| dc.description.abstract | The annual cycle of surface temperature is altered over time because the annual mean surface temperature trend is not constant throughout the year, but instead exhibits distinct seasonality. Over the last century, the observed annual Global Mean Surface Temperature (GMST) trend peaks in March, however historical simulations by General Circulation Models (GCMs) peak a few months earlier. This model to observation mismatch has been studied several times in the last 20 years, but the model discrepancy is still present in the latest generation of GCMs. This study quantifies the observed seasonal trends at individual grid points, and in the form of zonally averaged latitude bands, using surface air temperature (SAT) data over land and sea surface temperature (SST) for ocean regions for all regions of the globe. Three ensembles of coupled GCMs are compared to observations: the MPI Grand Ensemble, the CMIP5 and CMIP6 multi-model ensembles. The use of large climate model ensembles enables the quantification of forced trends and effects due to natural variability in the seasonality of long-term surface temperature trends. Long-term seasonal temperature trends are calculated as the annual harmonic of surface temperature trends. The distributions of the simulation ensemble members are compared to the observations using the Mahalanobis distance statistic. The largest mismatch between models and observations stems from the GCMs’ gross underestimation of the forced seasonal warming trends that occurs over Northern Hemisphere (NH) mid to high latitude regions. Large seasonal warming is observed in Southern Hemisphere (SH) mid-latitude SSTs with peak warming in March, thereby reinforcing the observed seasonality in the NH and GMST trends. The observed large boreal spring peak warming trend in NH land regions suggests the snow albedo feedback could be the primary mechanism that is altering the seasonal cycle of surface temperature, according to a conceptual model of energy balance. Using the same conceptual model, the simulated seasonality of warming over NH land in GCMs suggests that the sea-ice albedo is the dominant forcing mechanism driving changes to the seasonal cycle of surface temperatures. | en |
