Effects of urban expansion on ecosystem health in Southwest China from a multi-perspective analysis

Abstract

Effectively exploring the impact of urban expansion on ecosystem health has become a hot topic for sustainable development of cities. However, analyzing the relationship between urban expansion and ecosystem health from a multi-perspective view is relatively rare. Here, taking 438 cities in Southwest China as the study area, we attempted to quantify the relationship between urban expansion and ecosystem health, taking into account population density, economic structure, urban area size, and geological environment. First, the urban ecosystem health was assessed based on the vigor-organization-resilience-services framework. Second, we quantified urban expansion from three aspects: the intensity of expansion, growth modes and urban forms. Finally, the panel data analysis was used to focus on which aspects of urban expansion effect on ecosystem health. The result showed that when the intensity of urban expansion increased and the growth mode changed from edge-expansion type to outlying type, the ecosystem health experienced a significant decline. In addition, there was significant negative correlation between urban size and ecosystem health within all types of cities, except for the mega-scale cities (>100 km2). A regular and aggregated urban form was benefit for ecosystem health at the medium-sized (100–500 person·km−2) and large-sized cities (500–1000 person·km−2). Moreover, urban form complex had a significant negative impact on ecosystem health in industry cities (secondary industry accounts for more than 50%) than in the service cities (tertiary industry accounts for more than 50%). The ecosystem health of karst cities was more sensitive to the fragmentation of urban core than non-karst cities. These findings will help to further understand the influence mechanism of urban expansion on ecosystem health under different scenarios and could provide a scientific basis for formulating reasonable urban planning.