Estimation of land surface temperature of Kaduna metropolis, Nigeria using landsat images

Abstract

Understanding the spatial variation of Land Surface Temperature (LST), will be helpful in urban micro climate studies. This study estimates the land surface temperature of Kaduna metropolis, Nigeria. For the present study Landsat ETM+ images of 2001, 2006, 2009 and OLI 2015 was obtained from USGS of the study area. Normalized Difference Vegetation Index (NDVI) image was developed. The digital number of thermal infrared band is converted in to spectral radiance using the equation supplied by the Landsat user’s hand book. The effective at-sensor brightness temperature is obtained from the spectral radiance using Plank’s inverse function. The surface emissivity based on NDVI is used to retrieve the final LST. It was noted that 2006 has the highest maximum value with the highest mean value of 0.177 and standard deviation of 0.0903 while 2001 has the minimum value of NDVI. So also 2001 has the maximum value with highest mean value of 0.999 and standard deviation of 0.00161 while 2015 has the minimum value of surface emissivity. The coefficient determinant R² (0.837) show a strong positive correlation between mean of surface emissivity with date and season which shows downward trend in average over the study period. 2015 has the highest mean value of 39.42 with standard deviation of 1.92 of LST and coefficient determinant R² (0.46) show a positive correlation between mean of LST with date and season, with an upward trend in average LST over the study period. Lastly, NDVI is found to have negative correlation with LST. The Coefficient of determination (R²) (0.66) of surface temperature with NDVI and surface emissivity show a better prediction power of land surface temperature.

Keywords: LST, NDVI, Surface Emissivity, Landsat images, Temperature