Multi-year comparison of the effects of spatial pattern of urban green spaces on urban land surface temperature

Urban greening is increasingly viewed by cities worldwide as an effective measure to mitigate adverse impacts of urban heat island (UHI) effect. Despite the recognition of the possible influence of spatial pattern of urban green spaces (UGS) on their cooling effect by a small, yet growing body of literature, the results are inconsistent and inconclusive. This may be due to varying numbers and types of landscape metrics, and different statistical methods used in different studies, as well as scale-dependency of metrics and contextual differences between cities. We believe that comparative study across time and space using consistent methodology may overcome these challenges. We examined the relationships between spatial pattern and cooling effect of UGS in Singapore from 1973 to 2015 using a range of statistical and spatial methods. Landsat images were used to derive land cover and land surface temperature (LST) maps. Our results showed that: (1) four metrics of PLAND, SHAPE_AM, PD, and ENN_AM could effectively quantify UGS pattern and its relationships with LST across the years; (2) in addition to the amount of UGS (composition), four aspects of UGS configuration significantly influenced LST across the years namely, patch size, shape complexity, aggregation and fragmentation, and connectivity; (3) relative importance of composition versus configuration may depend on the existing UGS pattern; (4) relatively larger patches that are simpler in shape, more connected, and less fragmented were associated with lower LST. We discuss the significance of the results and insights they provide on optimizing UGS pattern for higher cooling effects.