The nonlinear impact of urban blue-green space landscape pattern on residential housing price in a new first-tier city of China

Abstract Urban green and blue spaces (UGBS) play an increasingly vital role in enhancing urban environmental conditions and have been proven to increase residential property values. However, the nonlinear relationships between landscape patterns of UGBS and residential property values were less considered. To fill the gap, we employed nonlinear random forest (RF) models and partial dependence plots (PDPs) to analyze the impact of the spatial pattern of UGBS on housing prices in a new first-tier city (i.e., Nanjing) in China. The results showed that (1) The RF model explained 71.9% of the variation in housing prices. (2) Landscape patterns of UGBS significantly contributed to housing price variation, with the sum of relative importance scores of all the landscape pattern predictors amounting to 60.56%. (3) The PDPs results revealed that the percentage and spatial distribution of green and blue spaces positively influence housing prices. (4) Combining green and blue spaces boosted property values, with an average increase of approximately 0.41% and 0.53% in the percentage of landscape and mean patch area compared to green spaces alone. The findings of this study can inform urban planning policies and guidelines to optimize recreation and leisure spaces.