CoAuthors

Benoit Meyssignac (LEGOS/CNES, France); Xavier Fettweis (Liege University, Belgium); Ben Marzeion (Innsbruck University, Austria); Giorgio Spada (Urbino University, Italy); Elsa Bourgeois (CNES/LEGOS, France); Angelique Melet (LEGOS/CNES, France); Anny Cazenave (LEGOS/CNES, France)

Over the 20th century, tide gauge records indicate a rise in global mean sea level of 1.7 mm.y-1 +/- 0.2 mm.yr-1 (Church and White 2011). This rise is essentially due to the warming of the ocean (which made the ocean water expand and sea level rise by ∼0.6 mm.yr-1) and the melt of mountain glaciers (which added more water to the ocean and made sea level rise by ∼ 0.6 mm.yr-1, Gregory et al. 2013). Land water changes and Greenland surface mass balance changes also played a role but of lesser importance ( ∼ 0.1 mm.yr-1 for the land water storage and ∼0.2 mm.yr-1 for the Greenland mass balance since 1900, Gregory et al. 2013). Each of these different contributors to the 20th century global mean sea level rise had also an impact on the regional sea level changes. This regional signal, which must be added to the global sea level rise to compute the total sea level signal, is essential when we want to assess the impacts of the 20th century sea level rise on coastal areas and low lying islands.

In this study we aim to estimate this 20th century regional sea level changes. We use historical runs of the CMIP5 (Coupled Model Intercomparison Project Phase 5) coupled climate models, which cover the period 1850-2005, and extend them until 2012 with the RCP8.5 run. With these simulations, we estimate the different contributors to the regional sea level changes since 1900. The ocean warming contribution is directly computed from the temperature and salinity outputs of the CMIP5 models while the mountain glaciers and the Greenland contributions are computed from offline models using the surface temperature and precipitations output of CMIP5 models. Concerning the landwater storage contribution, it is actually mainly of anthropogenic origin over the 20th century
(Ngoduc et al. 2005, Meyssignac et Cazenave 2012, IPCC AR5 2013) so it is not modeled in CMIP5 models, and we do not include them in this study. As a result, we present for each different contributors and for the sum of all, a map of associated sea level changes over the 20th century (average sea level changes over the period 1993-2012 with respect to the period 1860-1880). We then compare the models' results to observations. We show that over the altimetric era (1993-2012), the observed regional trends in sea level are not well reproduced by CMIP5 models. Large discrepancies are found in the tropical Pacific and seem to be due to a failure of CMIP5 models in reproducing the current regional trends in ocean warming.