CREATING AIR TEMPERATURE MODELS FOR HIGH- ELEVATION DESERT AREAS USING MACHINE LEARNING

The standard way to measure the air temperature (Ta) as key variable in climate change studies is at 2m height above surface a fixed location (weather station). In contrast, (Ts) can be measured by satellites over large areas. Estimation of Ta from Ts one potential overcoming shortages due uneven or irregular distributions weather stations. However, whether this successful has not been assessed high-elevation regions. This particularly important study, we estimate desert zone Kilimanjaro (>4500m) using four models (five including benchmark model) with unique sets inputs five machine learning (ML) algorithms. Note that different combinations and were tested evaluate proxy for Ta. Root Mean Square Error (RMSE) each model was compared ranked according their RMSE. Similarly, algorithms terms reliability consistency. Correspondingly, results model. Three out outperformed consistency ranking, while two ranking. Therefore, ML are efficient tools estimating environment. only accurate used an earlier time period inputs. highlights amount de-coupling between TS high elevations, which provides challenge alone zone.