Relationships of LST to NDBI and NDVI in Wuhan City based on Landsat ETM+ image

Based on four Landsat ETM+ images covering Wuhan City, Normalized Difference Vegetation Index (NDVI) and Normalized Difference Built-up Index (NDBI) in spring, summer, autumn and winter were calculated. The land surface temperature (LST) of the city in four seasons was estimated using the radiative transfer model. The spatial and temporal pattern and characteristics of NDVI, NDBI and LST and their relationship were analyzed in detail. Results show that the regions with higher NDBI and LST values were mainly in the municipal district of Wuhan City, including Hankou and Wuchang. The maximum NDBI of the district in four seasons was 0.669, 0.685, 0.659 and 0.703 respectively, while the maximum LST was 43.529°C, 59.849°C, 47.143°C and 23.875°C. Higher NDVI values were mainly distributed in the regions such as along the Han River, the lake beach and the wetlands area in the southwest and central hilly areas, and the maximum NDVI in different seasons was 0.432, 0.564, 0.442 and 0.158 respectively. There existed a strong positive correlation between NDBI and LST in all seasons. The Pearson's correlation coefficient in four seasons was 0.639, 0.717, 0.807 and 0.762 respectively. In autumn, NDBI and LST were most significantly correlated. The relationship between NDVI and LST changed with season, but without obvious regularity. The correlation between NDVI and LST was weaker in four seasons than that between NDBI and LST, with the Pearson coefficients of -0.515, -0.577, -0.246 and -0.038 respectively. NDVI values had negatively linkage with NDBI. The Pearson's correlation coefficients in four seasons was -0.568, -0.582, -0.536 and -0.242, which indicated that NDVI could be used to characterize the evolution of urban construction land. The throughout study suggests that NDBI not only can be used as an important indicator to analyze LST and urban heat island effects, but also will provide a reliable basis for urban construction and planning.