AbstractDecadal experiments' output of the coupled model inter‐comparison project phase‐5 (CMIP5) contains significant model drift. For practical use of CMIP5 decadal climate variables, it is necessary to correct this model drift. In previous studies, drift correction of CMIP5 decadal data of temperature and temperature‐based climate indices have been investigated, but there is no study that investigated the drift correction of decadal precipitation at a catchment scale. This study investigates the practical use of CMIP5 decadal precipitation data on seasonal scale using different drift correction alternatives for Brisbane catchment in Australia. For this, decadal monthly precipitation data from eight GCMs (EC‐EARTH, MRI‐CGCM3, MPI‐ESM‐LR, MPI‐ESM‐MR, MIROC4h, MIROC5 and CanCM4) were cut for the Australian region. By using the bi‐linear interpolation, the GCM data were re‐gridded to 0.05° × 0.05° spatial resolution matching with the observed gridded precipitation data collected from the Australian Bureau of Meteorology. Both model and observed data were subset for the Brisbane catchment and aggregated into seasonal means for Australian seasons. Four drift correction alternatives including one new modified method were employed for the selected GCM models and the models were categorized based on their performances using different skill scores. The results revealed that the performance of drift correction alternatives vary among different models and initialization years. Although some drift correction methods showed better performances than others for a given model but it is still difficult to suggest the most suitable method for seasonal precipitation. Drift correction approaches for other time scales such as monthly precipitation, and their application for individual ensemble members may be investigated further to better understand the implications of alternative drift corrections for decadal forecasting.