Some remarks on using the CMIP5 archive for sea-level projections

Tuesday, 16 December 2014
Luke P Jackson and Svetlana Jevrejeva, National Oceanography Center, Liverpool, L3, United Kingdom
The observed regional patterns of sea-level change differ widely from the global average, which is partly due to ocean dynamics. This spatial variability has ramifications for the mitigation of future sea-level change because some coastal regions will be affected more strongly by sea-level rise than others.

Quantitative projections of future sea-level rise rely on the outputs from climate models. The CMIP5 archive, which contains outputs from more than 60 climate models from 27 research groups, has been mined extensively for the purpose of sea-level projection based on four climate change scenarios. Previous studies show that regional patterns of sea-level change vary with each model. Not only that, but the individual model outputs vary widely in terms of actual resolution and model drift, and these factors have a secondary effect on the resulting projections.

We explore the influence of model resolution and drift correction methods upon sea-level projections and consider the projections from a limited set of high resolution model outputs compared to a limited set of low resolution model outputs. Furthermore, we consider the implications for the ensemble mean error for the different sea-level related parameters from the coastline to the sea, and show that projections near the coastline are highly sensitive to the method used for projection calculation.

This work contributes to the RISES-AM European Consortium Project, which seeks to provide a detailed holistic assessment of the physical and societal impacts of sea-level change due to high end climate change scenarios (>2°C warming relative to pre-industrial times).