Abstract: Based on the theoretical framework containing population agglomeration, greening level and environmental pollution, this paper constructs a spatial weight matrix of the combination of the inverse spherical distance and the area weighting and adopts the spatial Durbin model (SDM) and the spatial dynamic model of panel data (SDPD) to empirically examine the impacts of urban population agglomeration and greening level on environmental pollution. Then it employs the Geographically Temporally Weighted Regression (GTWR) to analyze the spatial heterogeneity of the environmental quality and its influencing factors. The findings show that: the population gathering would increase pollution emission, and its impact presents the form of the positive N type environmental Kuznets curve; the greening level can reduce the level of environmental pollution, the increased greening level per unit can reduce pollution by 0.204 unit which is caused by population space agglomeration through the interaction effect; their spillover effect on environmental pollution will be gradually decreased with the increase of geographical distance and strengthened with the increase of the urban area of the city; the cities with better environmental protection policies are more effective in avoiding the pollution caused by the agglomeration of population, the pollution treatment role of greening can be more fully played and the interaction between population spatial concentration and greening level can be promoted, while the cities without better environmental policies are just the opposite.

Key words: population agglomeration; greening level; environmental pollution; spatial heterogeneity