Evaluation of differences between gridded precipitation products in the Southern Prairies of Manitoba

This is the authors' abstract. We don't add key points for this paper.

Southern Manitoba, Canada Gridded precipitation products offer an alternative to sparse rainfall observations in the Prairies; however, the accuracy of the products significantly influences the accuracy of streamflow predictions. This study evaluates discrepancies among seven daily gridded precipitation datasets (2018–2023) against gauge observations. The error metrics chosen in this study are informed by the most impactful sources of error that affect hydrological modelling: systematic biases, non-stationary temporal variations, and event-related discrepancies. Seasonal analysis reveals distinct performance patterns across precipitation products. Winter estimates of precipitation exhibit the highest relative bias (RBias), with ERA5_L (199 %), GPM (161 %), and CaPA (137 %) substantially overestimating precipitation, despite having the lowest RMSE values (around 1 mm/day), reflecting the generally low observed precipitation amounts. Conversely, summer estimates of precipitation show lower RBias for MRMS (11.2 %) and ERA5_L (9.15 %), however, they exhibit the largest root mean square error (RMSE) values (up to 9.78 mm/day for GSMAP), indicating large absolute errors driven by intense convective storms. Fall and spring estimates present moderate RBias and RMSE values, with most products performing more consistently. Overall. MRMS and CaPA are the top-performing precipitation products, showing the lowest RMSE and highest correlation across seasons. NLDAS-2, ERA5-L, and PERSIANN have moderate accuracy, while GPM and GSMAP show higher errors and lower correlations. • Differences between seven gridded precipitation datasets are systematically evaluated. • Selected error metrics reflect the most impactful sources of precipitation error. • Seasonally dependent errors in precipitation estimates are dominant. • MRMS and CaPA outperform other products in accuracy and consistency.