Armin Köhl (CEN, Uni Hamburg, Germany); Detlef Stammer (CEN, Uni Hamburg, Germany); Benoit Messignyac (LEGOS, CNES, France); John Church (CSIRO, Australia); Jens Schröter (AWI, Germany); Manfred Wenzel (AWI, Germany)

Estimates of sea level variability and trends are being presented as they follow from tide gauge reconstructions (TGR) and ocean data assimilation (ODA) results over the 48-yr period 1960-2007. Statistical tide gauge reconstructions are based on empirical orthogonal functions, or are trained by altimetric variability patterns, and use tide gauge data data to extrapolate recent regional sea level changes backward in time. Reconstructions also exist from dynamical ocean modeling approaches using ocean data as constraints (data assimilation). Although tide gauge reconstructions match tide gauge data better than ODA results, they exhibit less variability in the open ocean where regional open-ocean sea-level variability found from altimetric data are not well reproduced by TGRs over substantial portions of the ocean. Nevertheless, TGRs match observed trends and variability better during the satellite-altimetry era, but less so for the entire 48-yr period. Over periods earlier than the satellite era, all reconstructed regional patterns deviate substantially from each other. Thus, an open question remains regarding how well any of these products reproduce historical open-ocean SSH prior to altimetry. An ensemble average of all products compares to tide gauge data nearly as well as the better tide gauge reconstructions over the pre-altimetry era, and compares well with the individual products at matching satellite altimetry variability.