Toward a Near Real-Time Product of Air Temperature Maps from Satellite Data and In Situ Measurements in Arid Environments

Earth observation data and in situ measurements were used to derive a method for the estimation of Air Temperature(AirT) and thereby produce maps. The methodology was developed and validated using data acquired at five stations located in the United Arab Emirates(UAE) during the year 2012. In the first step, the effect of different variables on the estimation of AirT values was investigated and the variables of high importance were used to build the final model. Land surface temperature (LST), relative humidity, global horizontal irradiance, direct normal irradiance, and diffuse horizontal irradiance were identified as the most important inputs for Air Testimation. Models were developed separately for four different cases based on the seasons (Winter/Summer) and the time of the day (Day/Night). The models were evaluated using jackknife validation on the stations resulting in root mean square errors of 2.25°C(Winter/Day), 2.24°C (Winter/Night), 2.56°C (Summer/Day), and 2.69°C (Summer/Night) and an overall average accuracy of 2.44°C. Finally, the model has been applied on a larger scale by assimilating the relevant Earth observation data and in situ measurements for the creation of AirT maps for the entire country. The resulting AirT maps were generated in near real time at a temporal resolution of 15 min from METEOSAT/SEVIRI LST and the above-mentioned in situ measurements. These maps can be considered as important resources for several applications (e.g, climate studies, solar energy applications, thermal comfort studies, etc.).