| 英文摘要: | Sea-level rise1 is one of the most pressing aspects of anthropogenic global warming with far-reaching consequences for coastal societies. However, sea-level rise did2, 3, 4, 5, 6, 7 and will strongly vary from coast to coast8, 9, 10. Here we investigate the long-term internal variability effects on centennial projections of dynamic sea level (DSL), the local departure from the globally averaged sea level. A large ensemble of global warming integrations has been conducted with a climate model, where each realization was forced by identical CO2 increase but started from different atmospheric and oceanic initial conditions. In large parts of the mid- and high latitudes, the ensemble spread of the projected centennial DSL trends is of the same order of magnitude as the globally averaged steric sea-level rise, suggesting that internal variability cannot be ignored when assessing twenty-first-century DSL trends. The ensemble spread is considerably reduced in the mid- to high latitudes when only the atmospheric initial conditions differ while keeping the oceanic initial state identical; indicating that centennial DSL projections are strongly dependent on ocean initial conditions. Globally averaged sea level has risen by about 20 cm since 1900 and at a rate of about 3 mm yr−1 during the past two decades, but with strong regional variation1, 2, 3, 4, 5, 6, 7. For example, the western tropical Pacific featured a much stronger rise than the global average during the recent decades, whereas even falling sea levels were observed in the eastern tropical Pacific and along the west coast of the Americas7. Dynamic sea surface topography, the departure from the Earth’s geoid, is influenced by ocean currents, local mass balance and density changes of the water column11, 12, 13, 14. The DSL, which is the focus of this study, has been introduced to describe the collective effect of the local steric (thermosteric and halosteric) and dynamical ocean adjustment contribution9, 11, 14. As the observed sea-level changes include the effects of both external forcing (natural, for example, solar; and anthropogenic, for example, CO2) and internal variability, we need to understand both drivers to assess twentieth and twenty-first-century sea-level changes. Globally averaged sea level has risen by about 20 cm since 1900 and at a rate of about 3 mm yr−1 during the past two decades, but with strong regional variation1, 2, 3, 4, 5, 6, 7. For example, the western tropical Pacific featured a much stronger rise than the global average during the recent decades, whereas even falling sea levels were observed in the eastern tropical Pacific and along the west coast of the Americas7. Dynamic sea surface topography, the departure from the Earth’s geoid, is influenced by ocean currents, local mass balance and density changes of the water column11, 12, 13, 14. The DSL, which is the focus of this study, has been introduced to describe the collective effect of the local steric (thermosteric and halosteric) and dynamical ocean adjustment contribution9, 11, 14. As the observed sea-level changes include the effects of both external forcing (natural, for example, solar; and anthropogenic, for example, CO2) and internal variability, we need to understand both drivers to assess twentieth and twenty-first-century sea-level changes. Climate modes, patterns with identifiable characteristics and specific regional effects, are prominent examples of internal variability. The El Niño/Southern Oscillation15, a quasi-periodic fluctuation of the equatorial Pacific sea surface temperature with a period of about 4 years, is the leading mode of tropical interannual variability. El Niño/Southern Oscillation is associated with zonal redistributions of heat, causing large sea-level anomalies across the equatorial Pacific and along the west coasts of the Americas15. The Pacific Decadal Oscillation, a decadal climate mode, also strongly affects sea level in the Pacific2, 16, 17 and tropical South Indian Ocean18. Other regions of strong internal decadal sea-level variations are the North19, 20 and South Atlantic19. Here we address the influence of the longer centennial variability21, 22, 23, 24, 25 on DSL projections for the twenty-first century. To this end we investigate five large ensembles of global warming integrations with climate models. We first discuss two ensembles performed with the Kiel Climate Model24, 25, 26 (KCM, Methods). The annual-mean DSL climatology from a millennial control run with fixed atmospheric CO2 (348 ppm) depicts the well-known features of the large-scale ocean circulation (Fig. 1a), and the pattern is in good agreement with that derived from satellite altimetry (Fig. 1b). However, there are noticeable biases. For example, the ‘northwest corner’ southeast of Newfoundland is missing and the North Atlantic subpolar gyre extends too far to the east (Fig. 1), which may influence the pattern of forced DSL changes near the east coast of North America and in the North Atlantic. Both ensembles conducted with the KCM consist of 22 global warming integrations, where each ensemble member was forced by increasing atmospheric CO2 at a rate of 1% per year (compound), approximately corresponding to the observed present rate of increase. Each ensemble member was integrated for 100 years. CO2 doubling is reached after about 70 years, and the integrations were continued for another 30 years with constant CO2. The ensemble-mean response is a measure of the CO2-forced signal, whereas the ensemble spread reflects the effects of internal variability.
|
![Comparison of the KCM[rsquor]s DSL, the local deviation from the global average sea level, with satellite observations.](http://www.nature.com/nclimate/journal/v5/n4/images_article/nclimate2569-f1.jpg)