Higher land surface temperature (LST) in cities than its surrounding areas presents a major sustainability challenge for cities. Adaptation and mitigation of the increased LST require in-depth understanding of the impacts of landscape features on LST. We studied the influences of different landscape features on LST in five large cities across China to investigate how the features of a specific urban landscape (endogenous features), and neighboring environments (exogenous features) impact its LST across a continuum of spatial scales. Surprisingly, results show that the influence of endogenous landscape features (Eendo) on LST can be described consistently across all cities as a nonlinear function of grain size (gs) and neighbor size (ns) (Eendo = βnsgs-0.5, where β is a city-specific constant) while the influence of exogenous features (Eexo) depends only on neighbor size (ns) (Eexo = γ-εns0.5, where γ and ε are city-specific constants). In addition, a simple relationship describing the relative strength of endogenous and exogenous impacts of landscape features on LST was found (Eendo > Eexo if ns > kgs2/5, where k is a city-specific parameter; otherwise, Eendo < Eexo). Overall, vegetation alleviates 40%-60% of the warming effect of built-up while surface wetness intensifies or reduces it depending on climate conditions. This study reveals a set of unifying quantitative relationships that effectively describes landscape impacts on LST across cities, grain and neighbor sizes, which can be instrumental towards the design of sustainable cities to deal with increasing temperature.
