Runtime bias corrected driving data for regional climate models: regional-scale impacts

An empirical runtime bias correction (ERBC) technique is applied to a global model used to drive two RCMs: CRCM5 (Ouranos) and CanRCM5 (CCCma). While the ERBC is constructed to improve the climatological annual cycle, due to its application at runtime, it also indirectly improves time-dependent circulation dynamics and maintains physical coherence among adjusted atmospheric variables. We analyze the regional impacts of this prognostic bias correction on the climatology of atmospheric and oceanic driving data, for the historical period (1981–2010) and the end-of-century (2071–2100) climate change signal, in RCM simulations of North America. Three 10-member ensembles are created using three configurations of CanESM5 as driving data: the original CanESM5 (no correction); CanESM5 with only diagnostically corrected sea surface temperature (SST) and sea ice concentration (SIC); and CanESM5 with ERBC-corrected atmosphere and adjusted SST/SIC. Results show clear advantages of using adjusted atmospheric and oceanic driving data, especially in regions with substantial raw biases. Significant improvements are noted in representing key meteorological phenomena, notably nor’easters (extratropical cyclones) and the North American monsoon, due to better oceanic data and enhanced representation of regional circulation. In future projections, the ERBC is found to alter the climate change response of both RCMs, offering the potential for uncertainty reduction in future projection of such regional scale features.