Basic Studies on the Estimation of Daily Mean Air Temperature by Using Surface Temperature Derived from Satellite Data-Relationship between Daily Mean Air. Temperature and Surface Temperature Measured on the Ground.

Supplemental analyses using Landsat data were performed to study the possibility of estimating daily mean air temperature using surface temperature derived from satellite IR data. It is recognized that daily mean air temperature can be estimated relatively accurately from surface temperature. Also, studies of the basic relation between daily mean air temperature and surface temperature measured on the ground (over forest and soybean fields) were performed. The results are as follows: (1)When there is comparatively small variation in vegetation density (e.g. forest), the correlation coefficients between daily mean air temperature and the temperature of surfaces receiving sunshine are large. (2)The correlation tends to be better in the morning and evening, and worse at noon. (3)When there is comparatively large variation in vegetation density (e.g. soybean fields), the correlation coefficient between daily mean air temperature and surface temperature for clear days is comparatively large, but the correlations are small for fine days. (4)RMSE over forest or soybean fields for both clear days and fine days are large in spite of a large correlation coefficient. (5)The correlation coefficients and RMSE can be improved by multi-regression analysis between two-hour surface temperature and daily mean air temperature. (6)The mean time at which surface temperature was closest to daily mean air temperature was 8-9 a.m. in the morning and 4-6 p.m. in the evening for May to October over a forest.