The estimation of air temperature from NOAA/AVHRR images and the study of NDVI-Ts impact: Case study: The application of split-window algorithms over (Perak Tengah & Manjong) area, Malaysia

The extraction of Land Surface Temperature Ts was the main objective in this study in order to estimate the ambient Air Temperature Ta over four locations in (Perak Tengah and Manjong) with variable surface cover types including: build-up areas, agricultural lands and forests. The Normalized Difference Vegetation Index NDVI was calculated from NOAA Advanced Very High Resolution Radiometer AVHRR images during June-August (2009) and surface emissivity ε was then calculated accordingly. From the brightness temperature in the images thermal bands 4&5 and using ε the surface temperature Ts was extracted using three different split-window algorithms. The application of the three formulas showed a homogenous results specially when using Becker et al (1990) and Uliveri et al. (1994) algorithms. A reasonable correlation was found between the satellite estimated Ts and air temperature Ta measured from ground weather stations over the study area especially when Uliveri et al (1994) algorithm was applied over the agricultural area with R² reached 82% as well as 87% over the residential and builds up areas while R² was 63% over area covered by forests. Moreover, a strong negative relationship was noticed among the NDVI - Ts representation. The relation was agreed with the so called “Universal Triangle” method which is being used in the study of soil moisture, vegetation cover and temperature interaction particularly over areas with biomass cover to indicate that fraction of vegetation cover has strong influence on the spatial value of Ts and its variability. In addition it was found that the NDVI-Ts relation is sensitive to surface moisture condition. Abnormality in NDVI-Ts was observed over urban and forest areas.