The Seasonality of Surface Temperature Warming: A Robust Comparison of Models and Observations.

Climate models exhibit systematic apparent errors in the phase and amplitude of seasonal trends over time in near-surface air temperature and sea surface temperature. These apparent errors are quantified using temperature analyses from NOAA, NASA, and the Hadley Centre and Met Office and CMIP5 historical simulations. A discrepancy of 2 to 6 months in the timing of the maximum warming rate is found, with that timing changing over the historical record and depending on latitude and land vs. ocean. To determine the robustness of these differences, we use the MPI 100-member large ensemble to investigate sensitivity to natural variability, the set of 100 surface temperature realizations from HadCRUT4 to investigate sensitivity to data analysis uncertainty, and RCP8.5 simulations to investigate sensitivity to model structure. The model-observation differences are found to be larger than can be explained by natural variability or analysis error, implying the existence of common errors in modeled surface energy budgets or systematic errors in historical observations.