The importance of spatial configuration of urban land cover in governing thermal response of urban form structure at design and masterplan levels.

Urban areas worldwide are affected by the urban heat island (UHI) effect whereby towns and cities are warmer than their rural backgrounds, having a negative impact on human health and well-being, energy use, and ecology. Appropriately distributed and spatially configured urban greenspaces can be used to mitigate the UHI, however, their efficacy so far has been investigated from either sparse air temperature measurements, micro-scale model simulations or coarse- resolution remotely sensed land surface temperature (LST), resulting in outcomes specific to particular urban fragments or averaged over areas relevant to masterplan and not urban design level. Additionally, the effect of the non- vegetated portion of land cover (LC) on the capacity of urban greenspaces to alleviate excess heat has largely been ignored. In this work, these gaps are addressed by using fine spatial resolution LST and LC data over the entire extents of three British towns to elucidate the relationship between LST and spatial configuration of urban form, taking into account both the spatial properties of greenspaces and their built-up neighbours. Spatial configuration of urban form was defined by aggregation of individual LC patches, size, elevation, and distance to LC patches of other types. Elucidation of the urban form-LST relationships required downscaling of available coarse resolution imagery with the use of high resolution ancillary data, and sub-division of main LC types into classes with distinct spatial aggregation and thermal properties. Random Forest regression allowed for determination of specific spatial configuration conditions leading to the formation of the hottest and coldest LC patches of a given type and highlighted the importance of neighbouring LC in their formation. Subsequently, the requirement for sophisticated spatial analyses for UHI-mitigating urban design was verified through assessment of the heat mitigation index generated by the InVEST 3.8.7 Urban Cooling model, which performed better at scales relevant to masterplans.