Validation of the uaqm-crcm5 model rainfall simulations over Northern Uganda

Abstract

Climate change has become one of the global challenges due to the different phenomenon like; extreme rains and droughts among others which have changed both in frequency and magnitude. To study climate change, climate models have been used to generate future climate information for climate change impacts and vulnerability assessments in many areas. Some climate models, however, fail to adequately capture observed rainfall pattern over East Africa particularly processes associated with topographical features especially on local scale for example in Northern Uganda. This study used Universitsé du Québec à Montréal-the Fifth Generation Canadian Regional Climate Model (UAQM-CRCM5) model over the northern region of Uganda. The overall objective of the study was to validate the UAQM-CRCM5 model rainfall simulations over study region. The data used in the study included the UAQM-CRCM5 model output and Climate Hazards Group Infrared Precipitation (CHIRPS) data for the period 1990 to 2008. The methodology employed included simple linear regression, correlation analysis and analysis of errors The results indicated that the model was unable to capture the pattern of the rainfall seasonality well over the region and it poorly reproduced the rainfall peaks as it depicts 2 rainfall peaks. However, rainfall anomaly was fairly reproduced by the model. The seasonal spatial distribution of rainfall was poorly reproduced by this model as it overestimated rainfall in the central part of Northern Uganda. It further exhibited wet biases for most parts of the region. On average, the correlation between observed and model output was weak ranging from 0 to 0.3. The model exhibited large errors with average mean error of 1.709 and Root mean square error of 3.106, an indication of the model’s inability to simulate the observed rainfall patterns. Overall, the model had a poor agreement with the observations.