Responding to Rising Seas

Chapter 1. Rising risks in coastal zones

Abstract

This chapter provides an overview of how coastal zones are facing growing risks from sea-level rise, the economic costs associated with this increasing risk and the implications for policy making. The chapter reviews the current scientific understanding of sea-level rise and coastal flood hazards. It then discusses the costs and benefits of adaptation under future sea-level rise, particularly focusing on coastal protection measures. It provides analysis of coastal adaptation from a robust decision-making perspective. Finally, the policy implications of current knowledge on coastal adaptation costs and benefits are discussed, along with priorities for future research to support coastal adaptation policy.

This chapter was written by Alexander Bisaro, Jochen Hinkel and Daniel Lincke, Global Climate Forum, Berlin and Division of Resource Economics, Humboldt University, Berlin.

1.1. Evolution of risks in coastal zones

Climate change-induced sea-level rise (SLR) will affect the world’s coasts by increasing flood and erosion risks, and potentially fully inundating some areas. As risks increase, so will the associated economic and human costs from extreme events and slow-onset changes. This will strain society’s capacity to maintain an acceptable level of risk at reasonable cost in coastal zones.

The core challenge of coastal adaptation is that decisions need to accommodate ongoing change, which is subject to deep uncertainty, in an area with contested stakeholder priorities. Coastal adaptation choices involve difficult trade-offs between different objectives and interests, and are constrained by existing institutional arrangements and the legacy of past decisions.

This report uses the Intergovernmental Panel on Climate Change (IPCC)’s 5th Assessment Report framework to describe risk.1 Risks are a function of the range of potential outcomes and the associated likelihoods of those outcomes materialising in a given period. In this context, risks arise from the interaction between hazards, exposure and vulnerability:

• Hazards are the potential occurrence of a physical event or trend (flooding, erosion) that may cause loss of life, injury, as well as damage and loss to property, infrastructure, livelihoods, service provision, ecosystems and environmental resources.

• Exposure refers to the presence of people, infrastructure, housing and other tangible human assets in hazard-prone areas. A measure of exposure can include the number of people or types of assets in a coastal flood zone.

• Vulnerability is the degree to which natural or social systems are susceptible to, and unable to cope with, exposure to hazard.

Figure 1.1. Risk-based conceptual framework

Source: Intergovernmental Panel on Climate Change (2014[1]), “Glossary, acronyms and chemical symbols”, https://doi.org/10.1017/CBO9781107415416.023.

Risks from SLR include high-probability, low-consequence events (e.g. nuisance flooding) and high-probability, high-consequence events (e.g. storm surges). There is robust evidence that storm surges are already penetrating farther inland than a few decades ago, with adverse impacts on communities and coastal ecosystems (Hoegh-Guldberg et al., 2018[2]).

1.1.1. Climate change and sea-level rise

Climate change-induced sea-level rise increases coastal risks by raising the likelihood of flooding events, and inducing land loss through inundation. Generally, sea-level increases are driven by changes in global mean temperature, which are in turn driven by atmospheric greenhouse gas concentrations. Projecting future sea levels thus requires developing SLR scenarios based on different greenhouse gas concentration pathways. Representative concentration pathways (RCP) cover a wide range of such potential future concentration pathways out to 2100.

Figure 1.2 illustrates a range of SLR scenarios based on recent global studies of sea-level rise impacts (Hinkel et al., 2014[3]). All SLR values are shown with respect to mean sea‑level in the 1985-2005 reference period. SLR projections in the range of 0.3 m to 1.3 m over the 21st century are based on scenarios that span three representative concentration pathways (RCP2.6, RCP4.5 and RCP8.5); four general circulation models (GCMs) (HadGEM2-ES, IPSLCM5A-LR, MIROC-ESM-CHEM and NorESM1-M); and a low, medium and high land-ice scenario (Hinkel et al., 2014[3]; Lincke and Hinkel, 2018[4]). However, future SLR outside of this range is also possible. For instance, Figure 1.2 also includes a scenario of up to 2.0 m of SLR by 2100, based on a high-end SLR scenario (H++) (Nicholls et al., 2013[5]). Such a high-end scenario can be located in the low‑probability, high-impact tail of possible 21st century SLR.

Figure 1.2. Sea-level rise scenarios to 2100

Source: Lincke, D. and J. Hinkel (2018[4]), “Economically robust protection against 21st century sea-level rise”, http://dx.doi.org/10.1016/J.GLOENVCHA.2018.05.003.

Different processes drive increases in global mean sea level, with the four main ones being: oceanic thermal expansion (Taylor et al., 2012[6]), melting from glaciers (Marzeion, Jarosch and Hofer, 2012[7]), the Greenland ice sheet (Fettweis et al., 2013[8]) and Antarctic ice sheets (Levermann et al., 2014[9]). Analysis comparing contributions of these components, using different GCMs and concentration pathways, found that the largest single contribution to global mean SLR comes from oceanic thermal expansion. Mountain glaciers and ice caps also contribute substantially, but less than thermal expansion. However, if considered as a whole, the melting of land ice is projected to contribute most to future sea-level rise (Hinkel et al., 2014[3]).

Gravitational and rotational effects from changes in ice masses (Farrell and Clark, 2007[10]) and ocean circulation (Hinkel et al., 2014[3]) additionally influence the regional distribution of sea-level rise. These effects lead to sea-level rise being higher in the tropics than at high latitudes (Perrette et al., 2013[11]).

Uncertainties around the contribution of these processes to SLR are the greatest regarding the melting of ice sheets. The contribution of the Antarctic ice sheet is the most uncertain, and gives rise to a long-tailed risk of very high sea-level rise. Recent studies find that 5th percentile of Antarctic ice sheet contribution is around 2 cm, and the 50th percentile is 10 cm. The 95th percentile of the Antarctic ice sheet contribution is, however, as high as 41 cm in the RCP 8.5 scenario (Hinkel et al., 2014[3]). A further uncertainty arises from model differences. For example, Hinkel et al. (2014[3]) find that median sea-level projections can differ by up to 20 cm by 2100 depending on the GCM used. These various uncertainties arising from physical processes are captured in the probability distributions illustrated in Figure 1.2.

Changes in local relative sea level can, however, vary significantly from changes in global mean sea level. Biophysical and geological processes such as vertical land movement, changes in ocean circulation patterns or natural glacial-isostatic adjustment influence local relative sea level. For instance, on the north coast of Finland and Sweden, the land is currently rising faster than the sea due to post-glacial uplift. Further, local relative sea level is also influenced by human activities, such as extraction of groundwater or oil, mining and changes in sediment supply from rivers due to dam building. In some areas, the contribution of these activities to SLR can be an order of magnitude higher than that from global climate change (Ericson et al., 2006[12]).

Densely populated deltas, which globally have a population of more than 500 million, are particularly susceptible to such human-induced subsidence due to their geological setting (Woodroffe et al., 2006[13]). Many of the world’s coastal megacities are also situated in deltas and several metres of human-induced subsidence have been observed during the 20th century (Nicholls, 1995[14]; World Bank, 2010[15]). For instance, in Jakarta, observed subsidence rates over the last three decades have been between 3 cm and 10 cm per year (Abidin et al., 2015[16]). Rural areas are also susceptible, as local subsidence rates of 250 mm per year have been observed in areas where intensive aquaculture activities require groundwater pumping to freshen fishponds (Higgins et al., 2013[17]).

Human-induced local relative changes in sea level can thus be a major source of uncertainty about the risks faced by coastal areas (Hinkel et al., 2014[3]). As such, geographically specific modelling is required to understand the potential impacts in a given area. While global models are available for natural glacial-isostatic adjustment (Douglas, Kearney and Leatherman, 2000[18]), information on annual rates of human-induced subsidence is extremely limited and both the drivers and responses are localised, making modelling extremely difficult (Hanson et al., 2011[19]).

Climate change-induced SLR beyond 2100 will, however, continue for thousands to tens of thousands years, even if greenhouse gas concentrations are stabilised during the 21st century (Levermann et al., 2013[20]). This has been termed commitment to SLR (Church et al., 2001[21]). However, the rate and magnitude of SLR over these long time frames is deeply uncertain and subject to some controversy. For example, IPCC AR5 estimates SLR in 2500 between 1.5 m and 6.6 m under high-concentration scenarios (> 700 ppm CO₂eq). In contrast, (Clark et al., 2016[22]) estimate 25-52 metres within the next 10 000 years under a stipulated equilibrium climate sensitivity of 3.5°C. Levermann et al. (2013[20]) estimate the committed SLR with rising temperature as approximately 2.3 m/°Celsius. Thus, the extent of long-term SLR could range from tens of centimetres up to several metres. Alternatively, a world with SLR much beyond present experience is also possible. These alternatives present radically different situations in future centuries.

1.1.2. Evidence of the reduced resilience of ecosystems, and the link with human activity

Estuarine and coastal ecosystems are some of the most heavily used and threatened natural systems globally, with significant deterioration due to human activities. For example, 50% of salt marshes, 35% of mangroves, 30% of coral reefs and 29% of seagrasses have been either lost or are degraded worldwide (Barbier et al., 2011[23]). Murray et al. (2014[24]) found that 28% of tidal flats bordering the Yellow Sea disappeared between 1980 and the late 2000s, at a rate of 1.2% annually.

Coastal development processes, such as land reclamation or hard coastal protection measures, can degrade coastal ecosystems (Hoegh-Guldberg et al., 2018[2]). Indeed, land reclamation has an extensive history in areas with dense populations and a shortage of land, e.g. southern North Sea countries and the People’s Republic of China (hereafter “China”) (Bisaro and Hinkel, 2018[25]). Globally, total land area gained from the sea in the last 30 years is approximately 33 700 km² (about 50% more than has been lost), with most land reclamation areas occurring in places like Dubai, Singapore and China (Donchyts et al., 2016[26]; Ma et al., 2014[27]). Wetlands loss or degradation due to land reclamation for urban or industrial uses reduces water storage areas. In such cases, high waters from storm surges can reach higher velocities and heights when forced into remaining channels (Wong et al., 2014[28]). Further, land reclamation may disrupt coastal ecosystems, negatively affecting coral reefs, mangroves or seagrass beds (Li et al., 2013[29]), while also disrupting natural morphological processes, leading to coastal erosion and increased flood risk (Murray et al., 2014[24]). Finally, as discussed above, coastal development often leads to increased groundwater extraction, causing land subsidence and increasing coastal risk (Wong et al., 2014[28]).

The loss of biodiversity, ecosystem functions and coastal vegetation has contributed to decreased coastal protection from flooding and storm events (Liquete et al., 2013[30]). Wetlands, mangroves, near-shore coral reefs and dunes can all reduce storm surges and stabilise shorelines (Spalding et al., 2014[31]). This protection has significant value; for example, globally, coral reefs are estimated to protect over 100 million people from wave-induced flooding. Further, it has been estimated that annual expected flood damage reduction from coral reefs exceed USD 400 million for Cuba, Indonesia, Malaysia, Mexico and Philippines alone (Beck et al., 2018[32]). In addition to coastal protection, healthy coastal ecosystems provide a suite of other valuable benefits (e.g. ecosystem services) on which humans depend. These include providing nursery habitat for fish and other marine species, water filtration, carbon storage, and opportunities for recreation (Mehvar et al., 2018[33]).

SLR itself poses a threat to coastal ecosystems, and (Spencer et al., 2016[34]) estimate that up to 78% of the global wetland area could be lost under a high SLR scenario. As coastal ecosystems change under SLR, the benefits that they provide in the form of ecosystem services are likely to decline and negatively impact the people who depend on them (Mehvar et al., 2018[33]).

1.2. The economic cost of sea-level rise

As risks from sea-level rise increase, so too will the associated economic and human costs from extreme events and slow-onset changes. This section uses new modelling to provide economic estimates of the impacts of rising sea levels on coastal assets, as well as the costs of adapting through protection.

Assessments of the costs of SLR must consider flood risk and adaptation costs, i.e. the costs of implementing protection, accommodation or retreat measures (further detailed in Chapter 2). The costs and benefits of adaptation have been assessed on a country level by a first generation of studies considering the gradual loss of land as the main impact of sea-level rise (Fankhauser, 1995[35]; Nicholls, Tol and Vafeidis, 2008[36]; Sugiyama, Nicholls and Vafeidis, 2008[37]; Yohe, Neumann and Ameden, 1995[38]). As mentioned above, these studies generally disregarded the adverse effects of extreme sea‑level events that are rising with mean sea levels, and which manifest even before land is lost permanently (Wong et al., 2014[28]). A second generation of studies have addressed this limitation, considering the expected damage caused by extreme sea levels as well as refining the scale of analysis to subnational levels based on segmentations of coastline into units (Diaz, 2016[39]; Hinkel et al., 2014[40]; Nicholls et al., 2011[41]; Vafeidis, 2008[42]).

The DIVA model, a global coastal SLR impact model, offers one of the most comprehensive and advanced representations of relevant processes for assessing coastal flood risk and adaptation costs, and detailed global scale representation of the coastal zone based on 12 148 coastline segments defined in the DINAS-COAST database (Vafeidis et al., 2008[43]). By using DIVA, it is possible to assess the costs associated with SLR under different adaptation scenarios for the 21st century. Flood damages are calculated by combining elevation-based population exposure with flood depths caused by extreme events and applying a depth-damage function. Expected annual flood damages are computed as the mathematical expectation of damages based on extreme event distributions, given protection levels (Hinkel et al., 2014[3]). Within DIVA, adaptation costs are assessed in terms of dike investment and additional maintenance costs.

Assessing the impacts of increased coastal flooding on population and assets requires a comprehensive sampling of state-of-the-art socio-economic and sea-level rise scenarios. Thus, a range of scenarios are applied in order to address uncertainties regarding the development of future coastal risk described above. For socio-economic scenarios, five population and gross domestic product (GDP) growth scenarios based on the shared socio‑economic pathways (SSPs) (Box 1.1) provide such a sampling.

Box 1.1. Shared socio-economic pathways and future coastal risk

Shared socio-economic pathways (SSPs) are widely used in climate impact assessment to describe future socio-economic development scenarios in a coherent and consistent manner (IIASA, 2012[44]; Lincke and Hinkel, 2018[4])The DIVA results described in this section have been obtained using SSPs 1-5. These can be described as follows:

• SSP 1 (Sustainability) reflects a world progressing towards sustainability with reduced resource intensity and fossil fuel dependency. SSP 1 attains the highest GDP and lowest population numbers.

• SSP 2 (Middle of the Road) reflects a world with medium assumptions.

• SSP 3 (Fragmentation) reflects a world fragmented into poor regions with low resource intensity and moderately wealthy regions with a high fossil fuel dependency. GDP is lowest and population highest in SSP 3.

• SSP 4 (Inequality) reflects a highly unequal world both within and across countries. GDP and population follow a similar, but lower, trend compared to SSP 3.

• SSP 5 (Conventional Development) reflects a world oriented toward rapid, equitable development that is dependent on fossil fuels.

Table 1.1. Global population and GDP in 2050 and 2100 under different shared socio-economic pathways

	Population (millions)		GDP (billion USD/yr)
	2050	2100	2050	2100
SSP 1	8 400	7 200	295 000	771 000
SSP 2	9 300	9 800	260 000	685 000
SSP 3	10 300	14 100	334 000	667 000
SSP 4	9 400	11 800	242 000	462 000
SSP 5	8 500	7 700	348 000	1 207 000

Source: Hinkel, J. et al. (2014[3]), “Coastal flood damage and adaptation costs under 21st century sea-level rise“, https://doi.org/10.1073/pnas.1222469111.

Different adaptation scenarios, which come on top of the main SSP storylines, can be considered in DIVA, and a baseline adaptation scenario is required for assessing sea-level rise impacts and adaptation. Generally, most approaches in the literature have considered a “no adaptation” case where coastal defences are not upgraded while sea levels rise, and socio-economic development in the flood plain continues. In such a constant protection strategy, dikes remain at their current height, so flood risk increases with time as relative sea level rises. In an enhanced protection strategy, dikes are raised following both socio‑economic development and relative sea-level rise (Hinkel et al., 2014[3]).

Figure 1.3 shows the coastal flooding costs for SSP 3 and SSP 5 and low and high-end RCPs 2.6 and 8.5 respectively, as analysed by Hinkel et al. (2014[3]). SSP 3 and SSP 5 are illustrative because they represent the low and high extremes respectively of annual flood costs over the 21st century. SSP 3 represents the second-lowest GDP of the SSPs (after SSP 4), distributed to most people, and thus coastal exposure is the lowest under SSP 3. Figure 1.3 illustrates that, without adaptation, flood damage costs will be very high by the end of the century. The median outcome for high-end SLR (1.3 m in RCP8.5) being approximately USD 50 trillion annually or ca. 4% of world GDP annually. Adaptation, through enhanced protection, can reduce these costs by two to three orders of magnitude, showing substantial benefits across all combinations of scenarios. Thus, one implication of the analysis is that for large parts of the world, coastal protection is economically attractive regardless of how SLR and socio-economic development proceed (see Section 1.4).

Figure 1.3. Global annual flood costs for different socio-economic and climate scenarios with and without adaptation

Note: The solid lines represent the median and the shaded area represents the range from the 5th to 95th percentile for a given scenario combination.

Source: Hinkel; J. et al. (2014[40]), “Coastal flood damage and adaptation costs under 21st century sea-level rise”, https://doi.org/10.1073/pnas.1222469111.

At the global level, cumulative residual flood damages of USD (2005) 0.3 trillion to USD (2005) 3.9 trillion for the 21st century are reported by Hinkel et al. (2014[3]). Considering high-end sea-level rise increases to this range, as Lincke and Hinkel (2018[4]) report, residual damage costs of USD (2014) 1.7 trillion to USD (2014) 5.5 trillion (undiscounted) over the 21st century. Higher damage costs (for both the low- and high‑end of the range) come from the fact that Lincke and Hinkel (2018[4]) consider SLR scenarios up to 2.0 m, while the previous study only considered them up to 1.3 m. For OECD countries only, by 2100, residual flood damages range from USD (2005) 2.5 billion to USD (2005) 29.8 billion. While still significant, the smaller share of overall global damages indicated for OECD countries by these numbers represent the relatively greater ability of OECD countries to invest in coastal protection.

Figure 1.4 shows the global costs of adaptation under an enhanced coastal protection strategy. Generally, protection costs increase significantly under high SLR scenarios regardless of socio-economic development. Further, protection costs are the highest under SSP 5, which represents a rich fossil-intensive world, as growing wealth leads to increasing exposure of assets, and thus increased protection costs. A range of USD (2005) 1.9 trillion to USD (2005) 4.2 trillion for protection considering SLR scenarios up to 1.3 m is reported by Hinkel et al. (2014[3]) over the 21st century. Including high-end SLR scenarios (up to 2.0 m) increases the high costs to USD (2014) 7.8 trillion (not discounted). These results are of the same order of magnitude as those reported in earlier global studies. For example, (Tol, 2002[45]) reports protection costs of USD (1995) 0.6 trillion to USD (1995) 1.06 trillion for 1 metre of SLR, excluding maintenance cost.

Figure 1.4. Global annual costs of adaptation under an enhanced coastal protection strategy for different socio-economic and climate scenarios

Source: Hinkel, J. et al. (2014[40]), “Coastal flood damage and adaptation costs under 21st century sea-level rise”, https://doi.org/10.1073/pnas.1222469111.

The relative costs of SLR are another important consideration because this indicates how significant SLR costs will be for a specific country or region. Relative SLR costs can be defined as the present value of protection and residual damage cost as a percentage of present value of GDP over the 21st century (Lincke and Hinkel, 2018[4]). Relative costs of SLR, provided an optimal protection strategy is pursued, represent a small proportion of GDP at the global level, but will be a significant share of GDP for some individual countries. Globally, under an optimal adaptation strategy, the relative costs of SLR lie between 0.02% of global GDP under the best-case scenario (0.3 m global mean SLR, SSP 5 and not discounted) and 0.07% of global GDP under the worst-case scenario (2.0 m global mean SLR, SSP 3 and 6% discount rate). While generally OECD countries do not experience high relative costs of SLR, there are some countries for which the relative cost of SLR exceeds 1% under the worst-case scenario combination. These are Iceland (2.3%), Korea (1.8%) and Norway (1.1%). Globally, small islands in particular will experience high relative costs of SLR, including the risk of inundation. There are in total 41 countries for which the relative cost of SLR exceeds 1% of GDP under the worst-case scenario combination.

Figure 1.5 shows the global number of people flooded under different socio-economic and climate scenarios under constant protection (“no adaptation”) and enhanced protection (“adaptation”) scenarios. Under constant levels of flood protection, the number of people flooded will grow throughout the century across all socio-economic scenarios. This is despite the fact that SSP 1 and SSP 5 project decreasing global population from 2050 onwards (Hinkel et al., 2014[3]). The expected annual number of people flooded is the highest under SSP 3 and the lowest under SSP 1, reflecting the highest and lowest population numbers under these scenarios (see Table 1.1).

Figure 1.5. Global annual number of people flooded under constant protection

Source: Hinkel, J. et al. (2014[40]), “Coastal flood damage and adaptation costs under 21st century sea-level rise”, https://doi.org/10.1073/pnas.1222469111.

In an enhanced protection scenario, the number of people flooded actually falls over the course of the century, as more regions become rich enough to build dikes. Further, the influence of socio-economic development on the number of people flooded is smaller as compared to under constant protection. An exception is the extreme scenario SSP 3, under which population grows fastest, but GDP and hence dike height grow the slowest.

While the modelled relative costs of SLR could represent a small proportion of GDP at the global level if an optimal protection strategy is pursued, there are other elements that should be considered. First, these costs will be unequally distributed, falling particularly heavily in some areas, which may not be well-equipped to adapt. Second, these cost estimates assume that an economically “optimal” protection strategy will be followed: the challenges of achieving this are explored in Chapter 2. Lastly, not all of the potential costs, including non-market impacts, are captured in this model.

1.3. Robust coastal adaptation to 21st century sea-level rise

The global studies reviewed above have applied a range of SLR and socio-economic scenarios. They have, however, all taken the general approach of assessing SLR impacts and thus adaptation decisions within a given SLR and socio-economic scenario. This is indeed appropriate for gaining understanding of the range of possible sea-level rise impacts, and understanding adaptation costs and benefits within a given scenario. However, such an approach does not reflect the actual decision framing for coastal planners at national to subnational levels because such decisions need to consider all scenarios.

The need for decision making to account for all different scenarios is acknowledged by local coastal adaptation studies and substantial literature on coastal adaptation decision making under “deep uncertainty” has emerged. Representative approaches include robust decision making (Lempert and Schlesinger, 2001[46]) and adaptation pathways analysis (Haasnoot et al., 2013[47]). These approaches are characterised by finding options that are robust in the sense that they satisfy given criteria, e.g. a flood safety level, for a sample of scenarios covering all the relevant uncertainties. Indeed, the Thames Estuary 2100 study provides a prominent example of the latter approach further examples are discussed in Chapter 3.

Such an approach to “deep uncertainty” can also be applied globally, and a first study does just this, applying the DIVA model introduced above (Lincke and Hinkel, 2018[4]). Based on the five scenarios of 21st century global mean SLR from 0.3-2.0 m, introduced above, and the five SSPs (see Box 1.1), the study assesses for which parts of the global coastline coastal protection is economically robust. Further details on this method are described in Box 1.2.

Box 1.2. Applying “deep uncertainty” to a global sea-level rise model

The 2018 study by Lincke and Hinkel is based on the five scenarios of 21st century global mean sea-level rise (SLR) from 0.3 m to 2.0 m, and the five shared socio-economic pathways (SSPs). The study assesses for which parts of the global coastline coastal protection is economically robust by considering all 25 combinations of SLR and socio-economic development, in order to account for the whole uncertainty space spanned by these scenarios. However, some combinations of SLR and socio-economic development are less likely to occur (O’Neill et al., 2014[48]). The study also considers a range of five discount rates from 0.0% to 6.0%. Discount rates represent a major uncertainty dimension in flood risk management decisions, particularly over the long term (Hall and Solomatine, 2008[49]; Lempert and Schlesinger, 2001[46]), though they have not been addressed in global studies on sea-level rise impacts.

The optimal protection level for each coastline segment under each scenario combination is determined by minimising the net present value of the sum of protection cost streams and residual damage costs streams (avoided flood damage). Coastal segments are defined as robust to protect when protection is economically desirable (i.e. protection produces a positive net present value) under all scenarios

The principle result is the level of robustness of the decision to protect across all possible combinations of different scenarios for each coastal segment. The study also analyses the length of coastline and coastal plain exposure (area, population, assets) for which it is robust to protect, i.e. where protection is desirable for all scenario combinations.

Source: Lincke, D. and J. Hinkel (2018[4]), “Economically robust protection against 21^st century sea-level rise”, http://dx.doi.org/10.1016/J.GLOENVCHA.2018.05.003.

The results show that coastal protection is economically robust across all scenario combinations for 13% of the world’s coastline. These coastlines account for 90% of the global coastal population and 96% of global assets situated in the 1-in-100-year event floodplain in 2015. Conversely, it is robust not to protect 65% of the world’s coastline, which corresponds to a small fraction of global coastal floodplain population (0.2%) and assets (0.2%). For the remaining 22% of the world’s coastline, the optimal adaptation strategy varies across scenarios.

Most of the locations for which protection is robust are located on the east coast of the United States, and in Europe. In Asia, China, Korea and Japan also have coastline for which it is robust to protect (Figure 1.6). The reason for this is that these areas have high levels of coastal urbanisation and are located in countries with high levels of existing protection standards. Further, large cities in Australia are also robust to protect due to their high population densities. Conversely, it is robust not to protect most of the coasts of countries with long and uninhabited coastlines such as in Australia, Canada, Chile and Norway.

Figure 1.6. Economic robustness of coastal protection globally

Source: Lincke, D. and J. Hinkel (2018[4]), “Economically robust protection against 21st century sea-level rise”, https://doi.org/10.1016/j.gloenvcha.2018.05.003.

Generally, for countries with a very short but densely populated coastline, it may be robust to protect the entire coast. Globally, there are 18 countries for which this modelling suggests that it would be robust to protect the entire coastline. Two of those are OECD countries (Belgium and Poland). Conversely, there are 30 countries for which it is economically robust not to protect any part of the coast.

Considering the different individual scenario combinations, one can observe the share of protected coast grows with higher sea-level rise, higher GDP and lower discount rates. Hence, the biggest share of protected coast is obtained under the highest SLR scenario (i.e. H++ scenario), the wealthiest socio-economic scenario and a zero discount rate.

Discount rates are clearly a significant factor influencing the level of robustness of protection, in that lower discount rates make it more robust to protect. Protection costs largely occur as upfront investments near the beginning of the century, while damage costs are the greatest at the end of the century due to rising sea levels. High discount rates therefore lower the present value of damage costs more than the present value of protection costs, and decrease the economic attractiveness of protection.

Under higher levels of socio-economic development and for higher SLR, it is economically robust to protect more of the coast. Damage costs are higher for socio‑economic development scenarios representing a rich world than for scenarios representing a poor world, as more assets are exposed in the rich-world scenarios. At the same, time, protection costs are not influenced by socio-economic development. It follows then that a larger share of coast will be protected in scenarios representing a rich world because more flood damage can be avoided by protecting is such scenarios, compared to poor-world scenarios. Relatedly, in high SLR scenarios, a larger share of the coast is protected. This is because SLR increases damage costs faster than protection costs. That is, protection costs grow linearly with SLR, while damage costs grow super-linearly with SLR.

Finally, comparing these results to previous studies, it is important to note that the share of protected coastline under the robust cost-benefit adaptation strategy explored by Lincke and Hinkel (2018[4]) is much lower than the ones found in previous studies under alternative protection strategies. For instance, Fankhauser (1995[35]) found that about 80% of the open coast, 98% of harbours and 99% of cities should be protected under a strategy that minimises 21st century costs of land loss, forced migration and protection for 0.2‑2.0 m of SLR until 2100. More recently, (Nicholls, Tol and Vafeidis, 2008[50]) provide a global estimate reporting 50-85% of protected coast under 1.0-6.0 m SLR until 2130. In contrast, Lincke and Hinkel (2018[4]) find only 15% of the US coast is protected under all scenario combinations while 66% are not protected under any scenario combination. The earlier studies report much higher lengths of protected coast than Lincke and Hinkel (2018[4]) because they do not provide subnational resolution of exposed people and assets. When protection decisions are evaluated for whole countries or regions, many parts of the coast are protected that would not be protected in approaches with more coastal segmentation, because local differences in exposure are averaged away.

1.4. Implications for future research and policy

To put these results into context, it should again be emphasised that the DIVA studies (Hinkel et al., 2014[3]; Lincke and Hinkel, 2018[4]), similar to other global studies reviewed above, focus on coastal protection and do not explore other forms of adaptation. Future assessment of coastal impacts at all levels should also explore other adaptation options, including accommodate and retreat measures. While including these measures is unlikely to significantly alter the global picture, locally it may be economically more efficient to retreat from the coast. For instance, the countries found by Lincke and Hinkel (2018[4]) to have high relative SLR cost all have sparse, but mainly coastal, population. In such sparsely populated places, though infrastructure and people might be concentrated on the coast, coastal planners and decision makers should explore alternative adaptation options that may be economically more efficient than protection (see Chapter 2).

In particular, the global results shown in Figure 1.6, only consider hard protection, but other adaptation options may be more cost-effective, and better maintain ecosystem health in any given area. These results should not be seen as advocating protection in only those areas where it is economically viable. Rather, at a highly aggregated level, these results show that protection is viable across a range of scenarios in various locations. This provides an entry point for national coastal decision makers and planners, who then must decide on adaptation measures, acknowledging that the timing of adaptation, and adaptation options that are flexible (i.e. which can be reversed or extended), are key to achieving efficient adaptation, given the uncertainties associated with future coastal risk (see Chapter 3).

Turning to policy implications of the above analysis, there are two overarching points to consider. First, the results showing that it is economically robust to protect the vast majority of the world’s assets in the coastal zone suggest that the world is likely to see bifurcating coastal futures. On the one hand, the large majority of coastal inhabitants live in densely populated urban coastal areas, and are likely to continue to protect themselves even under high-end sea-level rise due to the high cost-benefit ratios of coastal protection in these areas. Residual risks will remain with possible catastrophic consequences in the case of dike failure, but these can also be reduced by building stronger and wider defences (De Bruijn, Klijn and Knoeff, 2013[51]). A key point in this regard is that the SLR literature needs to account for adaptation to give a realistic picture of the coastal future. Where high concentrations of assets and people are present at the coast, we are likely to see increased protection, rather than large-scale damages. It is also worth mentioning that it will be important to consider alternatives to hard protection, such as nature-based solutions, in order to avoid escalating damage to ecosystems or loss of amenity value through hard protection alone.

A second policy implication concerns financing coastal adaptation. According to the global studies reviewed, it is attractive from an economic point of view to protect around 90% of the world’s coastal population (Lincke and Hinkel, 2018[4]). In practice, however, financing and implementing coastal protection gives rise to a number of challenges due to the public-good nature of coastal protection and its benefits being stochastic, long-term and distributed across diverse beneficiaries (Bisaro and Hinkel, 2018[25]; Bisaro and Hinkel, 2016[52]; Moser, Jeffress Williams and Boesch, 2012[53]). The governance and finance issues related to coastal adaptation are discussed in Chapter 2.

References

[16] Abidin, H. et al. (2015), “On correlation between urban development, land subsidence and flooding phenomena in Jakarta”, Proceedings of the International Association of Hydrological Sciences 370, pp. 15-20.

[23] Barbier, E. et al. (2011), “The value of estuarine and coastal ecosystem services”, Ecological Monographs, Vol. 81/2, pp. 169-193, http://dx.doi.org/10.1890/10-1510.1.

[32] Beck, M. et al. (2018), “The global flood protection savings provided by coral reefs”, Nature Communications, Vol. 9/1, p. 2186, http://dx.doi.org/10.1038/s41467-018-04568-z.

[25] Bisaro, A. and J. Hinkel (2018), “Mobilizing private finance for coastal adaptation: A literature review”, Wiley Interdisciplinary Reviews: Climate Change, Vol. 9/3, p. e514, http://dx.doi.org/10.1002/wcc.514.

[52] Bisaro, A. and J. Hinkel (2016), “Governance of social dilemmas in climate change adaptation”, Nature Climate Change, Vol. 6/4, pp. 354-359, http://dx.doi.org/10.1038/nclimate2936.

[21] Church, J. et al. (2001), “Changes in sea level: Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change”, in J.T., Ding, Y., Griggs, D.J., Noguer, M., Linden, P.J. van der, Dai, X., Maskell, K., Johnson, C. (ed.), Climate Change 2001: The Scientific Basis, Cambridge University Press.

[22] Clark, P. et al. (2016), “Consequences of twenty-first-century policy for multi-millennial climate and sea-level change”, Nature Climate Change, Vol. 6/4, pp. 360-369, http://dx.doi.org/10.1038/nclimate2923.

[51] De Bruijn, K., F. Klijn and J. Knoeff (2013), Unbreachable embankments? In pursuit of the most effective stretches for reducing fatality risk.

[39] Diaz, D. (2016), “Estimating global damages from sea level rise with the Coastal Impact and Adaptation Model (CIAM)”, Climatic Change, Vol. 137/1-2, pp. 143-156, http://dx.doi.org/10.1007/s10584-016-1675-4.

[26] Donchyts, G. et al. (2016), “Earth’s surface water change over the past 30 years”, Nature Climate Change.

[18] Douglas, B., M. Kearney and S. Leatherman (2000), “Sea level rise: History and consequences”, Elsevier.

[12] Ericson, J. et al. (2006), “Effective sea-level rise and deltas: Causes of change and human dimension implications”, Global and Planetary Change, Vol. 50, pp. 63-82, http://dx.doi.org/10.1016/j.gloplacha.2005.07.004.

[35] Fankhauser, S. (1995), “Protection versus retreat: The economic costs of sea-level rise”, Environment and Planning A, Vol. 27/2, pp. 299-319, http://dx.doi.org/10.1068/a270299.

[10] Farrell, W. and J. Clark (2007), “On Postglacial Sea Level”, Geophysical Journal of the Royal Astronomical Society, Vol. 46/3, pp. 647-667, http://dx.doi.org/10.1111/j.1365-246X.1976.tb01252.x.

[8] Fettweis, X. et al. (2013), “Estimating the Greenland ice sheet surface mass balance contribution to future sea level rise using the regional atmospheric climate model MAR”, The Cryosphere, Vol. 7/2, pp. 469-489, http://dx.doi.org/10.5194/tc-7-469-2013.

[47] Haasnoot, M. et al. (2013), “Dynamic adaptive policy pathways: A method for crafting robust decisions for a deeply uncertain world”, Global Environmental Change, Vol. 23/2, pp. 485-498, http://dx.doi.org/10.1016/J.GLOENVCHA.2012.12.006.

[49] Hall, J. and D. Solomatine (2008), “A framework for uncertainty analysis in flood risk management decisions”, International Journal of River Basin Management, Vol. 6/2, pp. 85-98, http://dx.doi.org/10.1080/15715124.2008.9635339.

[19] Hanson, S. et al. (2011), “A global ranking of port cities with high exposure to climate extremes”, Climatic Change, Vol. 104/1, pp. 89-111, http://dx.doi.org/10.1007/s10584-010-9977-4.

[17] Higgins, S. et al. (2013), “Land subsidence at aquaculture facilities in the Yellow River delta, China”, Geophysical Research Letters, Vol. 40, pp. 3898-3902.

[3] Hinkel, J. et al. (2014), “Coastal flood damage and adaptation costs under 21st century sea-level rise”, Proceedings of the National Academy of Sciences of the United States of America, Vol. 111/9, pp. 3292-3297, http://dx.doi.org/10.1073/pnas.1222469111.

[40] Hinkel, J. et al. (2014), “Coastal flood damage and adaptation costs under 21st century sea-level rise.”, Proceedings of the National Academy of Sciences of the United States of America, Vol. 111/9, pp. 3292-7, http://dx.doi.org/10.1073/pnas.1222469111.

[2] Masson-Delmotte, V. (ed.) (2018), Impacts of 1.5°C global warming on natural and human systems, Intergovernmental Panel on CLimate Change, https://www.ipcc.ch/site/assets/uploads/sites/2/2018/11/SR15_Chapter3_Low_Res.pdf (accessed on 10 December 2018).

[44] IIASA (2012), SSP Database..

[1] Intergovernmental Panel on Climate Change (2014), “Glossary, acronyms and chemical symbols”, in Climate Change 2014 Mitigation of Climate Change, Cambridge University Press, Cambridge, http://dx.doi.org/10.1017/CBO9781107415416.023.

[46] Lempert, R. and M. Schlesinger (2001), “Climate-change strategy needs to be robust”, Nature, Vol. 412/6845, pp. 375-375, http://dx.doi.org/10.1038/35086617.

[20] Levermann, A. et al. (2013), “The multimillennial sea-level commitment of global warming.”, Proceedings of the National Academy of Sciences of the United States of America, Vol. 110/34, pp. 13745-50, http://dx.doi.org/10.1073/pnas.1219414110.

[9] Levermann, A. et al. (2014), “Projecting Antarctic ice discharge using response functions from SeaRISE ice-sheet models”, Earth System Dynamics, Vol. 5, pp. 271-293, http://dx.doi.org/10.5194/esd-5-271-2014.

[4] Lincke, D. and J. Hinkel (2018), “Economically robust protection against 21st century sea-level rise”, Global Environmental Change, Vol. 51, pp. 67-73, http://dx.doi.org/10.1016/J.GLOENVCHA.2018.05.003.

[30] Liquete, C. et al. (2013), “Assessment of coastal protection as an ecosystem service in Europe”, Ecological Indicators, Vol. 30, pp. 205-217, http://dx.doi.org/10.1016/j.ecolind.2013.02.013.

[29] Li, Y. et al. (2013), “Coastal wetland loss and environmental change due to rapid urban expansion in Lianyungang, Jiangsu, China”, Regional Environmental Change, Vol. 14/3, pp. 1175-1188, http://dx.doi.org/10.1007/s10113-013-0552-1.

[7] Marzeion, B., A. Jarosch and M. Hofer (2012), “Past and future sea-level change from the surface mass balance of glaciers”, The Cryosphere, Vol. 6, http://dx.doi.org/10.5194/tc-6-1295-2012.

[27] Ma, Z. et al. (2014), “Rethinking China’s new great wall.”, Science , Vol. 346, pp. 912–914.

[33] Mehvar, S. et al. (2018), “Quantifying economic value of coastal ecosystem services: A review”, Journal of Marine Science and Engineering, Vol. 6/1, p. 5, http://dx.doi.org/10.3390/jmse6010005.

[53] Moser, S., S. Jeffress Williams and D. Boesch (2012), “Wicked challenges at Land's End: Managing coastal vulnerability under climate change”, Annual Review of Environment and Resources, Vol. 37/1, pp. 51-78, http://dx.doi.org/10.1146/annurev-environ-021611-135158.

[24] Murray, N. et al. (2014), “Tracking the rapid loss of tidal wetlands in the Yellow Sea”, Frontiers in Ecology and the Environment, Vol. 12/5, pp. 267-272, http://dx.doi.org/10.1890/130260.

[14] Nicholls, R. (1995), “Coastal mega-cities and climate change.”, GeoJournal, Vol. 37, pp. 369-379.

[5] Nicholls, R. et al. (2013), “Sea-level scenarios for evaluating coastal impacts”, Wiley Interdisciplinary Reviews: Climate Change, Vol. 5/1, pp. 129-150, http://dx.doi.org/10.1002/wcc.253.

[41] Nicholls, R. et al. (2011), “Sea-level rise and its possible impacts given a 'beyond 4 C world' in the twenty-first century”, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 369/1934, pp. 161-181, http://dx.doi.org/10.1098/rsta.2010.0291.

[50] Nicholls, R., R. Tol and A. Vafeidis (2008), “Global estimates of the impact of a collapse of the West Antarctic ice sheet: an application of FUND”, Climatic Change, Vol. 91/1-2, pp. 171-191, http://dx.doi.org/10.1007/s10584-008-9424-y.

[36] Nicholls, R., R. Tol and A. Vafeidis (2008), “Global estimates of the impact of a collapse of the West Antarctic ice sheet: An application of FUND”, Climatic Change, Vol. 91/1-2, pp. 171-191, http://dx.doi.org/10.1007/s10584-008-9424-y.

[48] O’Neill, B. et al. (2014), “A new scenario framework for climate change research: The concept of shared socioeconomic pathways”, Climatic Change, Vol. 122, pp. 387–400, https://doi.org/10.1007/s10584-013-0905-2.

[11] Perrette, M. et al. (2013), “A scaling approach to project regional sea level rise and its uncertainties”, Earth System Dynamics, Vol. 4/1, pp. 11-29, http://dx.doi.org/10.5194/esd-4-11-2013.

[31] Spalding, M. et al. (2014), “The role of ecosystems in coastal protection: Adapting to climate change and coastal hazards”, Ocean & Coastal Management, Vol. 90, pp. 50-57, http://dx.doi.org/10.1016/J.OCECOAMAN.2013.09.007.

[34] Spencer, T. et al. (2016), “Global coastal wetland change under sea-level rise and related stresses: The DIVA Wetland Change Model”, Global and Planetary Change, Vol. 139, pp. 15-30.

[37] Sugiyama, M., R. Nicholls and A. Vafeidis (2008), “Estimating the economic cost of sea-level rise”, Joint Program on the Science and Policy of Global Change Reports, Vol. 156, https://dspace.mit.edu/handle/1721.1/41522 (accessed on 29 November 2018).

[6] Taylor, K. et al. (2012), “An overview of CMIP5 and the experiment design”, Bulletin of the American Meteorological Society, Vol. 93/4, pp. 485-498, http://dx.doi.org/10.1175/BAMS-D-11-00094.1.

[45] Tol, R. (2002), “Estimates of the damage costs of climate change. Part 1: Benchmark estimates”, Environmental and Resource Economics, Vol. 21/1, pp. 47-72, https://doi.org/10.1023/A:1014500930521.

[42] Vafeidis, A. (2008), “A new global coastal database for impact and vulnerability analysis to sea-level rise”, Journal of Coastal Research, Vol. 24/4, pp. 917-924, http://dx.doi.org/10.2112/06-0725.1.

[43] Vafeidis, A. et al. (2008), “A new global coastal database for impact and vulnerability analysis to sea-level rise”, Journal of Coastal Research, Vol. 244/4, pp. 917-924, http://dx.doi.org/10.2112/06-0725.1.

[28] Wong, P. et al. (2014), Coastal Systems and Low-Lying Areas, IPCC, https://www.ipcc.ch/pdf/assessment-report/ar5/wg2/WGIIAR5-Chap5_FINAL.pdf (accessed on 17 August 2017).

[13] Woodroffe, C. et al. (2006), “Landscape variability and the response of Asian megadeltas to environmental change”, Global Change and Integrated Coastal Management, pp. 277-314.

[15] World Bank (2010), “The Global Report on the Economics of Adaptation to Climate Change Study”.

[38] Yohe, G., J. Neumann and H. Ameden (1995), “Assessing the economic cost of greenhouse-induced sea level rise: Methods and application in support of a national survey”, Journal of Environmental Economics and Management, Vol. 29/3, pp. S78-S97, http://dx.doi.org/10.1006/jeem.1995.1062.

Note

← 1. As the terminology used in this report is relevant to climate change adaptation, the definitions may differ slightly from other frameworks used to describe risk.