Characterizing uncertainty in CMIP5 projections of local sea level change

Climate-driven changes in sea level are highly temporally and spatially variable; most observed variability originates in the ocean's thermodynamic properties, mass distribution, and dynamics. Coastal planning efforts thus require localized projections of changes in dynamic sea level (DSL; sea surface height relative to the geoid). Ocean climate models included in the Coupled Model Intercomparison Project Phase 5 (CMIP5) provide projections of DSL and global thermal expansion (GTE) at approximately 1 degree horizontal resolution. At smaller scales, uncertainty in multimodel assessments of 21st century projections varies by location and over time. Characterizing this uncertainty at a local level is critical to understanding the unique coastal flooding risks faced by different locations, as well as the importance of GTE and DSL relative to other components of sea level change (e.g. glacier and ice sheet mass balance). The New York City Panel on Climate Change recently used CMIP5 projections of GTE and DSL (encompassing two representative concentration pathways, or RCPs, and 24 models) for inclusion in probabilistic projections of local sea level change at New York City. By the 2050's, GTE and DSL changes are responsible for approximately 30-50% of the total sea level rise projection; this combined contribution exhibits a wide spread (10-90th percentile range 7-35 cm) that is dominated by widely varying projections of DSL change across models. Here, we expand this analysis to other locations, using all available CMIP5 models and additional RCPs through 2100. The uncertainty in global mean and local sea level is partitioned into that arising from different RCPs, models, and internal variability. This uncertainty analysis yields insight into the influence of key choices in the assessment process: in particular, model and scenario exclusion and/or weighting. We then present fully probabilistic projections of GTE and DSL for a specific locations, highlighting their sensitivity to assessment choices.