A wavelet coherence approach to prioritizing influencing factors of land surface temperature and associated research scales

Land surface temperature (LST), as an effective indicator measuring urban thermal environment, is significantly influenced by a range of human and natural factors at different scales. However, the scale-dependence of LST influencing factors has not been fully explored, due to relatively discrete scales or single factor used in previous studies. It is a great challenge to explore the approach to prioritizing research scales in view of the influencing factors of LST. Taking the urban group of Xi'an City and Xianyang City in Western China as a case study area, this study proposed a wavelet coherence approach to identifying the prioritizing LST influencing factors and research scales. Based on the sample transects, the results showed that around 1 km could be used as the prioritizing research scale for simultaneously exploring multiple LST influencing factors. And the normalized difference build-up index was the dominant influencing factor with the strongest multi-scale stability. The coherence relationships with LST of the area percentage of blue land, and the mean patch area of blue land represented high spatial heterogeneity, with multi-scale stability in the area of widespread water body. The normalized difference vegetation index should also be highlighted due to the multi-scale stability and stable medium coherence with LST. This study proposed a wavelet coherence approach to exploring spatial heterogeneity and scale-dependence of the relationship between LST and multiple influencing factors.