Revisiting the reanalysis-model discrepancy in Southern Hemisphere winter storm track trends

Reanalysis datasets show wintertime storminess in the Southern Hemisphere (SH) has significantly increased since 1979.Previous work reported a reanalysis-model discrepancy whereby coupled and prescribed sea surface temperature (SST) models in CMIP6 were unable to reproduce the trend.Here we revisit the reanalysis-model trend discrepancy in SH winter storminess focusing on the impact of observational uncertainty, model ensemble size, a like-for-like comparison, and mechanisms underlying the discrepancy.A large spread is found across available reanalyses indicating observational uncertainty.When the storminess trends in reanalysis and model datasets are quantified on the same time and spatial grids, the reanalysis trends decrease, and a discrepancy between reanalyses and prescribed SST models is unlikely.However, a discrepancy between reanalyses and coupled models is still likely, particularly in the South Pacific.We test the importance of SST trend discrepancies in coupled models using Southern Ocean and tropical Pacific pacemaker simulations.Under Southern Ocean pacemaking, a zonal-mean trend discrepancy between reanalyses and coupled models is unlikely and the improvement is due to the coupled models capturing surface energy flux trends.However, a discrepancy is still likely in the South Pacific.Under tropical Pacific pacemaking, a trend discrepancy between reanalyses and coupled models in the South Pacific is unlikely due to the coupled models capturing the La Nina-like teleconnection trend.Our results show that reanalysis-model trend comparisons should involve all reanalysis and model datasets and like-for-like calculations.Furthermore, regional SST trend discrepancies can lead to non-local reanalysis-model circulation trend discrepancies.