the PCA-XGBoost model performed best in the three evaluation indicators. The main factors affecting haze are the economy and population density.