Evaluation of weather generator tools to estimate climate conditions in different agro ecological zones of North Shewa, Ethiopia

Abstract The information on climatic condition is difficult to obtain, expensive, and time-consuming so as to make timely decision on agricultural activities. As a scientific effort, this study was conducted to assess the temporal changes and trends of rainfall and temperature, to know the performance of weather generator (WG) tools in capturing the temporal and spatial distribution of rainfall, maximum temperature (Tmax) and minimum temperature (Tmin) and to evaluate the performance of WG in simulating the observed rainfall, Tmax and Tmin by using statistical methods. Mann–Kendall's trend analysis revealed that rainfall had non-significant (P < 0.05) decreasing trends, while Tmax and Tmin had an increasing significant (P < 0.05) trends at all stations. NASA POWER data followed by NewlocClim exactly capture the temporal changes of rainfall, Tmax, and Tmin in all stations except Debre Birhan and Mehal Meda. NewlocClim well captures rainfall at Alem ketema, while NASA well simulates rainfall at Debre Birhan and Majete stations. However Had-GEM2-ES, MRI-CGCM3, and CSIRO-Mk3.6.0 were not handling the spatial variability of observed rainfall at all stations. Similarly, some WGs showed moderate to good performance in capturing the spatial distributions of Tmax and Tmin. The smallest RMSE and CV, the highest R and d values were observed in NASA POWER and NewlocClim for rainfall, Tmax and Tmin. Therefore, NASA and NewlocClim are more accurate with goodness of fit to estimate rainfall, Tmax and Tmin in most of the stations to access weather data for ungagged stations for timely and reasonable decision-making on agriculture.