Estimation of monthly mean minimum and maximum air temperature using MODIS LST data over Yazd province

Continuous monitoring of temperature in extreme areas, such as mountainous and desert regions, is hampered with insufficient or sparse distribution of meteorological stations, as well as the complexity of interpolating existing station data. For this reason, the use of satellite remote sensing data has increased significantly in recent years. In this study, the daily LST data (daytime and nighttime) obtained from the MODIS Aqua satellite in the period from 2003 to 2019 at 1-km resolution, as well as the meteorological data of 11 synoptic stations, were used to estimate the air temperature in Yazd province. The assessment of the relationship between the mean monthly LST (day/night) and the mean monthly air temperature (maximum/minimum) indicated a strong correlation between them. The coefficient of determination (R2) values between LST daytime and maximum air temperature varied from 0.989 to 0.997 and between LST nighttime and minimum air temperature from 0.991 to 0.999. Therefore, according to the suitable spatial coverage of MODIS LST, it is possible to estimate the air temperature for different cells of Yazd province using the linear regression equation. In addition, the results indicated that the RMSE values at night were much smaller than the values during the day. Therefore, it is possible to retrieve the land surface temperature at night with a much higher accuracy than during the day. The findings of this research also showed that the estimated air temperature and LST have similar seasonal cycles. Although the difference between these two variables is greater during the day than at night. These differences between the air temperature and LST increase in the summer season (June to August).