John Fasullo (National Center for Atmospheric Research, United States)

The forced response due to greenhouse gases and aerosols is emerging in the 30-year satellite altimeter record of sea level change, but the differences with climate model large ensembles are also raising interesting questions about both the observations and the models. The broad patterns in the regional sea level trends are mainly being driven by the redistribution of heat in the oceans via wind-driven Ekman and Sverdrup dynamics. Climate model large ensembles (LEs) provide a useful framework for diagnosing the forced response in the observations, its emergence timescale, and a broad set of sea level sensitivities to climate forcings. While there is in general good agreement between the observations and the models, there are important differences. The forced response varies considerably across models and forcing uncertainty is as large an uncertainty as model structural uncertainty. Therefore, averaging those patterns across models leads to an offsetting of random model structural error and a significant underestimate of the likely magnitude of the forced pattern in nature. Also, the pattern in models fails to capture many aspects of the pattern estimated from altimetry, particularly in impact relevant regions such as the tropics. In addition, low resolution models, which are not eddy resolving, may underestimate the magnitude of spatial structures due to their coarse resolution. We will discuss what we have learned from the models, where the models might need improvement, and what this portends for projecting future sea level changes.