Estimating High Spatial Resolution Air Temperature for Regions with Limited in situ Data Using MODIS Products

Abstract

The use of land surface temperature and vertical temperature profile data from Moderate Resolution Imaging Spectroradiometer (MODIS), to estimate high spatial resolution daily and monthly maximum and minimum 2 m above ground level (AGL) air temperatures for regions with limited in situ data was investigated. A diurnal air temperature change model was proposed to consider the differences between the MODIS overpass times and the times of daily maximum and minimum temperatures, resulting in the improvements of the estimation in terms of error values, especially for minimum air temperature. Both land surface temperature and vertical temperature profile data produced relatively high coefficient of determination values and small Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) values for air temperature estimation. The correction of the estimates using two gridded datasets, National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis and Climate Research Unit (CRU), was performed and the errors were reduced, especially for maximum air temperature. The correction of daily and monthly air temperature estimates using the NCEP/NCAR reanalysis data, however, still produced relatively large error values compared to existing studies, while the correction of monthly air temperature estimates using the CRU data significantly reduced the errors; the MAE values for estimating monthly maximum air temperature range between 1.73 degrees C and 1.86 degrees C. Uncorrected land surface temperature generally performed better for estimating monthly minimum air temperature and the MAE values range from 1.18 degrees C to 1.89 degrees C. The suggested methodology on a monthly time scale may be applied in many data sparse areas to be used for regional environmental and agricultural studies that require high spatial resolution air temperature data.