Detecting anthropogenic footprints in regional and global sea level rise since 1900

Abstract:

While there is scientific consensus that global and local mean sea level (GMSL and LMSL) is rising since the late 19^th century, it remains unclear how much of this rise is due to natural variability or anthropogenic forcing. Distinguishing both contributions requires an extensive knowledge about the persistence of natural high and low stands in GMSL and LMSL. This is challenging, since observational time series represent the superposition of various processes with different spectral properties. Here we provide a probabilistic upper range of long-term persistent natural GMSL/LMSL variability (P=0.99), which in turn determines the minimum/maximum anthropogenic contribution since 1900. To account for different spectral characteristics of various contributing processes, we separate LMSL (corrected for vertical land motion) into a slowly varying volumetric (mass and density changes) and a more rapidly changing atmospheric component. Based on a combination of spectral analyses of tide gauge records, barotropic and baroclinic ocean models and numerical Monte-Carlo experiments, we find that in records where transient atmospheric processes dominate the spectra, the persistence of natural volumetric changes tends to be underestimated. If each component is assessed separately, natural centennial trends are locally up to ~1.0 mm/yr larger than in case of an integrated assessment, therefore erroneously enhancing the significance of anthropogenic footprints. The GMSL, however, remains unaffected by such biases. On the basis of a model assessment of the separate components, we conclude that it is virtually certain (P=0.99) that at least 45% of the observed increase in GMSL is of anthropogenic origin.