|Working Group I: The Scientific Basis|
|Other reports in this collection|
11.2 Factors Contributing to Sea Level Change
11.2.1 Ocean Processes
The pattern of sea level in ocean basins is maintained by atmospheric pressure and air-sea fluxes of momentum (surface wind stress), heat and fresh water (precipitation, evaporation, and fresh-water runoff from the land). The ocean is strongly density stratified with motion preferentially along density surfaces (e.g. Ledwell et al., 1993, 1998). This allows properties of water masses, set by interaction with the atmosphere or sea ice, to be carried thousands of kilometres into the ocean interior and thus provides a pathway for warming of surface waters to enter the ocean interior.
As the ocean warms, the density decreases and thus even at constant mass the volume of the ocean increases. This thermal expansion (or steric sea level rise) occurs at all ocean temperatures and is one of the major contributors to sea level changes during the 20th and 21st centuries. Water at higher temperature or under greater pressure (i.e., at greater depth) expands more for a given heat input, so the global average expansion is affected by the distribution of heat within the ocean. Salinity changes within the ocean also have a significant impact on the local density and thus local sea level, but have little effect on global average sea level change.
The rate of climate change depends strongly on the rate at which heat is removed from the ocean surface layers into the ocean interior; if heat is taken up more readily, climate change is retarded but sea level rises more rapidly. Climate change simulation requires a model which represents the sequestration of heat in the ocean and the evolution of temperature as a function of depth.
The large heat capacity of the ocean means that there will be considerable delay before the full effects of surface warming are felt throughout the depth of the ocean. As a result, the ocean will not be in equilibrium and global average sea level will continue to rise for centuries after atmospheric greenhouse gas concentrations have stabilised.
The geographical distribution of sea level change is principally determined by alterations to the ocean density structure, with consequent effects on ocean circulation, caused by the modified surface momentum, heat and water fluxes. Hsieh and Bryan (1996) have demonstrated how the first signals of sea level rise are propagated rapidly from a source region (for instance, a region of heat input) but that full adjustment takes place more slowly. As a result, the geographical distribution of sea level change may take many decades to centuries to arrive at its final state.
Previous IPCC sea level change assessments (Warrick and Oerlemans, 1990; Warrick et al., 1996) noted that there were a number of time-series which indicate warming of the ocean and a resultant thermal expansion (i.e. a steric sea level rise) but there was limited geographical coverage. Comparison of recent ocean temperature data sets (particularly those collected during the World Ocean Circulation Experiment) with historical data is beginning to reveal large-scale changes in the ocean interior. (Section 220.127.116.11 includes additional material on ocean warming, including studies for which there are no estimates of ocean thermal expansion.) However, the absence of comprehensive long ocean time-series data makes detection of trends difficult and prone to contamination by decadal and interannual variability. While there has been some work on interannual variability in the North Atlantic (e.g. Levitus, 1989a,b, 1990) and North Pacific (e.g. Yasuda and Hanawa, 1997; Zhang and Levitus, 1997), few studies have focused on long-term trends.
The most convincing evidence of ocean warming is for the North Atlantic. An almost constant rate of interior warming, with implied steric sea level rise, is found over 73 years at Ocean Station S (south-east of Bermuda). Comparisons of trans-ocean sections show that these changes are widespread (Table 11.1). On decadal time-scales, variations in surface steric height from station S compare well with sea level at Bermuda (Roemmich, 1990) and appear to be driven by changes in the wind stress curl (Sturges and Hong, 1995; Sturges et al., 1998). Variability in the western North Atlantic (Curry et al., 1998) is related to changes in convective activity in the Labrador Sea (Dickson et al., 1996). Over the 20 years up to the early 1990s there has been a cooling of the Labrador Sea Water (as in the Irminger Sea, Read and Gould, 1992), and more recently in the western North Atlantic (Koltermann et al., 1999). For the South Atlantic, changes are more uncertain, particularly those early in the 20th century.
A warming of the Atlantic layer in the Arctic Ocean is deduced by comparison of modern oceanographic sections collected on board ice-breakers (e.g., Quadfasel et al., 1991; Carmack et al., 1997; Swift et al., 1997) and submarines (e.g. Morison et al., 1998; Steele and Boyd, 1998) with Russian Arctic Ocean atlases compiled from decades of earlier data (Treshnikov, 1977; Gorshkov, 1983). It is not yet clear whether these changes result from a climate trend or, as argued by Grotefendt et al. (1998), from decadal variability. The published studies do not report estimates of steric sea level changes; we note that a warming of 1°C over the central 200 m of the Atlantic layer would result in a local rise of steric sea level of 10 to 20 mm.
Observations from the Pacific and Indian Oceans cover a relatively short period, so any changes seen may be a result of decadal variability. Wong (1999), Wong et al. (1999), Bindoff and McDougall (1994) and Johnson and Orsi (1997) studied changes in the South Pacific. Bindoff and McDougall (2000) studied changes in the southern Indian Ocean. These authors found changes in temperature and salinity in the upper hundreds of metres of the ocean which are consistent with a model of surface warming and freshening in the formation regions of the water masses and their subsequent subduction into the upper ocean. Such basin-scale changes are not merely a result of vertical thermocline heave, as might result from variability in surface winds.
In the only global analysis to date, Levitus et al. (2000) finds the ocean has stored 20x1022 J of heat between 1955 and 1995 (an average of 0.5 Wm-2), with over half of this occurring in the upper 300 m for a rate of warming of 0.7°C/century. The steric sea level rise equivalent is 0.55 mm/yr, with maxima in the sub-tropical gyre of the North Atlantic and the tropical eastern Pacific.
In summary, while the evidence is still incomplete, there are widespread indications of thermal expansion, particularly in the sub-tropical gyres, of the order 1 mm/yr (Table 11.1). The evidence is most convincing for the North Atlantic but it also extends into the Pacific and Indian Oceans. The only area where cooling has been observed is in the sub-polar gyre of the North Atlantic and perhaps the North Pacific sub-polar gyre.
Other reports in this collection