6.3.2 Climate and sea-level scenarios

In terms of climate change, the SRES scenarios in Section 6.3.1 translate into six greenhouse-gas emission ‘marker’ scenarios: one each for the A2, B1 and B2 worlds, and three scenarios for the A1 world – A1T (non-fossil fuel sources), A1B (balanced fuel sources) and A1FI (fossil-intensive fuel sources) (Nakićenović and Swart, 2000). B1 produces the lowest emissions and A1FI produces the highest emissions (see Chapter 2).

Table 6.2 summarises the range of potential drivers of climate change impacts in coastal areas, including the results from Meehl et al. (2007) and Christensen et al. (2007). In most cases there will be significant regional variations in the changes, and any impacts will be the result of the interaction between these climate change drivers and other drivers of change, leading to diverse effects and vulnerabilities (Sections 6.2 and 6.4).

Understanding of the relevant climate-change drivers for coastal areas has improved since the TAR. Projected global mean changes under the SRES scenarios are summarised in Table 6.3. As atmospheric CO₂ levels increase, more CO₂ is absorbed by surface waters, decreasing seawater pH and carbonate saturation (Andersson et al., 2003; Royal Society, 2005; Turley et al., 2006). A significant increase in atmospheric CO₂ concentration appears virtually certain (Table 6.3). Sea surface temperatures are also virtually certain to rise significantly (Table 6.3), although less than the global mean temperature rise. The rise will not be spatially uniform, with possible intensification of ENSO and time variability which suggests greater change in extremes with important implications for coral reefs (Box 6.1).

The global mean sea-level rise scenarios (Table 6.3) are based on thermal expansion and ice melt; the best estimate shows an acceleration of up to 2.4 times compared to the 20th century. These projections are smaller than those of Church et al. (2001), reflecting improved understanding, especially of estimates of ocean heat uptake. If recently observed increases in ice discharge rates from the Greenland and Antarctic ice sheets were to increase linearly with global mean temperature change, this would add a 0.05 to 0.11 m rise for the A1FI scenario over the 21st century (Meehl et al., 2007). (Large and long-term sea-level rise beyond 2100 is considered in Box 6.6.)

Importantly, local (or relative) changes in sea level depart from the global mean trend due to regional variations in oceanic level change and geological uplift/subsidence; it is relative sea-level change that drives impacts and is of concern to coastal managers (Nicholls and Klein, 2005; Harvey, 2006a). Meehl et al. (2007) found that regional sea-level change will depart significantly from the global mean trends in Table 6.3: for the A1B scenario the spatial standard deviation by the 2080s is 0.08 m, with a larger rise than average in the Arctic. While there is currently insufficient understanding to develop detailed scenarios, Hulme et al. (2002) suggested that impact analysis should explore additional sea-level rise scenarios of +50% the amount of global mean rise, plus uplift/subsidence, to assess the full range of possible change. Although this approach has been followed in the UK (Pearson et al., 2005; Thorne et al., 2006), its application elsewhere is limited to date.

Furthermore, coasts subsiding due to natural or human-induced causes will experience larger relative rises in sea level (Bird, 2000). In some locations, such as deltas and coastal cities, this effect can be significant (Dixon et al., 2006; Ericson et al., 2006).

Increases of extreme sea levels due to rises in mean sea level and/or changes in storm characteristics (Table 6.2) are of widespread concern (Box 6.2). Meehl et al. (2007) found that models suggest both tropical and extra-tropical storm intensity will increase. This implies additional coastal impacts than attributable to sea-level rise alone, especially for tropical and mid-latitude coastal systems. Increases in tropical cyclone intensity over the past three decades are consistent with the observed changes in SST (Emanuel, 2005; Webster et al., 2005). Changes in other storm characteristics are less certain and the number of tropical and extra-tropical storms might even reduce (Meehl et al., 2007). Similarly, future wave climate is uncertain, although extreme wave heights will likely increase with more intense storms (Meehl et al., 2007). Changes in runoff driven by changes to the hydrological cycle appear likely, but the uncertainties are large. Milly et al. (2005) showed increased discharges to coastal waters in the Arctic, in northern Argentina and southern Brazil, parts of the Indian sub-continent, China and Australia, while reduced discharges to coastal waters are suggested in southern Argentina and Chile, Western and Southern Africa, and in the Mediterranean Basin. The additional effects of catchment management also need to be considered (Table 6.1).

Box 6.2. Examples of extreme water level simulations for impact studies

Although inundation by increases in mean sea level over the 21st century and beyond will be a problem for unprotected low-lying areas, the most devastating impacts are likely to be associated with changes in extreme sea levels resulting from the passage of storms (e.g., Gornitz et al., 2002), especially as more intense tropical and extra-tropical storms are expected (Meehl et al., 2007). Simulations show that future changes are likely to be spatially variable, and a high level of detail can be modelled (see also Box 11.5 in Christensen et al. (2007).

Figures 6.4 and 6.5 are based on barotropic surge models driven by climate change projections for two flood-prone regions. In the northern Bay of Bengal, simulated changes in storminess cause changes in extreme water levels. When added to consistent relative sea-level rise scenarios, these result in increases in extreme water levels across the Bay, especially near Kolkata (Figure 6.4a). Around the UK, extreme high sea levels also occur. The largest change near London has important implications for flood defence (Figure 6.4b; Dawson et al., 2005; Lavery and Donovan, 2005). Figure 6.5 shows the change in flooding due to climate change for Cairns (Australia). It is based on a combination of stochastic sampling and dynamic modelling. This assumes a 10% increase in tropical cyclone intensity, implying more flooding than sea-level rise alone would suggest. However, detailed patterns and magnitudes of changes in extreme water levels remain uncertain (e.g., Lowe and Gregory, 2005); better quantification of this uncertainty and further field validation would support wider application of such scenarios.

Figure 6.4. Increases in the height (m) of the 50-year extreme water level. (a) In the northern Bay of Bengal under the IS92a climate scenario in 2040-2060 (K – Kolkata (Calcutta), C – Chittagong) (adapted from Mitchell et al., 2006). (b) Around the UK for the A2 scenario in the 2080s (L – London; H – Hamburg) (adapted from Lowe and Gregory, 2005).

Figure 6.5. Flooding around Cairns, Australia during the >100 year return-period event under current and 2050 climate conditions based on a 2xCO2 scenario. The road network is shown in black (based on McInnes et al., 2003).