Application of a new integrated landscape index to predict potential urban heat islands

Studies on the interactions between urban landscape patterns and land surface temperatures are the key to urban heat island (UHI) mitigation. However, the landscape pattern indices used in previous studies were inconsistent in type and number. Furthermore, few studies combined the composition and configuration indices into one integrated indicator. The description of landscape pattern is thus a great yet rewarding challenge. In this study, we used the integrated location weighted landscape index (LWLI) proposed by Chen et al. (2009) and revised it to indicate potential UHIs. Fifty-six circular landscape samples along four transects were created and the LWLI was derived based on the theory of the GINI index. The LWLI considers the type, composition and configuration of different land covers inside a landscape, where configuration is treated as location weights. The potential UHI was represented by the land surface temperature (LST) of each circle landscape center. The correlation analysis results showed that the LWLI was significantly positively correlated with summer potential UHIs, with a Pearson R equaling 0.736, and barely correlated with winter potential UHI. Moreover, the LWLI was slightly more strongly correlated with potential UHI than composition alone. These correlations weakened as the landscapes grew larger. The linear regression results further revealed that the LWLI explained about 53% of the summer potential UHI, which was slightly better than composition alone (49%). This indicated that the LWLI was as effective in predicting the potential UHI as the combination of several pattern indicators, echoing Tobler's first law of geography, which states that "all things are related, but nearby things are more related than distant things." This study also leaves room for improvement of the index by integrating more environmental ecological parameters as weights, and for further application of the index in other fields. (C) 2016 Elsevier Ltd. All rights reserved.