• Thermal expansion: For this the values of the IPCC are adopted (0.4–0.9 m: IPCC, 2001a, their Fig. 11.15a), which are derived from model simulations for a scenario with doubled CO2.
• Glaciers: For the volume of all glaciers outside of Greenland and the Antarctic the same IPCC report gives a sea-level equivalent of 0.5m; with 3 °C of global warming one could expect a loss of 80 per cent of the glacial mass for the year 2300. A more recent study (Raper and Braithwaite, 2006), however, uses half of this value; therefore a range of 0.2–0.4 m is applied.
• Greenland: The model presented by IPCC (2001a) for Greenland with a local warming of 5.5 °C (which is a plausible value with 3 °C global warming: Chylek and Lohmann, 2005) gives a sea-level rise contribution of 0.9 m by the year 2300. The dynamic mechanisms mentioned above, however, are not considered, so this value represents a lower limit; therefore 0.9–1.8 m is assumed here.
• Antarctic: The behaviour of the West Antarctic ice sheet (WAIS) is critical for the Antarctic. In 2001, IPCC considered the decay of this ice sheet to be very unlikely, because the then existing models suggested that the continental ice did not react to changes in the ice shelves floating in the adjacent sea. This now has to be considered as disproved, as the observations discussed above show. The disappearance of further ice shelves (like Larsen B: Fig. 3.1-3) due to warming of seawater means that the melting of the WAIS must be feared with a similar time frame as Greenland. For this, 1–2 m of sea-level rise is assumed by the year 2300. At a constant rate this corresponds to the disappearance of the WAIS in a time frame of 900–1800 years – some glaciologists consider that broad destruction is even possible within 300–400 years.
The net result is a rise of around 3–5 m by the year 2300. The value of 3 m corresponds to a loss of one-sixth of each, the Greenland and the West Antarctic ice sheets; 5 m corresponds to one-third of each (Table 3.1-1).

3.2   Impacts on coastal regions

The consequences of climate change, whether in the form of sea-level rise or through greater frequency and force of extreme weather events, will directly affect the future development of coastal regions. The worldwide length of coastlines (excluding small protrusions of less than a few kilometres) is on the order of around one million kilometres. Coastal regions are of extreme importance for humankind. They offer settlement areas, are centres of economic activity (Turner et al., 1996) and, not least, harbour a rich abundance of biological diversity.
     The direct effects of climate change, such as the extent and rate of sea-level rise, presently cannot be precisely determined. But it is very probable that the threat to coastal regions will increase considerably, as will the number of people affected by climate change. This is an obvious result of the fact that large numbers of settlements are located near the coasts. Eight of the world’s ten largest cities today lie on the coast (UN, 2004), and according to estimates 21 per cent of the world’s human population live less than 30 km from the sea (Cohen et al., 1997; Gommes et al., 1998). The great attraction of coastal regions is also reflected in the large growth rates of populations there, which is around twice the global average (Bijlsma et al., 1996). The worldwide trend toward urbanization will amplify this development in the future. By the year 2030 approximately 50 per cent of the world population could be living within 100 km of the coasts (Small and Nicholls, 2003).
     How sea-level rise and weather extremes due to climate change will affect coastal regions and societies depends primarily on the kind and number of affected natural and social systems. The natural systems will mainly be represented by river deltas, low-lying coastal plains, coral islands and atolls, barrier islands and lagoons, beaches, coastal wetlands and estuaries (IPCC, 2001b). The following sections explore in detail which biogeophysical and socio-economic impacts can be expected and to what extent people are threatened..

3.2.1   Biogeophysical impacts

3.2.1.1     Inundation due to sea-level rise

The rise in mean sea level will result in the inundation of coastal areas and island groups in several regions of the world. Inundation here is defined as the permanent covering of land areas with water (as opposed to temporary, episodic flooding). Without counter-measures the result will be the irretrievable loss of this land.
     In order to be able to estimate the total extent of the regions endangered by sea-level rise, Brooks et al. (2006) have compiled data on the global land-surface distribution with respect to the elevation above sea level. Figure 3.2-1 illustrates that large regions lie within the range of one metre above high water. Above the one-metre line, the land-surface distribution rises as an almost linear function of the elevation above the mean high water line. At an elevation of only 20 m above sea level a total land area of 8 million km2 would be affected.

Figure 3.2-1
Distribution of land area, excluding Antarctica, as a function of elevation above present mean high water (MHW).
Source: ISciences, 2003

     For purposes of illustrating the spatial distribution of these land areas, examples will be shown of regions that lie at elevations within 2 m and within 20 m above sea level. A rise of 20 m (Fig. 3.2-2) represents an extreme scenario that could result from anthropogenic warming over a time frame of around 1000 years, in the event that the ice sheets of Greenland and west Antarctica should melt for the most part (Section 3.1.1). This long time frame has to be considered with sea-level rise because the relevant processes, such as melting of the ice sheets and mixing of the ocean, are slow geophysical processes. Because of the physical inertia in the marine system these processes will first come to a standstill centuries after stabilization of the greenhouse gas concentrations and the surface climate.

Figure 3.2-2
Coastal areas in Europe, parts of western Asia and North Africa. Areas below 20 m elevation above the present mean sea level are coloured red (not taking future coastal defence measures into account).
Source: Brooks et al., 2006

     The particularly threatened areas in Europe with a rise of 20 m would be mainly eastern England, the Po Delta in northern Italy, and the coastal strips running through Belgium, the Netherlands, north-western Germany, and into northern Denmark (Fig. 3.2-2).
     A sea-level rise of 2 m (Figs. 3.2-3 and 3.2-4) could occur in the coming century. As an illustration of the effects, Figure 3.2-3 depicts regions on the North Sea and the northern European coast. Because this kind of illustration is based on the absolute elevation above sea level, it also includes areas that are protected by dikes today. Some densely populated areas in the Netherlands, England, Germany and Italy today already lie below the normal high-water level (EEA, 2005). For these regions the sea-level rise is especially threatening. Here the question of the rate of change takes on a special importance, because a more rapid rise could hamper the implementation of adaptive strategies (Brooks et al., 2006).

 

Figure 3.2-3
Coastal areas along the North Sea. Areas below 2 m elevation above the present mean sea level are coloured red (not taking future coastal defence measures into account).
Source: Brooks et al., 2006

     In Asia, with a sea-level rise of 2 m (Fig. 3.2-4), for example, the densely populated river delta of the Ganges-Brahmaputra-Meghna with its network of 230 rivers would be affected. The total river region covers an area of 175 million hectares and stretches from India and Bangladesh to Nepal, China and Bhutan (Mirza et al., 2003). Approximately 129 million people presently live in this river delta (Woodroffe et al., 2006), with a large portion of them in rural areas. With Dhaka and Kolkata (formerly Calcutta) there are already two fast-growing megacities here, that is, cities with more than ten million inhabitants.

Figure 3.2-4
Coastal areas along the Gulf of Bengal and in the Ganges-Brahmaputra-Meghna River Delta. Areas below 2 m elevation above the present mean sea level are coloured red (not taking future coastal defence measures into account).
Source: Brooks et al., 2006

3.2.1.2     Flooding as a result of storm surges

IIn most cases the most destructive results of sea-level rise will not be from the very slow rise of the mean water level, but in the increasing occurrences of storm surges.
The origin of storm surges is often related to the interplay of storm systems and tides. When storms push water onto the coasts at high tide it can lead to the flooding of large areas of land. Especially in river estuaries damage can occur over large inland distances (SwissRe, 1998). The word flooding here describes a temporally limited, partial or complete water cover of normally dry areas. This can be caused by the rise of surface water (still or flowing) over its banks, as well as by the results of strong precipitation (Münchener Rück, 1997).
     Sea-level rise increases the exposure of coastal inhabitants to storm surges and storm waves, and with it the risk of flooding. The destructive force of these kinds of weather extremes increases as a direct consequence of sea-level rise (Jimenez and Sanchez-Arcilla, 1997). Higher waves will more easily reach the original coastline and also penetrate farther inland. Even the water levels of the two-metre scenario exceed today’s standards for coastal defence structures. Although Great Britain, for example, has protective structures that reduce the wave height near the coasts, it is questionable whether these measures can provide long-term protection when the exceptional situation becomes the normal case. If water depths should change or shores become steeper, which would result in a direct energy increase for waves coming onto the land, then the existing structures would no longer be sufficient as coastal protection measures (Burgess and Townend, 2004).
     Additional factors could significantly increase the risks from flooding: changes in oceanic and atmospheric circulation patterns caused by climate change can influence storms and their destructive potential at regional and local scale. For example, an increase in the strength of tropical cyclones is anticipated (Section 3.1.2). Furthermore, climate warming could contribute to intensification of the hydrological cycle, which makes increases in the frequency and intensity of extreme precipitation events likely (IPCC, 2001a).
For the consequences of sea-level rise it is less critical how much higher the average water height is than how frequent certain high levels are reached during storm surges. This can be estimated by a comparison of the expected average rise with statistics of past storm surges. Accordingly, the return periods, i.e., the time interval between certain critical gauge levels, could be strongly reduced in the future (Lowe et al., 2001). A model by the Hadley Centre for a region in eastern England, based on a combination of meteorological data and an assumed sea-level rise of 0.5 m by 2100, shows a reduction in the return period of high-water events from 500 to 12 years (Lowe et al., 2001). Similar trends were calculated for the greater New York City area, based on various climate scenarios. According to these, with a sea-level rise of 24–95 cm the return period of a 100-year flood in the 2080s is shortened to 4 to 60 years (Gornitz et al., 2002; Section 3.3). When the return periods of destructive extreme events become too short, the repeated repair of damaged infrastructures would no longer make sense, and they would have to be abandoned.
     Land-use changes such as the clearing of forests, urbanization and the conversion of alluvial plains and wetlands can further increase the flooding risk, for example by weakening the water-retention capacity of the soil (Kundzewicz and Schellnhuber, 2004). Straightened or built-up rivers without natural forests and wetlands have less buffer capacity in extreme situations. The flow and sedimentation behaviour of rivers influenced by engineering measures often determines whether storm-caused flooding risks are amplified or attenuated.

3.2.1.3     Coastal erosion

In contrast to floods, which are relatively rare events with sometimes catastrophic results, erosion represents an episodically occurring process (Hall et al., 2002). During the erosion process, waves carry solid materials such as sand, mud and rocks away from the coast and redeposit them for the most part in other formations. A rise of sea level could accelerate these erosion processes (Zhang et al., 2004; Stive, 2004). In particular with a small rise, erosion may prove to be of greater importance than flooding (Smith and Lazo, 2001).
     The erosion rates depend on the local conditions. If undercutting occurs along with the resulting collapse of steep coasts or coastal protection structures, erosion can represent a serious danger. In this connection it is important to note that above all the rates of sea-level rise are relevant to changes in coastal morphology. If sedimentation rates can keep up with the sea-level rise, then a new equilibrium can be established and this can have a stabilizing effect on the coastal processes. Sedimentation processes have contributed to coastal development since the beginning of the Holocene, and have been responsible for the preservation of land areas, especially during the inundation of river deltas (Brooks et al., 2006). However, if the sea-level rise accelerates so quickly that a new equilibrium cannot be established, or if sedimentation rates are significantly reduced due to management measures, then a loss of coastal strips will probably result. A well-known example is the Nile, where the sedimentation rates were decreased, primarily by the construction of the Aswan Dam, which led to accelerated erosion of the northern Nile Delta by tides (Stanley and Warne, 1998).
     Many authors, including Zhang et al. (2004), refer to the Bruun Rule (Bruun, 1962) in their prognoses of the erosion of coastal regions caused by sea-level rise. This states that the erosion rates are approximately 50–100 times higher than the relative rate of sea-level rise, that is, a sea-level rise of 1 m would result in the loss of a 50–100 m wide coastal strip. Opinions about the general applicability of the Bruun Rule, however, vary widely, because it is based on the assumption of a simple, two-dimensional system, and the establishment of a sedimentation equilibrium in the bank area. These preconditions, however, can hardly be assumed in real situations. It therefore must be concluded that for estimating the results of erosion along the coastlines, more complex models have to be applied that also incorporate, for example, the sediment transport along the coasts and changes in sedimentation equilibrium, as would be the case with a sea-level rise.

3.2.1.4     Impacts on groundwater

The rise of sea level can also cause the groundwater level of a coastal region to rise. This is determined in part by geographic factors (e.g., elevation above sea level), and in part by geological factors (e.g., properties of rock and soil layers).
     Above all, a rise in the groundwater level caused by sea-level rise could impact on river deltas up to 20–50 km inland. This estimate is based primarily on the observation that groundwater along the coasts flows above a dense, landward-moving saltwater wedge. This balance between freshwater and saltwater is physically controlled by the relationship of the different water densities. With a rise of sea level, therefore, the overlying groundwater would also rise (Barlow, 2003). This can lead to a saturation of the soil. It produces impacts not only on the freshwater supply of a region, but also on agriculture (salinization risk), the stability of foundations, and the safety and functioning of dewatering and other underground systems such as subways.
In addition, a rise in sea level can promote the invasion of saltwater in coastal aquifers (seawater intrusion).      Model simulations by Sherif and Singh (1999) concluded that a rise of 0.5 m in the Mediterranean Sea would result in the intrusion of saltwater 9 km into the coastal aquifers of the Nile Delta. With the same rise, however, in the Gulf of Bengal, only a zone of 0.4 km landward would be affected. These processes would result in an increased salinization of the groundwater and surface waters, with considerable impacts on agriculture and the drinking-water supply. The intrusion of saltwater into the groundwater reservoirs of coastal zones can already be observed worldwide today, e.g. in China and India (Shah et al., 2000). Through the increasing over-exploitation of freshwater resources in the densely populated coastal zones, this process can also be considerably amplified.
     Although sea-level rise can cause the salinization of river estuaries and near-coastal groundwater reservoirs, this process is determined by a number of other factors. The run-off behaviour of precipitation and its contribution to groundwater recharging also controls seawater intrusion in coastal regions. An increase of freshwater runoff can counteract seawater intrusion. Seawater intrusion would have long-lasting effects. Some of these aquifers would require hundreds to thousands of years to reach a new harmonic equilibrium with the new sea level (Barlow, 2003).

3.2.1.5     Biological impacts

Besides temperature increase and acidification, the expected sea-level rise is an important additional stress factor for the often highly species-rich terrestrial coastal ecosystems or near-coast ecosystems. Two particularly relevant ecosystem types are coral reefs (Section 2.4) and mangrove forests, because they not only harbour great biological diversity, but at the same time play an important role in coastal protection. This latter was illustrated by the tsunami catastrophe in December 2004 in the Indian Ocean: on coasts with intact coral reefs and mangrove forests the flood wave was slowed considerably so that the damage was less disastrous (Fernando and McCulley, 2005; Dahdouh-Guebas et al., 2005; Danielsen et al., 2005).
     How coral reefs will respond to sea-level rise can be derived by reconstructions of the past or by model simulations. The adaptive capacity of corals in prehistoric times varied greatly (Montaggioni, 2005). The average vertical growth rate of coral reefs since the last ice age is reported as at most 10mm per year (IPCC, 2001b). But because the growth rate of corals is influenced by many factors (Section 2.4), and corals in this century will also be impaired by warming, acidification and other environmental factors, a prognosis for the adaptive capacity of these ecosystems cannot be made with respect to the rising water level.
     Around 8 per cent of the coastlines worldwide today are bordered by mangroves. More than half of the mangrove forests have already disappeared (WRI, 2001). The observed decline can be attributed in large part to changes in human uses of coastal zones. A study on the changes of mangrove belts in the Amazon region (Cohen and Lara, 2003) shows that sea-level rise can also have a local effect on the distribution of mangroves. The rise of sea level in the future will force the near-coastal mangrove belts farther inland. The mangroves, however, will only be able to survive in areas where enough space is left for them adjacent to the intensive human land use. For the preservation of this valuable ecosystem it is therefore urgently necessary to maintain protected areas or create new ones that include a wide buffer zone on solid land. With the help of the HadCM3 model, Nicholls (2004) evaluated the sensitivity of coastal regions to flooding under the different SRES scenarios (IPCC, 2000). In every case the sea-level rise results in the loss of wetland areas. This study also shows, however, that the direct destruction of wetlands by people could exceed the losses caused by climate change.
     Changes in the tidal ranges and high-water levels caused by sea-level rise are an additional burden for coastal ecosystems. The consequences include changes in water depths, light and temperature, and current speeds, and a shift in the freshwater-saltwater distribution. These can lead to physiological burdens for some animal and plant species that could then require a habitat change. Studies show that even minor seawater intrusions into coastal seas lead to large disturbances in the structure and diversity of zooplankton populations. Accordingly, small salinity changes can result in a decline in the biodiversity of coastal ecosystems (Schallenberg et al., 2003). The functioning and preservation of ecosystems are therefore not only threatened by flooding because of sea-level rise, but also by changes in the frequency and strength of seawater intrusions.
     The DIVA model (DINAS-COAST Consortium, 2004) is a new interactive tool for integrated analysis of the results of sea-level rise. The model simulates the effects of local sea-level rise (including tectonic rises and falls) on the ecosystems and populations of the coastal regions of the world, and incorporates different adaptive strategies. It is based on the analysis of the worldwide coast lines in more than 10,000 homogenous segments according to morphological and socio-economic aspects, a self-developed extensive worldwide database, and a series of coupled modules. For a scenario with a mean sea-level rise of 50 cm by the year 2100 the model reports a loss of more than half of the freshwater wetlands in coastal regions, around 20 per cent of the coastal forests, and a quarter of the mangroves.

3.2.2   Socio-economic impacts

3.2.2.1     Impacts on people

The multiple effects of sea-level rise on the natural environment will have a major impact on people and the systems they depend on. It is likely that some of these effects will interact, intensifying each other, such as floods and erosion-related events. For inhabitants of coastal regions, sea-level rise will be the biggest challenge posed by global climate change (IPCC, 2001b).
    The extent of climate-related hazard will also depend on the extent to which the ecosystems of the affected coastal regions have been exposed to prior damage. Pre-existing environmental problems often interact with the impacts of climate change. For instance, flood risk may be increased by changes in land use (deforestation, settlement, etc.) in hydrological catchment areas, or degradation of coastal ecosystems (coral death caused by marine pollution, logging of mangrove forests for building materials and to clear land for aquaculture installations, etc.). Moreover, it has been observed that in some cities the land mass is sinking below sea level. Contributing factors in this case include both the physical pressure of buildings and infrastructure and intensive urban management practices, combined with groundwater extraction, drainage and building activity. Nicholls et al. (1995) estimate that, at their most extreme, local rates of subsidence may be as much as 1 m per decade. A rise in sea level then makes the risk of flooding in these regions even greater. The fact that a variety of factors are superimposed on each other – disappearance of natural barriers, sinking of land masses below sea level, and rising sea levels resulting from climate change – increases the risk to populations (Nicholls, 2003).
    Based on 1995 population figures, there are currently 60 million people living below 1 m elevation and 275 million below 5 m elevation above mean sea level. If estimates are adjusted to take into account forecasts of population growth, the figures for the end of the 21st century rise to 130 million (below 1 m elevation) and 410 million (below 5 m elevation; Nicholls et al., 2005). A more recent study by Brooks et al. (2006) arrived at similar findings (Fig. 3.2-5).

Figure 3.2-5
Population living below a certain elevation above mean high water (MHW) in 1995.
Source: Brooks et al., 2006

    How people in threatened areas will ultimately deal with the challenges of an accelerated sea-level rise is a complex and dynamic process. Migration away from threatened areas will depend on the particular situation in a given locality, and can range from planned migration based on risk assessments and economic considerations to the sudden displacement of people fleeing floods, storms, or sudden erosion-related events. Due to the likely increase in extreme weather events, the incidence of spontaneous migration following natural disasters will probably exceed that of planned migration (Brooks et al., 2006). This is especially likely to happen where a radical change in the landscape occurs and the costs of protecting the affected population become disproportionately high. People in low-lying coastal regions, especially river deltas and small island states, are particularly at risk in this regard (Nicholls, 2003). Studies show, for example, that unless costly protective measures are put in place, a sea-level rise of around 0.5 m would put 1.5 million people at risk in the Egyptian governerates of Alexandria and Port Said (El-Raey et al., 1999). In the case of Europe, estimates suggest that 13 million people would be at risk in the event of a sea-level rise of 1 m (EEA, 2005).
    There is a whole range of model simulations designed to obtain a more precise estimate of the number of people exposed to flood risks. Using the FUND model, Nicholls et al. (2006), for example, have simulated the consequences of disintegration of the West Antarctic ice sheet and the resulting sea-level rise of 5 m over a period of 100 to 1000 years beginning in the year 2030. The impact of coastal protection measures was evaluated using cost-benefit analysis. In all scenarios, population displacement reaches a peak between 2030 and 2060. Based on the (extreme) assumption that the ice sheet will disintegrate rapidly within 100 years, up to 350,000 persons per year will be forced to leave their homes over a period of ten years. This would give a total of 15 million people. However, these figures account for a mere 2–3 per cent of the total number of people at risk, because they are based on the assumption that coastal protection measures will be implemented on a large scale. In another study aimed at estimating flood risks, Hall et al. (2005) arrive at the conclusion that, under the A1 and A2-SRES scenarios, the number of people at risk in Britain in the 2080s compared to 2002 would double, rising from 0.9 million to 1.8 million.

‘Sea-level refugees’

Whether coastal dwellers who are forced to leave their homeland due to climate-related environmental changes (‘sea-level refugees’) return home or settle further away from the coast will depend on a whole host of factors. On the one hand, the decision will be influenced by whether coastal protection measures are put in place and how effective or reliable these are. On the other hand, the position adopted by local and regional government will also play a role, for example by discouraging or even prohibiting return to evacuated areas (Brooks et al., 2006). Actual numbers of sea-level refugees will ultimately be determined by the interplay of these factors and measures.
    In any case, in the long term sea-level refugees will need to be resettled elsewhere, and this poses new challenges for policy. This is especially true in the case of the inhabitants of some of the low-lying atolls such as the Maldives, the Marshall Islands, Kiribati, Tuvalu or Tokelau. These island states, with a total population of more than 500,000 (CIA, 2005), lie a mere 2 m above sea level on average and are therefore at risk of becoming uninhabitable or disappearing completely as a result of climate change. Their inhabitants face a constantly increasing risk of salinization and drinking water shortages and higher risk of storms and floods even if the 1 m guard rail (Section 3.3) is successfully adhered to (Barnett and Adger, 2003). These factors are already making their impact felt: the first relocations to higher-lying land took place in December 2005 on the Pacific island of Vanuatu. In this particular case, decreasing intervals between storm surges had made it necessary to relocate the village of Lateu. The United Nations Environment Programme regards this case as probably the first formally recorded resettlement measure of its kind, resulting directly from the consequences of climate change (UNEP, 2005).
    Official programmes are already in place to tackle the problem of sea-level refugees. New Zealand has reached agreement with the governments of Tuvalu, Fiji, Kiribati and Tonga on immigration regulations for their inhabitants under the ‘Pacific Access Category’. Each year, a certain number of refugees whose status is a direct result of the consequences of climate change are granted a New Zealand residence permit. A whole set of conditions is attached to obtaining a residence permit under these arrangements, however, and older people and poor people are currently excluded (Friends of the Earth, 2005). The right of sea-level refugees to be granted refuge in other countries needs to be enshrined in international law (Section 3.4.2.3).

Threats to human health

    In coastal areas, the primary threat to the lives and health of large numbers of people is posed by storms and floods. Even today, a total of 75 million people in coastal regions are exposed to the risk of storm-induced floods. Assuming a moderate climate change scenario with a sea-level rise of 0.4 m by the 2080s, this figure would rise to an estimated 200 million (IPCC, 2001b; Patz et al., 2005).
    When assessing health-related consequences of storm tides and floods, a distinction can be drawn between immediate, medium-term and long-term impacts. Immediate impacts refers to impacts arising during the event itself and which are due to the effects of flooding. These include death and injury due to drowning or collision with hard objects, and to hypothermia and cardiac arrest (WHO, 2002). In this context, the World Health Organization (WHO) has calculated that in the year 2030 the relative risk of death due to flooding in the coastal areas of the EUR-B Region will be 6.3 times higher than in the base years 1980–1999 (McMichael et al., 2004). Affected countries in the EUR-B Region include some of the former Soviet republics, several Balkan states, Turkey, Poland and the EU accession states Bulgaria and Romania.
    The medium-term impacts of floods manifest themselves most notably in an increase in infectious diseases resulting from ingestion of or contact with contaminated water (e.g. cholera, hepatitis A, or leptospirosis), or respiratory infections due to overcrowded accommodation (IPCC, 2001b). The lack of properly functioning sanitary installations and public healthcare provision makes these risks even greater in poorer countries. Following the floods in Bangladesh in 1988, for example, the most common diseases were diarrhoea and respiratory infections, while the most frequent cause of mortality for all age groups under 45 years was acute watery diarrhoea (Siddique et al., 1991).
    In the longer term, the consequences of sea-level rise could influence the frequency and distribution of disease vectors. Inundation of coastal regions, for example, affects the incidence of mosquito species that breed in brackish water, e.g. the malaria vectors Anopheles subpictus and A. sundaicus in Asia. Floods could, however, also destroy the natural habitat of some pathogens, such as the EEE virus (eastern equine encephalitis virus) found in freshwater swamp areas along the US coastline (IPCC, 2001b).
    In addition, the rise in sea level and the consequences of storm surges and floods pose a risk to drinking water supplies and food security. This is a matter of increasing salinization of freshwater reservoirs, which not only affects drinking water supply, but can also adversely affect agricultural productivity in the vicinity of the coast. At the same time, floods can also lead to considerable crop losses, as for example in the case of the 1998 floods in Bangladesh, where rice losses accounted for more than half of total agricultural losses and resulted in annual agricultural production falling to a mere 24 per cent of the expected total. Potential consequences include food shortages and undernourishment (del Ninno et al., 2001; WHO, 2002).
    As a result of the shock and the consequences of such events, floods can also have long-term effects on the psychological wellbeing of the people affected. Loss of family members and friends, social networks, property and employment can lead to post-traumatic stress syndrome. This manifests itself in feelings of anxiety, depression, psychosocial disorders, and indeed can lead to an increase in suicide rates. It must be taken into account that psychological problems of this sort may not emerge until months or years after an event of this sort (WHO, 2002).

	Categories	Damage or losses
	Infrastructure	Buildings, transport infrastructure (roads, rail networks, ports, airports), energy infrastructure, coastal protection structures
	Economic sectors	Fisheries, agriculture, forestry (timber in mangrove forests), tourism, transport
	Human wellbeing	Mortality, spread of diseases, migration/displacement of people, loss of landscapes and cultural assets
	Ecosystems	Services from coastal ecosystems, biological diversity, including some species-rich islands, disruption of the freshwater/saltwater balance

Table 3.2-1
Classification of damage caused by a rise in sea level.
Source: adapted from Fankhauser, 1995

    According to a study by the World Health Organization, more than 150,000 people are already dying every year due to the consequences of climate change (WHO, 2002). The primary causes of mortality in this context are increased incidence of diarrhoea, malaria and undernourishment. WHO estimates suggest that additional health risks resulting from climate change will more than double worldwide by the year 2030 (McMichael et al., 2004). These estimates are based on forecasts of a sharp increase in the relative risk of floods, with smaller increases in malaria, undernourishment and diarrhoea. Although smaller relative changes in these phenomena are forecast, they have the potential to bring about disease on a much bigger scale. Infectious diseases thus seem to present a greater risk to humanity than the direct impact of sea-level rise. However, the models used at present do not take account of potential interactions among the various health risks.

3.2.2.2     Economic damage

AAssessing the economic impact of climate change on coastal areas also presents scientists with considerable challenges. To be able to make any statement on the overall costs of the impact of sea-level related climate change requires detailed analysis that is also highly disaggregated in geographical terms so as to enable estimation of the expected damage. Such damage may take a wide variety of forms, ranging from damage to property to costs arising due to loss of human life or loss of biological diversity and ecosystem services. Table 3.2-1 gives examples of sectors of society affected by sea-level rise and of the damage and losses to be expected.
   In order to assess potential physical damage and impacts on people, it must be borne in mind that a large number of megacities will be affected by a rise in sea level. Of the 20 megacities throughout the world, 15 are exposed to the sea (calculated on the basis of data from Klein et al., 2002; UN, 2004). These include Tokyo, Mumbai and New York. Since development of megacities often entails exacerbation of existing local environmental problems, such as lowering of the groundwater level, these areas often lack any natural buffering capacity to balance the consequences of a rise in sea level. In such cases, drinking water supplies could be jeopardized. This is an example of what is termed ‘critical infrastructure’ (Bruneau et al., 2003; DRM, 2006), a category that also includes transport, telecommunications and energy supply networks, and emergency, rescue and health services; it further includes the retail sector, public administration, banking and finance. Critical infrastructure refers to institutions that fulfil vital needs and guarantee public safety, uphold law and order and ensure provision of basic public services and a functioning economy (Commission of the European Communities, 2005a). Disruption or damage to this infrastructure can result in supply bottlenecks and significantly impair public safety (BBK, 2006), and may even have a destabilizing effect on a whole region.    For example, a gradual rise in sea level and extreme events resulting from it could interfere with the functioning of major ports and at times bring them to a halt altogether, with a knock-on effect on regional trade and transport networks. Geophysical changes to coasts are thus also likely to have large-scale economic impacts on neighbouring and inland regions (Brooks et al., 2006).
   In addition to the costs arising from physical damage or disruption to production, there are also costs resulting from the loss of ecosystem services. For example, the negative impact of sea-level rise on coastal ecosystems can adversely affect local fishing yields (Brooks et al., 2006). In many countries, especially poorer countries, the security of people’s livelihoods often depends directly on the yield from these ecosystems. Any disturbance in the freshwater balance, for example by seawater intrusion (Section 3.2.1.4), can also affect agriculture. Increasing groundwater salinization has already damaged common agricultural land on the islands of Tuvalu (Friends of the Earth, 2005). As well as jeopardizing food supply, this also brings about a decline in local economic activity.
   The overall costs of climate change include on the one hand the damage caused by climate change in monetary units, and on the other the costs of adapting to climate change. Adaptation measures must be implemented in accordance with the principle of economic efficiency so that the benefits of the measures (in the form of damage prevented) outweigh the costs (for example costs associated with construction and maintenance of sea walls). In other cases, it may be more sensible in strict economic terms to forego adaptation measures altogether and accept climate change-induced damage. A cost-effective portfolio of strategies will also depend on environmental and socio-economic conditions in a given region, and these can change over time. For planning strategies and decision-making, it is important that categories of costs and benefits associated are thoroughly explored and taken into account (Section 3.4.1.1).
   A great deal of data is required in order to assess the overall costs of climate change worldwide. Detailed information is particularly needed for identifying and assessing potential damage. Available information, however, is often far from comprehensive and indeed may be quite rudimentary, especially in developing countries. As shown in Table 3.2-1, damage may take a variety of forms and also includes goods that are not traded on the market and therefore have no price. This applies particularly to loss of ecosystem services and biodiversity, which may be quantified in economic terms with the aid of surveys and economic methods of estimation. However, this is an area that is fraught with uncertainty.
   People will not simply put up with the effects of climate change. They will protect themselves from damage by putting measures in place to adapt. Economic analysis must therefore include exploring cost-effective combinations of strategies. To do this, models are needed that can simulate not only climate change but also national economic development worldwide. Although some models of this sort are already in use (Fankhauser, 1995; Yohe et al., 1999; Darwin and Tol, 2001), they are based on highly simplified assumptions, with the result that estimates of global costs can only be calculated very roughly and are therefore of limited expressiveness. Data from regional vulnerability analysis, however, enable more accurate estimation of the costs of climate change due to sea-level rise, at least in smaller areas (e.g. Box 3.4-2).

3.3   Guard rail: Sea-level rise

3.3.1   Recommended guard rail

WBGU recommends the following guard rail: absolute sea-level rise should not exceed 1 m in the long term (even over several centuries), and the rate of rise should remain below 5 cm per decade at all times. For comparison: total anthropogenic sea-level rise up to now has been 20 cm; current rates are around 3cm per decade (Section 3.1).

3.3.2   Rationale

The recommended levels are based on WBGU’s estimation that a higher or more rapid rise in sea level would in all probability cause damage and losses to humankind and nature that exceed tolerable levels. As is generally the case with guard rails, this estimation contains a normative component and is not solely based on scientific principles (Box 1-1), given that there continues to be considerable uncertainty surrounding the actual consequences of sea-level rise. WBGU hopes that this proposal will stimulate broad debate within society on what is an acceptable degree of sea-level rise and stimulate further research on its consequences.
   As in the case of WBGU’s climate guard rail on the increase in global temperature (a total of 2 ºC and not more than 0.2 ºC per decade; Box 1-1), the consequences of sea-level rise depend both on the overall figure and on the rate. Effects on structures that are non-moveable in the long term, such as cities and world cultural heritage sites, depend to a greater extent on the absolute figure, while the rate of rise tends to be more important for dynamic systems such as ecosystems, beaches and some coral atolls, which are able to adapt to some degree. Between the two – in other words between the overall figure and rate – there is a variable degree of trade-off, in the sense that a higher absolute value may be tolerated if the rate is slower, while the maximum rate is tolerable at best for a very short time.

Absolute rise

In order to justify setting an absolute guard rail for sea-level rise that must be adhered to even in the long term, one must consider the consequences of a possible very slow rise in sea level. Based on current knowledge, in the view of WBGU a rise of more than 1 m would be intolerable, because severe consequences would be virtually unavoidable even with a very long period of adaptation. This applies, for example, to a whole series of megacities in close proximity to the coast, such as New York, Lagos or Kinshasa.
New York City consists of several islands and peninsulas and has around 1000km of coastline (Bloomfield et al., 1999). Figure 3.3-1 shows the areas of southern Manhattan that would be inundated in the event of a ‘one-hundred year flood’ (water levels 3 m above normal levels) at today’s sea levels. In this case, massive damage could be expected to occur, with flooding of important infrastructure including some subway stations. Statistically, if there is a sea-level rise of 1m, this storm surge level would be attained not just once a century, but every four years. A ‘one-hundred year flood’ would then extend correspondingly further into the streets of Manhattan.
    Similar storm problems are to be expected in other cities and in large river deltas (e.g. the Yellow River, the Yangtze, the Ganges-Brahmaputra, the Mississippi or the Nile). In developing countries, poor population groups are often concentrated in these endangered areas.

Figure 3.3-1
Inundated areas (blue) in lower Manhattan (New York) in a statistically typical one-hundred year storm event based on the present sea level. A sea-level rise of 1 m would result in storm tides of this height approximately every four years.
Source: Rosenzweig and Solecki, 2001; data based on USGS, U.S. Army Corps of Engineers, Marquise McGraw, NASA GISS

     In its first report, the IPCC listed a whole series of island states that would face a considerable threat from sea-level rise. Many small island states would lose a significant proportion of their land if the sea level rose by 1 m (IPCC, 1990). Some of the islands are at risk of becoming uninhabitable due to storm surges resulting from a sea-level rise of this magnitude. Affected islands include, for example, the Maldives, Kiribati, Tuvalu and the Marshall Islands, with a total population of 523,000 people. These problems are exacerbated by the increase in tropical cyclone intensity (Section 3.1.2). Around another 380,000 people living on the Caribbean islands of Anguilla, Cayman Islands, Turks and Caicos Islands and the island state of the Bahamas would also be affected by this. Although some of these islands have high ground of up to 65 m above sea level, with the rise in sea level, storm floods would penetrate further and further inland. In many cases, virtually the whole of the island’s infrastructure (e.g. airports, roads) is located directly on the coast.
    If the sea level rises by more than 1m, there is an additional risk that cultural heritage sites will be irretrievably lost. Cultural goods from the past possess ‘outstanding universal value’ (UNESCO, 1972). In view of this fact, in 1972 UNESCO adopted the International Convention for the Protection of the World Cultural and Natural Heritage. Some 180 countries are now signatories to this Convention. An important component of world heritage is its universality; it belongs to all individuals and peoples of the world, irrespective of the territory in which it is located.
    Great importance should therefore be given to protecting these world heritage sites. A sea-level rise of more than 1 m would pose a direct threat for example to the 12th century Shinto shrine of Itsukushima in Japan and the 8th century Shore Temple in Mahabalipuram in India. Both are important religious sites whose special character derives from their coastal location. To protect these sites from sea-level rise, one option might be to consider removing the monuments to another site. This would involve at least some loss, as the monuments are symbolically and historically rooted in their present environment.
    A sea-level rise of 1 m would also put, among other places, Venice and St. Petersburg at considerable risk. In the storms of 1966, when flood levels peaked at 2 m above normal, large areas of Venice were submerged. Homes and businesses were destroyed as a result, but so too were valuable works of art (Nosengo, 2003). In St. Petersburg too, storms could have devastating consequences. A researcher at the European Bank for Reconstruction and Development (EBRD) suggests that a storm-induced rise in water levels of 2.5 m would inundate around 10 per cent of the city, while a rise in excess of this level could affect up to one-third of the city (Walsh, 2003). As a result of these dangers, extensive projects are currently under way to build protective structures; in the case of St. Petersburg, international funding is also involved.
    Many valuable coastal ecosystems would also be threatened by a sea-level rise of this sort, for example the Kakadu National Park in Australia and the mangrove forests of the Sundarbans National Park in Bangladesh and India (UNESCO, 2006).

Rate of rise

The rate of sea-level rise should not overstretch the adaptive capacity of human society or marine and coastal ecosystems.
    The adaptive capacity of ecosystems can be estimated using the example of coral reefs, mangrove forests and beaches. The last great rise in the sea level occurred at the end of the last Ice Age, over the period from 18,000 to 5,000 years before present. Since then, the rate of rise has always been less than 20cm per hundred years, and usually well below this (Walbroeck et al., 2002; Peltier, 2004). In the Holocene era, after this last great sea-level rise came to an end, coral reefs, beaches, mangrove forests and other ecosystems were able to become established again along the newly formed coastline.
The maximum vertical growth of coral reefs is estimated at 10 cm per decade (IPCC, 2001b). If the conditions are highly favourable, they could therefore presumably keep pace with this rate of sea-level rise. Future growth rates will be markedly slower, however, due to ocean acidification and warming and other environmental stresses (Section 2.4).
    The adaptive capacity of mangrove forests and beaches is highly dependent on sediment accretion. Sand beach loss already observed along many coastlines is considered to be a consequence of sea-level rise (Leatherman, 2001). Ellison and Stoddart (1991) analyse the development of mangrove forests during the Holocene and arrive at the conclusion that in a situation where there is little sediment accretion, even the current rate of sea-level rise places excessive demands on adaptive capacity and will result in the loss of mangrove forests. Other authors (Snedaker et al., 1994), meanwhile, argue that if the habitat is favourable, retreat of mangrove forests further inland could enable them to accommodate an even higher rate of sea-level rise. In many cases, however, such favourable conditions will not be present. Based on a scenario with an almost linear sea-level rise of 5 cm per decade, the global DIVA model (Section 3.2.1.5) projects a continuous loss of mangrove forests whose adaptive capacity has thus already been exceeded. By 2100, according to this projection, a quarter of all mangrove forests would disappear.
    According to the scenarios postulated by IPCC (2001a) the rate of sea-level rise towards the end of this century will be 3–7 cm per decade, with up to 13cm per decade in the worst-case scenario. In view of these facts, WBGU recommends setting the guard rail for maximum sea-level rise at no more than 5cm per decade. It must be borne in mind, however, that even compliance with this guard rail will not provide protection from damage that is already significant, as is also the case with other WBGU guard rails
(Box 1-1).

3.3.3   Feasibility

The current and foreseeable rise in sea level is almost entirely anthropogenic, and hence its future development can also be influenced by humankind. The ability to control it is limited on the one hand by the long time-scale required (centuries) for a response in terms of sea-level change. It is also limited by forecasting difficulties and by the potential for strongly non-linear behaviour on the part of the great continental ice sheets. Nevertheless, the recommended guard rails can be implemented, according to current knowledge, by means of an appropriate climate change mitigation strategy.
    Stabilizing the global temperature at 2 ºC above pre-industrial levels, according to the mathematical models, would result in a sea-level rise of around 50 cm in the long term (after 1000 years) simply due to thermal expansion. Mountain glaciers would add approximately another 20 cm to this (Section 3.1.1.4). Prevention of large-scale melting of the continental ice sheets in Greenland and Antarctica would therefore be critical for compliance with the guard rail. Further research must establish the limit that needs to be set as regards the rise in global mean temperature in order to achieve this. It is conceivable that, in the long term, it may be necessary to reduce the temperature to below the 2 °C threshold again.
    In this century, the guard rail for the rate of sea-level rise would only be breached by the more pessimistic half of the IPCC scenarios (2001a); the more optimistic scenarios comply with the guard rail even without climate protection measures. It should be borne in mind, however, that the currently observed rate of rise of 3 cm per decade is already clearly higher than all of these scenarios (Fig. 3.1-4). It must therefore be assumed that, in all likelihood, the IPCC (2001a) has underestimated sea-level rise, and that climate mitigation measures are indeed required to comply with this guard rail. Based on the assumption that the change in sea-level rise will be relatively smooth and gradual, as depicted in all the scenarios, compliance with the guard rail for the rate of sea-level rise would mean a maximum rise in sea level of around 40 cm in the 21st century. This would be double the sea-level rise that has taken place to date as a result of human activity.
    The climate and sea-level rise guard rails are closely interlinked, since sea-level rise is directly caused by global warming. In the next few decades, the climate protection strategies required to meet the 2 °C goal and to comply with the guard rails relating to sea-level rise will most likely be similar and compatible. Despite the long-term nature of sea-level rise and the uncertainties surrounding the behaviour of the continental ice sheets, these guard rails are not redundant. Even if the global warming guard rail is obeyed and lasting climate warming of 2 °C takes place, this would be enough to cause melting of the Greenland ice sheets, thereby breaching the guard rail on sea-level rise. For this reason, it is conceivable that the guard rail on sea-level rise will lead to the imposition of strict limits on emissions, especially in the long term, in other words in the coming centuries, in order to stabilize the continental ice sheets.
    This is why it is vital, as regards emissions, to embark on a path that will lead to stabilization of the global temperature at a low level after 2100, and if possible well below 2 ºC above the pre-industrial level. The guard rail on sea-level rise therefore determines in particular longer-term climate protection goals from the second half of the century onwards. In the coming decades, it is a key additional justification in support of the 2 °C goal. If, on the other hand, the continental ice sheets of Greenland and Antarctica were to shrink suddenly and unexpectedly, the guard rail on sea-level rise could require tougher climate protection measures than the 2 °C climate guard rail even sooner. It thus gives particular grounds for closer observation of the ice sheets in order to identify dangerous developments in time.

3.4   Recommendations for action: Develop and implement adaptation strategies

In its previous reports on climate policy, WBGU has made it clear that priority should be given to strategies for preventing greenhouse gas emissions. However, even if there is substantial success in preventing greenhouse gas emissions and complying with the guard rail on sea-level rise, it will no longer be possible to prevent some of the effects of climate change on coastal areas. Appropriate adaptation measures are required in order to cope with these effects. As regards strategies for adapting to sea-level rise and extreme weather events, WBGU focuses on two questions in particular:

3.4.1   Adapting coastal regions to the consequences of climate change

3.4.1.1   Adaptation options: Classification and assessment

The extent to which the consequences of climate change will give rise to damage in coastal areas and turn hazards into disasters varies considerably from one region to another and depends on the vulnerability of the areas affected. This in turn depends on the susceptibility and resilience of natural, social, infrastructural, economic, institutional and cultural subsystems (Titus et al., 1991; Klein et al., 1999). Resilience in this context means the ability of subsystems to cope with repeated disruption so that key structures and processes remain intact (Burton and Lim, 2001; Burton et al., 2002; Adger et al., 2005).
Industrial countries will be better able to deal with hazards than developing countries, because they have more extensive capacities at their disposal, such as an efficient institutional infrastructure, technical know-how and financial resources. Hurricane Andrew, for example, a category 5 event on the Saffir-Simpson hurricane scale, cost the lives of 23 people in the USA in 1992. A typhoon of comparable strength that hit Bangladesh in 1991, meanwhile, led to extensive flooding that resulted in 100,000 deaths and millions of refugees (Adger et al., 2005).
    The large number of influencing factors and interactions makes it essential to develop adaptation strategies that are tailored to the given context. Adaptation in this context needs to fulfil two purposes: on the one hand to reduce damage, and on the other to increase resilience of the above subsystems. There are basically three different options for adaptation in response to the hazards outlined above: ‘protection’, ‘managed retreat’ and ‘accommodation’ (IPCC, 2001b).

Protection

Protection involves protecting coasts from rising sea levels by means of structural measures. These might include ‘hard’ engineering measures such as construction of sea walls, dykes, or flood defence systems, and ‘soft’ measures such as conservation or introduction of protective coastal ecosystems (e.g. wetlands, mangroves, islands) or beach nourishment as natural barriers. Hard structural adaptation measures are exceedingly cost-intensive in terms of construction and maintenance. In addition, they increase stress on neighbouring ecosystems; for example, they increase the threat of wetland loss. Without intervention, wetlands will tend, as a rule, to migrate inland in the event of floods. This autonomous adaptation is impeded by sea wall construction, because areas on the seaward side of sea walls become inundated, while on the landward side new wetlands are prevented from forming. In the case of US coasts, it is estimated that 50 per cent of all wetlands have disappeared as a result of this process (Titus, 1990). In the coastal regions of the EU, moreover, it has been observed that adaptation involving hard engineering measures can trigger or accelerate erosion processes in neighbouring coastal areas. This in turn can significantly impair the functioning of hard protection measures (Commission of the European Communities, 2005b; Brooks et al., 2006). Due to the multitude of problems associated with hard structural adaptation measures, preference is given nowadays to ‘soft’ measures wherever possible. Soft strategies interfere less with coastal ecosystems and permit a more flexible response to sea-level rise, the extent of which is fraught with uncertainty. Ultimately, however, the effectiveness of soft and hard measures will depend on the environmental and societal context.

Managed retreat

Managed retreat means that use of areas in proximity to the coast is reduced, or certain areas are relinquished completely. In this context, strategies might include moving buildings and settlements, and introducing government regulation on the use of vulnerable areas. Retreat may be enforced by means of public order legislation, e.g. by regulating land use under national construction and planning law. Another means is to provide incentives in favour of the decision to retreat voluntarily. Measures of this sort encourage households and private businesses to take account of all costs relating to use of the coast in their decisions to invest. A targeted information policy implemented by local public bodies could help to enhance awareness of the implications of climate-induced risks.
    In certain cases, a sensible option may be to actively support resettlement of people from the coast to less threatened areas, for instance by providing grants via the regional administrative bodies, or within the framework of international development cooperation.
    The issue of resettling communities and their residents arises in a very concrete way in the aftermath of a natural disaster, in other words, when infrastructure has been destroyed over a large area. A decision must then be made as to whether reconstruction is economically sensible according to the prognoses relating to future sea-level rise and the incidence and intensity of extreme weather events. The more residents can rely on the government to share the costs of protection measures, the more they will be inclined to remain in threatened regions. If, however, each individual is confronted with the costs of protection, the attractiveness of reconstruction decreases and more people will opt to migrate to less threatened areas. In order to provide the right incentives in such instances, therefore, government (and international) aid for reconstruction must be tied to a corresponding relocation condition. Municipalities, too, must weigh up their adaptation options and decide between protection and retreat. After a natural disaster, they will tend to rebuild destroyed infrastructure rapidly to be able to ensure continuity in public services. This is why it is important to develop strategies for resettlement for threatened areas before a natural disaster strikes (Brooks et al., 2006).
It is conceivable that, despite government incentives to encourage migration away from threatened coastal areas and an adequate information policy on the part of public institutions, some people will not agree to relocate on a voluntary basis. In such a situation, the government must decide whether it will permit the affected people to remain and face damage to property and life at their own risk, or whether it will forcibly relocate sections of the population. The latter option, of course, carries considerable potential for conflict (Box 3.4-1).

Box 3.4-1 Potential for conflict over resettlement Depending on regional scenarios of threat, policy-makers must consider the option of planned resettlement of population groups. However, a large number of projects in a great variety of socio-economic and political contexts demonstrate the many problems that can arise as a result of such measures. As examples, dam projects, mining and infrastructure projects (e.g. the Three Gorges Dam in China, lignite open-cast mining in Garzweiler, Germany, road construction in metropolitan Manila, etc.) can be cited.         Although resettlement of endangered coastal residents is generally a necessity in order to protect the affected people, there is also considerable potential for conflict in this situation. For example, decisions to protect important infrastructure installations may amount to unequal treatment of different population groups (populations in the proximity of a protected installation will be protected along with it, while other settlements are evacuated). In addition, exacerbation of conflicts relating to land use in the target area for resettlement may occur (conflicts between long-standing residents and new settlers). Massive resistance is most likely to occur, however, in regions where resettlement programmes were used in the past as a repressive measure by the government.

Government measures encouraging migration away from coastal areas should go hand in hand with measures limiting migration of people into these areas. For example, levying a tax that reflects the social costs resulting from migration of people into coastal areas ensures that these costs are included in an individual’s calculation of the costs associated with migrating into the area, and thus become relevant for his decision.
Government regulation can thus fundamentally support the relocation of people in the desired direction. It is nevertheless possible to provide false incentives, for example, via interventions in insurance markets. Due to the increasing incidence of floods and cyclones, economic adjustments are in fact to be expected in the insurance markets: insurance premiums for flood damage will rise, and some private insurers will withdraw from the market. As a result, coastal areas lose their attractiveness as areas for settlement. If insurance premiums are kept artificially low by government subsidies, however, as is the case in the USA, prices are distorted and incentives to migrate away from coastal areas are reduced.

Accommodation

The third strategy, termed accommodation, involves modifying land use and subsystems to ensure that they take account of the new threats. Residents of threatened regions continue to use the threatened land, but without trying to protect it from inundation. This can take place, for example, by instituting disaster management systems (constructing emergency refuges, formulating plans of action, undertaking targeted public education and communication work). It is likewise possible to modify land use, for example by cultivating varieties of grain that are resistant to increasing salinization and inundation of the soil or by converting arable land to fish farming facilities. In addition, accommodation also includes engineering measures (such as increasing the height of buildings, making cellars and buildings water-tight).

Portfolio approach

It frequently happens that these options are implemented as a combined set of strategies rather than as alternatives. A ‘portfolio approach’ is pursued in order to respond adequately to the given conditions in a particular region. One possible combination of strategies involves partial retreat, where protective measures are applied only to areas where there is a high concentration of people, assets and functions. Flooding is allowed to take place in the other areas. Using this approach, implementation of protection measures would be prioritized, focusing on political and economic hubs such as cities, towns and industrial areas. A particular focus of attention in this context is protecting the ‘critical infrastructure’, in other words, infrastructure so essential that its destruction has a destabilizing impact on a country’s public life and economy.
Another strategy that might be considered is to combine protection with accommodation. This could involve, for example, aiming to increase coastal resilience through conservation of mangrove forests as natural barriers. In the context of local land use planning, setback areas could be created or extended to permit ecosystems to shift landward, thereby enhancing an area’s capacity for autonomous adaptation (Nicholls, 2003).

3.4.1.2   Choosing adaptation strategies

Cost-benefit analysis may be used (Box 3.4-2) to help choose appropriate adaptation strategies for a specific region. This requires comprehensive information on the state of coastal areas and on the impact of human activities. Assessment of the interaction of land and sea for commerce and industry, for port facilities, buildings, groundwater and extraction of construction material is also called for in this context (Kullenberg, 2001; SEEDS, 2005). The data required for this purpose are gathered and evaluated in the framework of vulnerability studies (Burton and Dore, 2000).
    In contrast to prevention strategies, the effects of adaptation projects are essentially local; in other words, they have no direct global environmental benefit. Moreover, because the extent of the environmental effects is fraught with uncertainty, ‘no-regret’ measures should initially be identified and put in place. These are measures that bring a net benefit for stakeholders irrespective of any actually occuring climate-induced losses. Such measures tend to win greater support among affected stakeholder groups because they take into account the uncertainties of climate change and yield desirable results even if climate change were not to occur. An example might be a coastal region with pre-existing damage and a high population density, for which a rise in sea level would exacerbate existing problems. Improving planning relating to use of coastal areas would be an appropriate strategy here for dealing with sea-level rise. Moreover, it would still bring a positive net benefit even if the anticipated effects of climate change failed to materialize.
    In practice, transaction costs, institutional failure or lack of information have frequently led to the shelving of such projects. Adaptation projects can help to dismantle these obstacles (Fankhauser, 1998). Implementation of integrated coastal zone management, for example, could help to improve the exchange of information among the different policy-makers and thereby make it easier to carry out projects.

3.4.1.3     Implementing adaptation strategies

AAdaptation requires more than simply implementing engineering options. Not only is the choice of strategies influenced by a multitude of factors; the strategies themselves impact on the subsystems of the region in which they are implemented. It is also necessary to reconcile the various responsibilities and interests of participating or affected groups in society (Nicholls, 2003).

Risk management

Risk management provides an ideal means of implementing adaptation strategies. Risk management plans designate persons responsible (public and private, at municipal, national and international level) for all stages of an event – before, during and after. They describe what measures should be taken (strategic versus tactical measures) at what point in time, and the manner in which the responsible persons should respond and to whom they should report (Boyd et al., 2005). In many instances, policy-makers do not treat the issue of climate change as a priority, and as a result, climate-induced changes in the risk situation are not adequately taken into account. Risks are thus often graded as low and threats are considered as rather unlikely, with the result that the design of available risk management plans is inappropriate. The example of Hurricane Katrina, which caused destruction on an unprecedented scale on the US coast in August 2005, shows that inadequate planning can heighten the socio-economic impact of such events.
    Formulating an appropriate risk management plan should ideally be a cyclic process. In advance of an extreme event, a planning phase (Phase 1) – in which preventative and reactive measures are devised – is followed by a preparatory phase (Phase 2). Measures included in this second phase are aimed at reducing the likelihood of potential hazards resulting in disasters. This may be done by establishing action plans, providing emergency training and conducting targeted public information and education campaigns, or by establishing agreements for international cooperation in the area of disaster assistance and for dealing with environmental refugees. If an event actually occurs, the response phase becomes relevant (Phase 3). This phase involves measures to be carried out during and after an event. Such measures include emergency response, measures to prevent consequential damage such as outbreaks of epidemics, or implementing measures aimed at speeding the recovery of affected areas. The reconstruction phase (Phase 4) concludes the process of managing such an event. All activities in this phase are aimed at restoring normal system functioning, via disbursal of insurance payments, setting up temporary emergency shelters or reconstructing the physical infrastructure. In addition to these four phases, problems are identified with regard to the handling of the event and mistakes are analysed. The experience gathered is then evaluated in a new planning phase and implemented in the form of improved strategies (Boyd et al., 2005).
    In the case of slow onset hazards, on the other hand, the priority of risk management lies in regularly assessing the potential risks and identifying the most vulnerable individuals and regions. Risk management involves adaptation to constantly changing conditions. A high degree of flexibility in terms of strategies is needed to achieve this. Such strategies include in particular scientific monitoring, public education and communication and legislative provisions (Boyd et al., 2005).

Integrated coastal zone management

In order to take account of the highly complex, interconnected nature of impacts, adaptation measures should be very broad, in other words, they should be enshrined in all key areas of policy. Examples are coastal protection plans and strategies for sustainable development. Another term used in this context is ‘integrated coastal zone management’. As part of this system of management, data on both ecosystems and social systems are collected and processed. Integrated coastal zone management as an instrument for managing risk in this context refers to a dynamic process developed and implemented on the basis of a coordinated strategy with the aim of managing environmental, socio-cultural and institutional resources so as to ensure sustainable conservation of coastal areas and ensure that they can be used in a variety of ways in future (Fankhauser, 1998; Yeung, 2001).
    Coordinating the sectoral, competing and in some cases overlapping competences of the various decision-making tiers and specialist areas within the administration presents a major challenge as regards devising an integrated coastal management system. Institutional fragmentation often gets in the way of providing adequate responses. WBGU therefore recommends creating integrated institutions that bring together all the key competences. Such institutions would also facilitate reconciliation of the diverse interests of affected groups in society. Municipalities and local administrative departments can play a key coordinating role. Providing for a high degree of local responsibility could help to ensure that available knowledge on coping strategies on the ground is used efficiently, that affected groups in society are appropriately involved in planning and decision-making processes, and, in doing so, that coastal management systems are accepted by the local population (SEEDS, 2005; WCDR, 2005; Box 3.4-2).

Box 3.4-2 Coastal management on the German North Sea coast Global forecasts on the impact of sea-level rise are not directly applicable to regional or local circumstances. Even within Germany’s coastal areas, major differences can be observed as regards the hazard situation and socio-economic resilience. To be able to put expedient measures in place to adapt to future consequences of climate change, therefore, small-scale, scenario-based studies need to be conducted that analyse both the natural and social characteristics of a given area. Studies of this sort have already been carried out for two regions of the German North Sea coast: for the island of Sylt and for the north-west German coastal region.     As a result of its particularly risky situation and economic productivity, the North Sea island of Sylt was analysed in the context of a study commissioned by the Federal Ministry for Education and Research (BMBF) entitled ‘Climate change: consequences for people and coasts’. The island is an open system with a negative sediment balance in which erosion processes occur that result in a continuous reduction in land mass. Sea-level rise is expected to accelerate these processes. The economic structure of Sylt is heavily biased towards tourism, which is concentrated on the western side of the island. Various scenarios were devised in order to estimate the consequences of climate change up to the year 2050. The scenario presented here is based on an assumed local sea-level rise of up to 25 cm and changes in wind conditions, tidal range and swell (i.e. wave height, direction of approach and period). Extreme weather and its impacts on the natural environment and socio-economic structures were not considered, so further research is needed to explore these aspects. Based on the results of the model simulations, changes in sediment transport are to be expected on the west coast of the island, which would have a negative effect on the wave-attenuating impact of an offshore reef. The three municipalities of Rantum, Hörnum and Wenningstedt are likely to be worst affected by this.     The study recommends adaptation of these coastal areas by adopting a ‘portfolio approach’, in other words, by a combination of different component strategies. Consideration is given to the three components: protection, managed retreat and accommodation. A combination of ‘soft’ and ‘hard’ coastal protection measures was identified as the optimum strategy for protecting Sylt’s current coastal form. The measures focus particularly on the environmentally sustainable option of beach nourishment, which is currently already being implemented on the west coast of the island. To evaluate the recommended adaptation measures in economic terms, cost-benefit analysis was carried out for the west coast of the island. According to this analysis, adaptation costs would consist primarily of the costs of additional beach nourishment measures. At current values, the costs involved in the period up to 2050 are estimated at E 33 million. The benefits of coastal protection in terms of prevented losses of assets, infrastructure, beaches and dunes must be balanced against this. The current value of this benefit is estimated at E 381 million. The results of the analysis for this period thus show that the benefit-cost ratio of coastal protection for the island of Sylt is clearly positive. The scenario examined illustrates that the island of Sylt can probably be protected effectively against a small rise in sea level of 25 cm by beach nourishment measures.     At the same time, it must be noted that Sylt represents a special case, with its particular geographical characteristics and its very high concentration of assets resulting from tourism. The recommended adaptation measures are unlikely to be applicable to the majority of the world’s other coastal regions. Beach nourishment, for example, is only possible, and only makes economic and environmental sense, if sand is available within the coastal region in sufficient quantities. In addition, the estimated adaptation costs are based on a coastal protection strategy aimed primarily at protecting the main part of the island, which is higher-lying compared to the shore, from erosion. Beach nourishment would be unlikely to provide adequate protection for flatter sections of coast in the face of rising sea levels – especially in the event of extreme weather situations. Building on the experiences of the Sylt study, the project ‘Climate change and preventive risk and coastal protection management on the German North Sea coast’ (KRIM) not only examines the consequences of accelerated sea-level rise for various stretches of coastline, but also the concomitant risks of extreme weather events. Using the same time horizon of 2050, this study analyses future consequences of climate change together with possible social adaptation measures and their impact. The KRIM project assumes a regional increase in temperature of 2.8 °C, a local sea-level rise of 55 cm, and changes in mean tidal range, precipitation, swell, and winter wind velocity and direction. In addition, extreme weather conditions with flood levels of +200 cm are also taken into account.     In order to analyse the consequences of this climate scenario for Germany’s north-western coastal region, the resulting risks of extreme weather events were calculated and the costs and knock-on effects for the regional economy of possible coastal protection strategies were set against these. Alternative strategies were compared using cost-benefit analysis techniques. When assessing potential damage from storm surges, the KRIM project not only considered environmental damage and economic losses from damage to assets and infrastructure, but also the resulting losses to the national economy in terms of value-added activities, income and jobs. A mesoscale modelling technique was used to assess the value of economic losses; in other words aggregated data based on official regional statistics were used. Subsequent to the value assessment, losses were calculated as a function of flood levels.     Among other things, the study examined dyke raising and construction of a second line of dykes as elements of a protection and accommodation strategy for the study area of Wangerland. The location of Wangerland (a region between river Weser and river Jade at the North Sea) is geographically very exposed, with long coastlines requiring protection, but it has relatively few assets. Based on the year 2010 as the assumed investment year (base scenario), the costs of dyke raising were estimated at E 10.5 million (for an average increase in height of 0.75 m over a 28 km stretch of dyke), while the costs of constructing a second line of dykes (variant II) were estimated at E 20 million (for a 17 km stretch of dyke with a height of 3 m above mean sea level). The current value of economic losses in the KRIM climate change scenario, meanwhile, is estimated at E 63 million (2000) – calculated using the flood simulations and for the period up to 2050. According to these calculations, the benefit-cost ratio is highest for the dyke-raising option, with the result that action was recommended to implement this coastal protection measure.     The procedure followed in the KRIM project provides a guideline for handling climate change-induced uncertainties with regard to coastal management and shows how the economic future of coastal regions could be forecast and planned. There is still a great need for further research in this area, however: (1) to explore scenarios involving higher rises in sea level; (2) to extend existing insights regarding regional losses and the costs of different preventative strategies (e.g. for the managed retreat option); (3) to include other sections of coastline in this analysis. It will ultimately require a great many small-scale studies of this type in order to be able to make more reliable supra-regional predictions regarding the financial impacts of climate change, and to be better able to work out the possible options for action. The cases described show that in these instances, where the rise in sea level is clearly below the WBGU guard rail on sea-level rise, the problems can probably be overcome by means of appropriate adaptation measures. Unfortunately, a rise in sea level of more than 1 m was not studied for these regions; in many places, successful adaptation in the event of such a high sea-level rise would no longer be possible at an acceptable cost. In developing countries, meanwhile, the problem of financial feasibility would arise even with the scenarios and strategies presented above; the adaptation measures discussed above, therefore, do not represent strategies that are universally applicable. Sources: Daschkeit and Schottes, 2002; Mai et al., 2004; Elsner et al., 2005

     There is still a considerable degree of catching up to do in order to integrate information on the potential impact of climate change systematically into implementation of coastal management systems. Despite sound scientific findings regarding the potential consequences of climate change, there has been scant political effort so far to devise adequate strategies for action.
     Against this background, the German Federal Government’s national strategy for integrated management of German coastal areas is laudable (Bundesregierung, 2006). This strategy takes into account the many different players involved and brings together the competing interests of protection and utilization of Germany’s coastal areas under a single, integrated concept. It certainly emphasizes climate change as a major component in the long-term orientation of precautionary planning at regional level. However, in view of the gravity of the anticipated consequences of climate change, it is necessary to improve the scientific basis for developing this strategy further. Measures aimed at adapting to the consequences of sea-level rise and extreme weather events will need to become the primary focus of future strategy.

3.4.1.4     Future challenges

Two issues relating to implementing adaptation strategies need to be emphasized here: the significance of proactive measures and the special challenges of implementing adaptation strategies in developing countries.

Early warning systems

Risk management plans encompass both proactive and reactive components of adaptation. Proactive components are particularly important for cost-effective adaptation design, because they prevent or at least reduce the chance of a risk translating into a disaster. This is true particularly with regard to sudden onset hazards. In the past, priority in financing adaptation strategies was given to reactive measures such as financing reconstruction of damaged infrastructure in the wake of a natural disaster (WCDR, 2005). What appears to be needed, therefore, is a reorientation of funding resources combined with a shift in priorities when choosing appropriate adaptation strategies. At the World Conference on Disaster Reduction (WCDR) in 2005 in the Japanese city of Kobe, it was decided that 10 per cent of funds hitherto used for ex-post measures in the aftermath of natural disasters should be diverted into preventive measures over the next ten years (WCDR, 2005; Münchener Rück, 2005a). The significance of proactive measures is underlined by the plan adopted by the WCDR to promote an International Early Warning Programme (IEWP). The IEWP is aimed at identifying and closing existing gaps as regards early warning (UN ISDR, 2005c). Key elements for improving early warning systems include developing national, integrated risk reduction strategies, capacity-building in the field of risk management and improving technical equipment and training. In addition, strategies are to be developed to improve communication of warnings to affected communities. Early warning therefore comprises a range of aspects, from technical capacity to preparatory measures at municipality level. To date, however, this linking of planning and precautionary measures with adequate response strategies has often been flawed. In future this deficit in existing systems is also to be eliminated. In order to achieve the goals set out in Kobe, international cooperation is required particularly in the area of data exchange, dissemination of warnings and developing institutional structures. At the present time there is a particularly urgent need to raise governments’ awareness of the problem and establish priorities for developing appropriate strategies of risk reduction.

Special challenges in developing countries

Climate change will have a major impact on developing countries in particular. These countries account for 97 per cent of fatalities from natural disasters (Freeman et al., 2003). Damage resulting from natural disasters is a considerable impediment to economic development in these countries. Adaptation therefore has particular significance for these regions. Technical know-how, appropriate institutions and especially financial resources are lacking, however, to enable the necessary measures to be put in place. There is broad consensus in the international community that support should be given to help developing countries cope with the impact of climate change. In Article 4, para. 3 of the Framework Convention on Climate Change, the Parties to the Convention commit themselves to provide financial and technical support to affected countries. In the context of the ‘Hyogo Framework for Action’, the 10-year programme of action adopted at the WCDR, this commitment was reiterated (WCDR, 2005). In addition, in recent years there has been increasing recognition of the fact that strategies for adapting to natural disasters and slow onset hazards need to be made an integral part of sustainable development cooperation (UNFCCC, 1992; UN ISDR, 2005a, b).

3.4.1.5     Financing adaptation measures in developing countries

To provide financial support to enable developing countries to adapt to the general consequences of climate change, a variety of international funding institutions offer financial transfers at multilateral level.

International Funds

In recent years, international funding bodies have been set up to promote adaptation measures in developing countries. In the context of the Framework Convention on Climate Change, three funds have been established that provide funding for adaptation to climate change in general, in other words not specifically related to oceans: the Special Climate Change Fund (SCCF), the Least Developed Countries Fund (LDCF) and the Adaptation Fund (GEF, 2005b).
     It is the explicit mandate of the SCCF to provide funding for adaptation projects and technology transfer. The fund was set up in 2003 to complement the Global Environment Facility (GEF) with a specific focus on climate change. By late 2004, the volume of funds in the hands of the SCCF in the form of voluntary contributions from OECD countries and other industrialized countries totalled US$34.7 million. SCCF has been in a position to provide effective support for projects since early 2005.
     The LDCF gives particular priority to providing support for developing countries to formulate and implement National Adaptation Programmes of Action (NAPA). NAPAs identify areas where action relating to adaptation is most needed. Of the US$32.5 million already contributed to the fund, US$11 million has already been disbursed for the formulation of NAPAs.
     The Adaptation Fund, lastly, was set up with the aim of implementing Article 12, para. 8 of the Kyoto Protocol. Its primary source of funds is a share in the proceeds of Clean Development Mechanism (CDM) project activities amounting to 2 per cent of the certified emission reductions issued for a project activity. Disbursal of payments from this fund is unlikely to begin before 2008, that is, before the start of the first commitment period under the Kyoto Protocol. While the revenue effect of this de facto taxation of prevention projects is welcome, its allocation effect must be viewed with considerable criticism.
In addition to the above, GEF also provides funds for projects under its Climate Change Focal Area. In this case, however, the focus is on prevention projects rather than adaptation projects.

Efficient use of development cooperation funds

As well as the above funds, international donors provide financial assistance to developing countries affected by natural disasters in the context of development cooperation. In recent years, for example, the share of funds made available by the World Bank for dealing with the consequences of natural disasters such as tropical cyclones has increased markedly, from 3 per cent to 8 per cent of the World Bank portfolio (Freeman et al., 2003). Financial resources are thus increasingly being earmarked for projects not aimed at fulfilling the original goal of promoting economic and social development.
     If the aim of international development cooperation is to support the development of adaptive capacity in developing countries, then assistance must focus to a greater extent on preventative strategies than has hitherto been the case, e.g., on developing early warning systems. A partial shift of this sort from aftercare to hazard prevention takes on added significance against the background of expected intensification of climate-induced extreme events. In order to prevent a loss of efficiency, development cooperation should be brought into line with the policies of the special adaptation funds described above.
     At the same time, while ironing out the issue of financing adaptation measures, it is important not to lose sight of the actual goal of development cooperation. Economic and social development in itself remains the best form of adaptation strategy, because it generally increases the adaptive capacity of a developing country and thereby reduces its vulnerability to the impacts of climate change (Schelling, 1992).

Complementary instruments: Prioritizing micro-insurance

The funding required for adaptation measures cannot be quantified in any robust manner due to a lack of even moderately reliable damage estimates (Section 3.2). It can nevertheless be assumed that the above-mentioned financial resources will not be sufficient and that it would consequently be sensible to secure funding for adaptation measures in the broadest possible manner. For this reason, new funding mechanisms should be considered alongside existing funding instruments and reallocation of currently available resources (WBGU, 2002).
     Another means is to promote micro-insurance in order to disperse the individual risk of hardship; in countries with low per capita incomes, this takes on added significance. Micro-insurance aims to provide insurance protection at extremely low premiums for households and small businesses with a low, and in some cases irregular, income and to increase available financial resources in the event of losses occurring. Micro-insurance, therefore, is not concerned with the national or international level, but is aimed at protecting individual assets (Münchener Rück, 2005b).
     Micro-insurance experience already exists in some areas where individual risks occur independently of each other, e.g. risks relating to illness or accidents (Brown and Churchill, 1999, 2000; Ahmed et al., 2005; Cohen et al., 2005). Case studies carried out in India, Kenya or Uganda show that life insurance and health insurance in particular are already being used successfully (Brown and Churchill, 1999, 2000; Athreye and Roth, 2005). Micro-insurance for risks relating to natural disasters, on the other hand, is still being piloted. Applying micro-insurance to natural disasters is particularly difficult because large numbers of people are usually affected and the individual risk of loss to local policyholders thus depends on the risk for all the others. As a result, demands on the insurer are very high in the event of loss occurring, and may even exceed the insurer’s capital stock. If the insurance provider opts for increasing his capital stock or reinsuring as a means of solving the problem, his capital costs increase and this is reflected in the price of the insurance policy. In these circumstances, many households and businesses with low incomes will ultimately forego private insurance altogether.
     In order to be able put a reasonably-priced and effective insurance product within their reach despite the difficulties, existing micro-insurance systems for independent risks could be extended to cover losses arising as a result of natural disasters. The costs of insurance cover are kept low by developing effective institutional capacities and ‘bundling’ policyholders in groups and municipalities. In addition, governments could make it compulsory to take out insurance against natural disasters. This would enable a large number of policyholders to be recruited swiftly and achieve a broad geographical distribution of policyholders, which would greatly reduce the problem of correlated risks of individual losses. The question of whether compulsory insurance of this sort would be a sensible option – especially in countries where social insurance systems are still inadequate – should be investigated in the context of future research.
     In order to ensure that providers of insurance for natural disasters operating at national or regional level are successful in the long term, it is important that they are linked to the international capital market. For example, against payment of a premium, reinsurance companies act as ‘insurers of the insurers’, assuming a proportion of the insurance provider’s risk. Risks are thus spread more broadly and insurers are freed from the risk of facing extremely high payouts.
     Micro-insurance programmes should be actively promoted by governments (public co-financing): alongside establishing the necessary legal framework, providing financial support might also be considered in the early stages, especially to develop the necessary institutional infrastructure, for example in the context of public-private partnerships and in cooperation with development organizations (Linnerooth-Bayer and Mechler, 2005).

3.4.2   The adoption of provisions governing loss of territory in international law

Adaptation strategies raise a number of legal issues as well. With steadily rising sea levels, it is likely that managed retreat will be the only option in many cases. In particular, national territories may well be lost completely or partially as a result of flooding, with people being forced to abandon settled areas. In terms of international law, this poses various challenges which relate, firstly, to the resettlement of the people displaced by sea-level rise, and secondly, the question of financial compensation in cases when states which are affected by the impacts of climate change-induced sea-level rise have not contributed significantly to its causes.

3.4.2.1   Reduction in the size of national territory

If a state’s territory shrinks as a result of sea-level rise, this does not have any specific implications in terms of international law, aside from the issue of compensation (see Section 3.4.2.4). According to the relevant provisions of international law, in such a scenario, the constituent national territory is simply reduced in size. In individual cases, however, it may be necessary to amend specific commitments arising under international law, primarily those relating to the territory which is now submerged. In general, the relevant provisions of international law supply satisfactory solutions to the legal problems which can be anticipated here. It must be borne in mind that a reduction in the size of a state’s national territory may result in a shift in the boundaries of its maritime jurisdiction as well, if the points used to position them have changed.

3.4.2.2   Submersion of (island) states

According to current knowledge, the survival of island states lying only a few metres above sea level is in acute jeopardy as a result of climate change-induced sea-level rise (CSD, 2004). These island states include the Maldives, an island group lying no more than 2 m above sea level, and the Tuvalu, Kiribati and Tonga island groups, which are located on coral reefs. These small island states, which are also developing countries (Small Island Developing States – SIDS), have formed a community of interest which is making its presence felt as a political alliance in the international negotiations on the United Nations Framework Convention on Climate Change (UNFCCC) (Burns, 1997; Neroni Slade, 2001). Admittedly, the SIDS (along with countries with low-lying coastal areas) have been given special consideration in the UNFCCC; for example, Article 4, para. 8 (a) and (b) calls for consideration to be given to actions, including funding, insurance and the transfer of technology, which may be necessary to meet the specific needs and concerns of these countries. However, this vague reference comprises the full extent of the special consideration of island states contained in the UNFCCC. Indeed, Article 4, para. 8 of the UNFCCC defines the specific needs of other categories of developing countries in such broad terms that almost any developing country Party could claim to be particularly vulnerable in some way. In other words, no specific rights for the island states can be derived from these provisions of the Convention. The island states are not mentioned specifically in the Kyoto Protocol. In the supplementary agreements adopted by the Parties to operationalize the Kyoto Protocol, notably the Marrakech Accords, the needs of the island states are emphasized repeatedly, but this has yet to result in the adoption of any institutional or other specific provisions.
     Other regional or global agreements, especially in the law of the sea, also fail to recognize, in any legally meaningful way, the status of island states as countries with special ecological or other problems. The same applies to the United Nations Convention on the Law of the Sea, even though islands play a key role in this Convention as a maritime geographical category of importance in determining maritime zones and related sovereign rights (Jesus, 2003).
     From the perspective of international law, the existence of a national territory is a constituent element of the state, which means that submersion of a state’s territory could result in its extinction. Nor does international law currently grant any entitlement to the allocation of any kind of ‘replacement territory’, although this would be possible in political terms. However, experience in the Middle East, not least, has shown that the creation of a state or new national territory may trigger considerable potential for conflict, especially given that hardly any unsettled territories are now.

3.4.2.3   Dealing with ‘refugees from sea-level rise’

If a state is submerged, its citizens become stateless. ‘Refugees from sea-level rise’ will probably seek refuge in neighbouring countries, perhaps greatly exceeding these countries’ absorption capacities. WBGU therefore considers that formal provisions are required to regulate the legal status of these people.
     WBGU recommends that the adoption of relevant provisions under international law be guided by the following principles. Basic provisions should establish the affected population’s right to regulated refuge/resettlement. This raises the question of the obligations which would thus arise for potential host countries, whereby a distinction must be made between the practical reception of the refugees and the covering of costs. From a humanitarian perspective, the best option is for refugees to be received by countries in the geographical vicinity of, or with specific links to, the submerged state. The refugees should have a say in choosing their new living environment; forced resettlement should be avoided as far as possible.      At the same time, however, an allocation formula should be developed in a process involving the wider international community, in order to ensure that individual host countries’ capacities are not overstretched. Fair and efficient burden-sharing requires that the costs of receiving the refugees be allocated according to the ‘polluter pays’ principle. The allocation formula should thus be guided by the principle of common but differentiated responsibility enshrined in international law. This means that the heaviest burden must be borne by those countries which are making the largest contribution to global greenhouse gas emissions and which also have the greatest financial resources at their disposal (Principle 7 of the Rio Declaration, Article 3, para. 1 and Article 4, para. 1 of the UNFCCC; Kellersmann, 2000; Stone, 2004). It is important to bear in mind that the issue of sea-level refugees is universal, for it arises not only when an individual state is submerged but also when major climate change-induced flooding and devastation occur in states which continue to exist.
     The development and application of the relevant legal provisions may prove to be problematical in practice, however. How can refugees who have lost their living environment as a result of climate change, making them dependent on assistance from others, be distinguished from other refugees? And how can the fundamental problem of causality be resolved? After all, hurricanes or extreme weather conditions which trigger refugee flows may not necessarily be caused by climate change but may simply be the result of the natural variability of the climate system (Section 3.1.2; Stone and Allen, 2005). Solutions to these problems must be found when formulating legal provisions governing the treatment of sea-level refugees. Against this background, WBGU recommends a significant increase in research in this area, especially the analysis and exploration of fair and effective burden-sharing systems.
     A further difficulty arising in this context is that ‘environmental refugees’ do not fit into any accepted category in international refugee and migration law (GCIM, 2005). According to the Convention relating to the Status of Refugees (Geneva Refugee Convention), the term ‘refugee’ only applies to persons persecuted for reasons of race, religion, nationality, [or] membership of a particular social group or political opinion. It does not create any specific obligations under international law for the treatment of ‘sea-level refugees’. In WBGU’s view, this gap in international refugee law must be closed. One option is to establish bilateral agreements, e.g. with neighbour states, or to adopt a multilateral agreement. This raises the question whether the existing conventions, especially the Refugee Convention, can be amended appropriately without renegotiating the definition of ‘refugee’ itself, or whether the conclusion of a separate convention would be more appropriate. In line with the non-refoulement principle, persecuted persons may not be deported to a country where they may be subjected to torture or inhumane treatment. By the same token, states should undertake not to return sea-level refugees to their country of origin if climate change has rendered the conditions of life in these countries unsustainable, i.e. if the living conditions are incompatible with human dignity or basic economic survival cannot be guaranteed. The scope of such a new norm must therefore extend beyond the specific problem of sea-level refugees to encompass other forms of environmentally related migration as well.

3.4.2.4   Compensation for loss of land

Compensation issues play a key role in relation to the loss of territory and the submersion of island states. A distinction must be made between various scenarios here.
     In cases when only the national level is affected, i.e. the damage is sustained by private individuals through the loss of their property or its value, or loss of income, national law applies; such cases are not relevant to this report. However, possible international conventions may have an impact on private actors if, for example, a state passes the responsibility for collecting the resources to cover international agreed compensation payments through taxes and levies to the private sector.
     What is relevant, however, is whether and to what extent the international community or other individual states have an obligation to pay compensation if a country sustains damage directly or indirectly as a result of sea-level rise. According to current international law and practice and prevailing opinion, no such obligation exists: even though the problem of rising sea levels is rarely caused by the affected island or coastal states but is primarily due to greenhouse gas emissions in the industrialized and newly industrializing countries, an obligation to pay reparations or damages does not arise under current international law. The background to this issue is the problem of the cumulative effects of certain types of conduct – a question which has yet to be satisfactorily resolved in international law – and the causal links, which are sometimes difficult to establish. As international law stands, the ban on causing major transboundary environmental injury, recognized in customary international law, thus does not apply (Epiney, 1995; Beyerlin, 2000; Wolfrum, 2000; Sands, 2003). Nonetheless, cause and effect have been established in many instances, and there is no doubt that climate changed-induced sea-level rise is presenting some developing countries with problems which they lack the financial resources to cope with unaided.
     Against this background, WBGU recommends the conclusion of an international convention which would oblige the industrialized countries in particular to guarantee adequate funding for an internationally administered compensation fund. Funding would be disbursed from this fund to countries particularly affected by rising sea levels. A country’s contribution commitments should be weighted according to the greenhouse gas emissions it produces, so that payments can be regarded as compensation for a country’s actual contribution to climate-related damage (Section 3.4.1.5). Once this compensation fund has been established as a means of providing assistance to the affected states, it could also take on a role in burden-sharing within the international community, e.g. managing the reception of refugees fleeing from sea-level rise and the payments made to host countries (Section 3.4.2.3).
     Utilizing the mechanisms for the transfer of financial resources and technology established for the climate regime might also appear, at first sight, to be a viable option. For example, the Mauritius Strategy for the Further Implementation of the Programme of Action for the Sustainable Development of Small Island Developing States (para. 78(a)) posits, in the context of climate change adaptation and sea-level rise, that strategies can be developed with support from the Least Developed Countries Fund and the Special Climate Change Fund set up within the framework of the United Nations Framework Convention on Climate Change. The key objection here is that such support cannot be regarded as genuine compensation for climate-induced damage. A further possibility is for the United Nations Compensation Commission to take action in this area; for example, it recently adjudicated compensation for environmental damage caused in the 1990–1991 Gulf War (Sands, 2003). However, this particular instrument is not precise enough to pay targeted compensation for the damage caused by climate changed-induced sea-level rise. An existing body could at best be entrusted with the task of administering the separate compensation scheme outlined above.

3.5   Research recommendations

Hurricane formation and strength
The links between hurricane activity and global warming need to be researched more thoroughly, both through further analysis of data gathered from past developments and by modelling the future development of the hurricane climate, including potential threats to areas not affected previously (South America, southern Europe).

Extent and rate of sea-level rise
The greatest uncertainty surrounding future sea-level rise concerns the behaviour of continental ice sheets in Greenland and Antarctica. To reduce this uncertainty, there is a need to gain an improved understanding of ice dynamics; major progress is needed in continental ice modelling. These activities include researching the stability of the ice shelves as well as their interplay with continental ice. Further uncertainties surround ocean dynamics, especially the intensity of ocean mixing. Such dynamics greatly influence sea levels. There is a need to improve their characterization within global climate models.

Global potential for damage caused by sea-level rise
The issue of ‘dangerous sea-level rise’ forms a sub-set of the wider question of ‘dangerous climate change’ and must be answered quantitatively if possible. To do so, there is a need to aggregate globally the health, socio-economic and ecological consequences associated with various scenarios (x metres rise in y years). Present assessments are not robust in this respect. They must be replaced by a new generation of impact analyses. This could produce a more precise definition of the provisional absolute guard rail proposed by WBGU (maximum of 1 m sea-level rise).

Vulnerability of coastal megacities in developing countries
Climate change and urbanization are dominant trends of global change. The interplay of the two trends in the major coastal cities of the developing world could cause an almost unmanageable situation, particularly if the arsenal of responses is limited by social, economic and institutional deficits. There is an urgent need to conduct interdisciplinary studies in order to assess the severity of the problems for particularly critical megacities such as Lagos, Mumbai or Havana.

Regional portfolio strategies for coastal management
The dramatic geophysical impacts of climate change upon coastal zones (which will arise even if vigorous measures are taken to reduce global greenhouse gas emissions) mean that the traditional approaches to coastal management need to be revisited. There is a particular need to determine the priority given to the various strategic elements of protection, managed retreat and accommodation. To be able to conduct such an assessment, types of cost-benefit analysis must be developed that take account of the novel potential for damage. At present such analyses have only been carried out for limited sections of coasts, e.g. in Great Britain. There is an urgent need to conduct an integrated re-appraisal of robust and effective portfolio strategies for German coasts.

‘Sea-level refugeese’: Legal and institutional aspects
The threats presented to coastal regions and the potential destruction of entire state territories by climate-induced sea-level rise generate a novel migration problem whose legal dimensions have yet to be explored. There is a particular need for research on how to shape provisions under international law with respect to the reception of ‘sea-level refugees’, the payment of compensation, and burden-sharing in line with the ‘polluter pays’ principle. To resolve the legal problems, it is also very important to make progress in the scientific attribution of damage or territorial loss arising as a consequence of human-induced climate change. There is also a need to conduct operative appraisals, for instance to evaluate the capacity of the existing United Nations institutions to cope with refugee flows, especially given that needs will presumably grow exponentially in the future.