Derivation of air temperature of agricultural areas of Morocco from remotely land surface temperature based on the updated Köppen-Geiger climate classification

• Published in: Modeling earth systems and environment 2019-12, Vol.5 (4), p.1883-1892
• Main Authors: Hadria, R., Benabdelouahab, T., Elmansouri, L., Gadouali, F., Ouatiki, H., Lebrini, Y., Boudhar, A., Salhi, A., Lionboui, H.
• Format: Article
• Language: English
• Subjects: Air temperature; Chemistry and Earth Sciences; Classification; Climate; Components; Computer Science; Earth and Environmental Science; Earth Sciences; Earth System Sciences; Ecosystems; Environment; Geiger counters; Image classification; Land surface temperature; Math. Appl. in Environmental Science; Mathematical Applications in the Physical Sciences; Model accuracy; Original Article; Physics; Remote sensing; Spatial distribution; Stability; Stability analysis; Statistics for Engineering; Surface temperature; Temperature; Temperature effects; Temperature measurement; Temperature requirements; Weather stations
• ISSN: 2363-6203; 2363-6211
• DOI: 10.1007/s40808-019-00645-4

Air temperature is an important meteorological variable in many fields of our life. However, the availability of air temperature measurements over large geographic areas is often limited by the weather stations spatial distribution inadequacy, their low density and difficulties of data quality and collection. In this context, this study consists to develop four simple models to estimate the three components of air temperature ( T min , T max and T mean ) from remotely sensed land surface temperature ( T s ) derived from NOAA-AVHRR images, and based on the international Köppen-Geiger climate classification of Morocco. The results confirmed the existence of good relationships between the three components of measured air temperatures and land surface temperature derived from NOAA-AVHRR images for the main four climate classes of Morocco. The coefficient of determination, R 2 , varied between 0.69 and 0.80 for T min versus T s , between 0.62 and 0.74 for T max versus T s , and between 0.69 and 0.79 for T mean versus T s . The root mean square error varied between 3.1 °C and 3.3 °C for T min versus T s , between 3.2 and 4.1 °C for T max versus T s and between 2.7 and 3.4 °C for T mean versus T s . K-fold cross validation method was performed to assess the accuracy and the stability of proposed models. The limited number of proposed models is a great advantage to carry further studies requiring air temperature’s components at larger scale.