Online Marine & coastal projections 3.2 High++ scenario
In Chapter 1 we introduced the concept of an range for vulnerability testing above our estimated uncertainty range. Here we describe the development of the time-mean sea level component of this scenario. Sea level increases are given from present day (1980–1999) to 2095 for H++, but no time series is presented.
Data which relates to climate changes over the past hundreds of thousands of years can be found in proxy records, for example, in deep ocean sediments, corals or ice cores from the ice sheets. Some such records can be used to infer estimates of past sea level changes. These are indirect estimates, but they provide a possibility of looking at past climates which may bear some relation to projections for the future. Records relating to the last interglacial period climate (about 125,000 years ago), at which time the major continental ice sheets were similarly located to today and the global mean surface temperatures were comparable to those projected for coming decades (Otto-Bliesner et al. 2006), may offer some insight into possible future sea level changes.
Some of the proxy data suggests the possibility that future sea level rise might be greater than the maximum given in Section 3.5 (based on regionalisation of the IPCC Fourth Assessment Projections). Such inferences made from proxy data, together with known limitations in the physics of ice sheet models used in the projections, have led us to provide here a low probability, high impact range for sea level rise around the UK, which we call the High-plus-plus (H++) scenario. This might be used for contingency planning and to help users thinking about the limits to adaptation. We think it very unlikely* that the upper limit of this scenario will occur during the 21st century but cannot yet rule it out completely given past climate proxy observations and current model limitations.
Using Red Sea sediment data, Rohling et al. (2008) estimate average rates of sea level rise during the last interglacial period of 1.6±0.8 m per century. From this we derive an upper limit of 2.5 m sea level rise for our maximum global mean sea level rise over the 21st century in the H++ scenario (from 1990–2095). We reiterate that while we cannot rule out this amount of global sea level rise, recent observations and model projections do not provide any evidence to suggest it will occur. This amount of sea level rise would require a massive increase in the current observed contribution of ice sheets to sea level rise.
For the maximum sea level rise around the UK in the H++ scenario, we need to adapt the global 2.5 m sea level rise to consider regional deviations from the global mean. For scenarios dominated by thermal expansion components, such as those which give the relative sea level rise estimates in Section 3.5, regional deviations from the global mean are mainly caused by ocean circulation and regional variations in expansion of the ocean. For the maximum sea level rise in the H++ scenario, however, where the global mean sea level rise is dominated by ice sheet melt, changes in the ice load on Greenland can affect regional sea level though mechanisms. Spatial patterns for this have been estimated for particular changes in the Greenland and Antarctic ice sheets (for instance, Tamisiea et al, 2001). Allowing for these regional adjustments gives an estimate for the average sea level rise around the UK, under the H++ scenario, of 1.9 m.
One piece of evidence which may relate to the potential for long term acceleration of loss of ice from the ice sheets is from recent observational studies (Rignot and Kanagaratnam, 2006). This suggests that the loss of freshwater from the Greenland ice sheet to the sea has doubled in the last ten years. It is not yet clear, however, if this observed change is part of a long term trend or decadal variability. This increased contribution to sea level rise is from accelerated loss into the sea of ice at the margins, as well as from increased liquid water runoff relative to the accumulation of snow. Parts of the West Antarctic ice sheet rest on bedrock below sea level and so the melting of its fringing glaciers is sensitive to increases in the surrounding ocean temperature. Many of these glaciers have also seen increased speeds as their floating ice tongues (ice shelves) have thinned and, in some cases, broken up entirely (Rignot, 2006). However, even if the tide water glaciers, the fastest flowing glaciers around Greenland, were to increase their discharge of ice to the ocean by an order of magnitude, they would still only raise sea level of order 10–20 cm by 2100 (estimated using values given in Rignot and Kanagaratnam, 2006). The fastest flowing glaciers around Antarctica are currently a factor of about 4 slower than those in Greenland (Rignot and Kanagaratnam, 2006; Rignot et al. 2008). Using a simple scaling of the estimated recent contribution to sea level changes from accelerated ice flow with global mean surface temperature, the IPCC Fourth Assessment estimated that this might give up to 17 cm (for the High emissions scenario) additional global mean sea level rise during the 21st century. However, whilst they did not rule out larger increases, they noted that rapid ice sheet changes, such as the collapse of the West Antarctic Ice Sheet are not considered likely to occur in the 21st century, and we support this view here. Adding the 17 cm scaled discharge contribution to our maximum previous estimate for the UK (Table 3.3; 95th percentile) gives us 92.8 cm of sea level increase, which we take to be the bottom of the H++ range.
Recently Pfeffer et al. (2008) provided an alternative estimate of constraints on 21st century SLR. They consider the degree of acceleration of outlet glaciers and ice streams on Greenland and Antarctica that would lead to large increases in SLR. After considering maximum observed glacial movement rates they concluded SLR in excess of 2 m was physically untenable. When our slightly larger thermal expansion estimate is combined with Pfeffer et al.’s (2008) ice melt and GIA is allowed for, a worst case risk rise is again estimated at approximately 2 m for the UK region. This alternative evidence for 2 m as a sensible maximum value in sensitivity testing adds extra confidence in our earlier estimates for the top of the H++ range.
In summary, our H++ scenario range for time-mean sea level rise around the UK is 93 cm to approximately 1.9 m. Beyond our qualitative statement that the top of this range is very unlikely to occur in the 21st century we make no attempt here to assign a precise probability to this event. Improvements in models and continued monitoring may, in the future, help us to estimate the likelihood of this type of event or rule it out completely.