The cooling effect of hybrid land-use patterns and their marginal effects at the neighborhood scale

Abstract

Mitigating urban heat islands in neighborhoods is one of the most important issues in landscape planning. Identifying land-use patterns with cooling effect and applying them in practice is a way to reduce neighborhoods’ temperature. However, there are relatively few relevant theoretical and practical studies at present. A combination of factors influences the thermal environment of a neighborhood. With the increasing complexity of neighborhood land-uses, the cooling effect of these factors is difficult to measure individually and need to be considered from a holistic perspective. What kinds of hybrid land-use patterns (HLPs) in neighborhoods have the cooling effect and are prone to creating cold islands? This is a big question to environmental optimization and sustainable development, and there has little discussion about it. To address this, this paper proposed a technical framework for delineating and analyzing the cooling effect of HLPs at the neighborhood scale. Taking Jinan City as an example, four indicators were selected to describe the characteristics of HLPs, including building surface ratio (BSR), green-blue space surface ratio (GBSR), impervious surface ratio (ISR), and average building height (MAH). With the K-Medoids clustering algorithm, ten HLPs were delineated, and their marginal effects were analyzed by the Boosting Regression Trees (BRT) model. The study found that HLPs dominated by high-rise buildings and by green-blue spaces had the most pronounced cooling effect; low- and mid-rise dominated HLPs showed a weak cooling effect when the mean GBSR reached 34%, and the cooling effect of green-blue space and impervious surface dominated HLPs are more obvious when the mean GBSR reached 48%. The technical framework proposed in this paper provides a new perspective on thermal environment research and heat island mitigation at the neighborhood scale, which is applicable to cities with various climatic backgrounds and landscape characteristics and of generalization value.