RESUMO
Wall surface temperatures are important components of urban climates but are under-sampled by satellite and airborne remote sensing and at the microscale are under-sampled in observational studies. In urban canopy models, they are represented with simplistic geometries. This study examines the effect of microscale (sub-facet) surface structure geometries on wall surface brightness temperature distributions at micro- to neighbourhood scales using mobile sampling traverses of two suburban neighbourhoods with different sub-facet geometries. Visible and thermal imagery were recorded simultaneously and combined and classified to create a database of temperatures with associated geographic and thermal attributes. This study investigates (1) if sub-facet scale geometries affect temperature distributions, (2) if these cause canyon scale biases, and (3) if there are therefore inter-neighbourhood biases. It is shown that sub-facet geometries modify wall surface temperatures predominantly by cooling due to self-shading. Surface-sun geometry thus leads to intra- and inter-neighbourhood temperature differences of several degrees Celsius. The observed effects have important implications for modelling of urban surface temperatures, where simplified geometries may overestimate wall surface temperatures.