Analyses of land surface temperature (LST) variability among local climate zones (LCZs) comparing Landsat-8 and ENVI-met model data

Abstract

The acceleration of urbanization has caused the urban heat island (UHI) phenomenon because it has changed the built and natural structure. In recent years, the common concept of a local climate zone (LCZ) has become the top priority for UHI studies for eliminating the lack of standardization. This paper presents the statistical relation between LCZ and land surface temperature (LST) for the hot and arid summer period in Adana City of Turkey. It combines vector-based analysis to map LCZ classes and raster-based research to obtain LST derived from a Landsat-8 and ENVI-met 4.3.0 microclimate model. Firstly, LCZs were classified using five parameters including building height, building surface fraction, aspect ratio, pervious and impervious surface fraction. Secondly, the mean LST for each LCZ was obtained from Landsat-8 and ENVI-met. Finally, the statistical significance of temperature differences of each LCZ was analysed by pairwise comparisons using the ANOVA Tukey-HSD test. The results show that built types had generally warmer LSTs than land cover types. LST difference was 0.5–1.0 °C in built types, while it was 1.0–2.0 °C in land cover types in Landsat 8. On the other hand, the differences of LST range from 1.5 to 3.0 °C for built types, and 3.0–8.0 °C for land cover types in ENVI-met. Subsequent post-hoc tests demonstrated that the mean LSTs obtained from Landsat-8 and ENVI-met exhibited statistically significant differences (p < 0.05). Moreover, the ENVI-met data had more obvious relationships with the LCZ-LST statistical results when compared with the Landsat-8 results. This study’s results will act as a guide for urban planners and future researchers to maintain a standardized UHI and ensure sustainable development, and have approved the current LCZ classification.