Influence of trees on landscape temperature in semi-arid agro-ecosystems of East Africa

In the Dodoma region of Tanzania, overexploitation of natural resources has led to a severe land degradation and, consequently, to the need to restore local ecosystems. Within the possible strategies, the farmer managed natural regeneration agroforestry is strongly being promoted in the area. To validate the utilisation of this practise, this research aims (i) to assess the benefits of trees on microclimate studying the relationship between land surface temperature (LST) and tree canopy cover (TCC) in farmlands and (ii) to propose an operative tree density threshold to be set as objective by restoration activities in the area. Data collected with an unmanned aerial vehicle were integrated with Landsat 8 and Sentinel 2 satellite imagery. LST was retrieved through the practical single channel algorithm and TCC was calculated with a supervised classification. Pixel-based analyses at a resolution of 30 m and 10 m were performed, and the resulting data statistically analysed through the coefficient of determination, the Kruskal–Wallis test, and the Dunn test. Results showed that the TCC of the surveyed areas was 5.1%, and a significant decrease in LST of 1.32 °C (p < 0.01) was found in the areas with the highest TCC, during the late dry season. From these preliminary analyses, a threshold of 10% TCC is suggested for farmlands in this agroecological zone to bring beneficial microclimate changes in terms of temperature. Considering the average plant phenotype characteristic of the area, it corresponds to a tree density of 50 trees ha−1.