Ecosystem service value (ESV) is crucial for sustainable development. However, the spatiotemporal patterns of different types of ESV and their non-linear responses to influencing factors remained unclear at the municipal scale. By adopting an urban-rural comparative perspective and taking Wuhan as the case, this study integrated the information from remote sensing (RS) images and scenic points of interest (POI) into a revised method to investigate the spatiotemporal patterns of provisioning ESV (PESV), regulatory and supporting ESV (RSESV) and cultural ESV (CESV) during 2014–2022, and employed gradient boosting decision tree models to explore the key influencing factors and their non-linear effects. The results revealed heterogeneous spatiotemporal patterns of different ESVs. In urban areas, RSESV and CESV were low in the urban core. Meanwhile, RSESV decreased notably in suburban areas. In rural areas, PESV and RSESV demonstrated significant decreases near urban-rural fringes, whereas CESV remained stable in most analysis units. As for the influencing mechanisms, urban areas exhibited weakened anthropogenic effects and strengthened natural effects. Urban green infrastructure and forest were significantly associated with the increases of RSESV in 2022, especially when their coverages exceeded approximately 14%. In rural areas, the influence of land-use factors was dominant. Cropland and pond were the key influencing factors of PESV. Forest exerted substantial positive effects on both RSESV and CESV, especially when its coverage exceeded approximately 30%. The revealed heterogeneous urban-rural spatiotemporal patterns and non-linear influencing mechanisms of different ESVs are beneficial for municipal-scale urban-rural ecological planning. • An ESV estimation method was proposed by integrating RS and scenic POI information. • Each ESV showed heterogeneous spatiotemporal patterns in urban and rural areas. • In urban areas, anthropogenic negative effects weakened but natural effects strengthened. • In rural areas, the influence of land-use factors on ESV was dominant. • Different ESVs' non-linear responses to influencing factors were found in urban and rural areas.
Fan et al. (Mon,) studied this question.