Utilizing MODIS Products and Vegetation Indices to Derive Near Surface Air Temperature

Knowledge of spatial and temporal variation of air temperature is regarded essential in environmental studies. The objective of this study was to utilize the estimate of the kinetic temperature of the earth's surface skin (Land Surface Temperature, LST) along with Normalized Difference Vegetation Index (NDVI) in order to derive the air temperature in Columbia Valley, Washington. The estimated values were evaluated against the North American Land Data Assimilation System (NLDAS) dataset which is a Land Surface Model (LSM) composed of both observational data and model outputs. Furthermore, the relationship between RMSE, average bias, and associated land cover type for each individual pixel in the study area was analyzed. The results indicated that the estimated air temperature, using vegetation indices, is sensitive to the land cover type. The results indicated a highly significant relationship between the landcover type and the estimated air temperature values (P value= 0.0019). The lowest bias and RMSE values were associated with evergreen forest pixels. The highest coefficient of determination was obtained in pixels with a landcover of fallow cropland. The results of this study suggested the implication of this methodology for estimation of near surface daily air temperature particularly where there is limited access to recorded meteorological data. Hence, the impact of the landcover type on the accuracy of the predicted values should be considered to achieve a more reliable estimate for air temperature.