A spatially explicit approach for analysing the landscape pattern of urban vegetation using remotely sensed data and its impacts on urban surface temperature.

Abstract

The landscape pattern of urban green spaces and vegetation plays a significant role in supplying essential benefits and ecological services including sequestering and storing carbon, purification of air and water, regulating climate and providing recreational opportunities. However, due to the negative impacts of land cover change and rapid rates of urbanization, vegetation in an urban landscape typically becomes isolated and highly heterogeneous in space and time, relative to non-urban landscapes or natural areas. This research aimed to develop a spatially explicit approach based on remotely sensed data to quantify and monitor vegetation fragmentation and landscape structure of urban vegetation over time and its related impacts on the urban thermal environment using Harare metropolitan city in Zimbabwe as a case study. Specifically, multi-temporal Sentinel 2, Landsat 8 and Aster data were used in achieving the above objectives. Results based on the forest fragmentation model showed that the patch vegetation conditions, which represents the highest and severe vegetation fragmentation level, were dominant across the landscape, followed by edge, transition and perforated, whilst the core vegetation covered a small portion of the city. The decrease of large, connected and contiguous vegetation to a more scattered and fragmented vegetated patches was common across the city but more dominant in the heavily built-up areas of western, eastern and the southern parts of the city, indicating the significant impact of urban development. The small, isolated and scattered vegetation patches were associated with low positive and negative spatial autocorrelation of Local Indicators of Spatial Association (LISA) indices. On the other hand, the more homogeneous (clustered) vegetation was associated with high positive spatial autocorrelation in the northern part of Harare metropolitan city. Furthermore, the study showed that clustered, highly connected vegetation produces stronger cooling effects than dispersed, isolated and smaller patches of vegetation. Overall, spatial explicit approach and tools including the forest fragmentation model and LISA indices could play a significant role in landscape ecology with significant implications for conservation and restoration efforts based on the delineation of spatially explicit clusters of high or low vegetation cover, core or patch or edge vegetation conditions.