Performance evaluation of bias correction methods for CMIP6 precipitation datasets in the drylands of the Wolaita Zone, Southern Ethiopia

Purpose Reliable climate projections are required for effective climate change adaptation and management. However, raw general circulation model (GCM) outputs are usually beset by systematic bias that can be harmful to decision-making. This study aims to evaluate the performance of five bias correction methods (BCMs) in correcting precipitation data from six CMIP6 models over southern Ethiopia’s Wolaita Zone drylands. Design/methodology/approach The BCMs evaluated were distribution mapping (DM), multiplicative linear scaling (MLS), local intensity scaling (LIS), multiplicative delta change (MDC) and power transformation (PT). Their performance was evaluated using the Nash–Sutcliffe efficiency (NSE), mean absolute error (MAE) and coefficient of determination (R²). Findings The BCM performance varied across the models and metrics. The MDC was consistently the best, recording decreases in MAEs to 9.61–96.82 mm, R² to 0.99 and NSE of 0.71–0.99. Model-specific reductions in error ranged from 49.5% (ACCESS-CM2) to 89.3% (MPI-ESM1 – 2-HR), whereas the ensemble recorded 93.7% improvement. MLS and LIS improved mean rainfall and low-end extremes, respectively, but both failed to predict high rainfall quantiles. DM and PT exhibited fragile and unstable improvement. In general, the ensemble mean provided a more reliable improvement over the individual models. Research limitations/implications The study recommends using the MDC method for bias correction of precipitation data from six CMIP6 GCMs in the Wolaita Zone’s drylands. However, it is important to acknowledge that biases arising from imperfect modeling remain and cannot be fully eliminated by BCMs. Practical implications Using the suggested bias correction methods in the study area, it would be easy to protect future rainfall variability and change, as well as impacts on crop and livestock production in Ethiopia. Social implications The proactive adaptation measures suggested based on better accuracy data improve the farmers’ resilience to climate variability and change, especially the rainfall. This, in turn, maintains the stability of societies in the area by minimizing the level of migration. Originality/value This study provides the first comparative evaluation of several BCMs for CMIP6 precipitation data over Ethiopia’s drylands. The MDC and ensemble approaches were determined to be particularly robust for regional climate applications, with significant implications for climate adjustment, water resource management and policy planning.