Editing IPCC:AR6/SROCC/Chapter-4 (section)

== Box 4.2 Methodological Advances in Exposure and Vulnerability Assessments ==

<div id="section-4-3-1introduction-block-1"></div>

This box highlights recent advances in methodologies in assessing exposure and vulnerability to sea level rise (SLR) and its physical impacts, such as coastal flooding since the IPCC 5th Assessment Report (AR5). In few cases it also leverages methodological advances, which have not been yet applied in the coastal context but have great potential to inform coastal assessments.

'''''Improved spatial-temporal exposure assessments'''''

Exposure assessment is frequently based on census data, which is available at coarse resolutions. However, new technologies (e.g., drones and mobile phone data) and more available satellite products provide new tools for exposure analysis. Exposure assessment is increasingly based on the combination of high resolution satellite imagery and spatio-temporal population modelling as well as improved quality of digital elevation models (DEM; Kulp and Strauss, 2017 <sup>[[#fn:r879|879]]</sup> ). This is used to understand better exposure to coastal flooding (Kulp and Strauss, 2017), diurnal differences in flood risk exposure (Smith et al., 2016 <sup>[[#fn:r880|880]]</sup> ), dynamic gridded population information for daily and seasonal differences in exposure (Renner et al., 2017 <sup>[[#fn:r881|881]]</sup> ), a combination of remotely-sensed and geospatial data with modelling for a gridded prediction of population density at ~100 m spatial resolution (Stevens et al., 2015 <sup>[[#fn:r882|882]]</sup> ), or open building data using building locations, footprint areas and heights (Figueiredo and Martina, 2016 <sup>[[#fn:r883|883]]</sup> ). In addition, methods based on mobile phone data (Deville et al., 2014 <sup>[[#fn:r884|884]]</sup> ; Ahas et al., 2015 <sup>[[#fn:r885|885]]</sup> ), and social media-based participation are increasingly available for population distribution mapping (Steiger et al., 2015 <sup>[[#fn:r886|886]]</sup> ). Some of these methodologies have been already applied in coastal assessments (Smith et al., 2016 <sup>[[#fn:r887|887]]</sup> ). Integrating daily and seasonal changes with the distribution of population improves population exposure information for risk assessments especially in areas with highly dynamic population distributions, as shown in high tourism areas in mountain regions (e.g., Renner et al., 2017), which would have advantages at touristic coastal areas as well.

'''''Projections of future exposure'''''

Recent studies assess exposure considering not only projected sea levels but also expected changes in population size (Jongman et al., 2012 <sup>[[#fn:r888|888]]</sup> ; Hauer et al., 2016 <sup>[[#fn:r889|889]]</sup> ). It involves different socioeconomic scenarios together with changing growth rates for coastal areas and the hinterland (Neumann et al., 2015 <sup>[[#fn:r890|890]]</sup> ) and using spatially explicit simulation models for urban, residential and rural areas (Sleeter et al., 2017 <sup>[[#fn:r891|891]]</sup> ). Migration-based changes in population distribution (Merkens et al., 2016 <sup>[[#fn:r892|892]]</sup> ; Hauer, 2017 <sup>[[#fn:r893|893]]</sup> ) are also considered, as well as simulated future land use (specifically urban growth) to investigate future exposure to SLR (Song et al., 2017 <sup>[[#fn:r894|894]]</sup> ). Other studies assess future exposure trends by accounting for the role of varying patterns of topography and development projections leading to different rates of anticipated future exposure (Kulp and Strauss, 2017 <sup>[[#fn:r895|895]]</sup> ), which influence how effectively coastal communities can adapt. Recent studies aim to account for the sociodemographic characteristics of potentially exposed future populations (Shepherd and Binita, 2015 <sup>[[#fn:r896|896]]</sup> ), and anticipate future risk by projecting the evolution of the exposure of vulnerable populations and groups (Hardy and Hauer, 2018 <sup>[[#fn:r897|897]]</sup> ). Using social heterogeneity modelling when developing future exposure scenarios enhances the quality of risk assessments in coastal areas (Rao et al., 2017 <sup>[[#fn:r858|858]]</sup> ; Hardy and Hauer, 2018 <sup>[[#fn:r899|899]]</sup> ). Subnational population dynamics combined with an extended coastal narrative-based version of the five shared socioeconomic pathways (SSP) for global coastal population distribution was used for assessing global climate impacts at the coast, highlighting regions where high coastal population growth is expected and which therefore face increased exposure to coastal flooding (Merkens et al., 2016 <sup>[[#fn:r900|900]]</sup> ). SSPs have also been used to estimate future population in regional coastal-hazard risk exposure studies (Vousdoukas et al., 2018b <sup>[[#fn:r901|901]]</sup> ).

'''''Advances in vulnerability assessment'''''

Since the IPCC Special Report on Managing the Risks of Extreme Events and Disasters (SREX) report, vulnerability has been more consistently considered in climate risk assessments ( ''medium confidence'' ). It is recognised that climate risk is not only hazard-driven, but also a sociopolitical and economic phenomenon that evolves with changing societal and institutional conditions ( ''high confidence'' ). Many studies related to climate risk and adaptation include vulnerability assessments, most of them considering vulnerability as a pre-existing condition while some interpret vulnerability as an outcome (Jurgilevich et al., 2017 <sup>[[#fn:r902|902]]</sup> ).

'''''Increasing importance of dynamic assessments'''''

The dynamic nature of vulnerability, and the need to align climate forecasts with socioeconomic scenarios, was a key message of IPCC SREX. Challenges in methodology and data availability, particularly of future socioeconomic data is overcome by extrapolating empirical information of past trends in vulnerability to flooding (Jongman et al., 2015 <sup>[[#fn:r903|903]]</sup> ; Mechler and Bouwer, 2015 <sup>[[#fn:r904|904]]</sup> ; Kreibich et al., 2017 <sup>[[#fn:r905|905]]</sup> ), downscaling global scenarios, for example, the SSPs (Van Ruijven et al., 2014; Viguié et al., 2014 <sup>[[#fn:r906|906]]</sup> ; Absar and Preston, 2015 <sup>[[#fn:r907|907]]</sup> ), or by using participatory methods, surveys and interviews to develop future scenarios (Ordóñez and Duinker, 2015 <sup>[[#fn:r908|908]]</sup> ; Tellman et al., 2016 <sup>[[#fn:r909|909]]</sup> ). The uncertainty of the downscaled projections needs to be considered along with the limitation that, even if population data projections are available, the future level of education, poverty, etc. is hard to predict (Jurgilevich et al., 2017 <sup>[[#fn:r910|910]]</sup> ). Suggestions to overcome these shortcomings entail the use of a combination of different data sources for triangulation and inclusion of uncertainties (Hewitson et al., 2014 <sup>[[#fn:r911|911]]</sup> ), or the meaningful involvement of stakeholders to project plausible future socioeconomic conditions through co-production (Jurgilevich et al., 2017 <sup>[[#fn:r912|912]]</sup> ). Recent innovations in (flood) risk assessment include the integration of behaviour into risk assessments (Aerts et al., 2018b <sup>[[#fn:r913|913]]</sup> ) as well as vulnerabilities related to cascading events (Serre and Heinzlef, 2018 <sup>[[#fn:r914|914]]</sup> ).

'''''Social-ecological vulnerability assessments'''''

Especially in rural, natural resource-dependent settings, where the population directly rely on the services provided by ecosystems, the vulnerability of the ecosystems (e.g., fragmented, degraded ecosystems with low biodiversity) directly influence that of the population. Since AR5, several methods have been developed and piloted to assess and map social-ecological vulnerability. Examples include the use of i) the sustainable livelihood approach and resource dependence metrics for Australian coastal communities (Metcalf et al., 2015 <sup>[[#fn:r915|915]]</sup> ), ii) integration of local climate forecasts for coral reef fisheries in Papua New Guinea (Maina et al., 2016 <sup>[[#fn:r916|916]]</sup> ), iii) ecosystem supply-demand model for an integrated vulnerability assessment in Rostock, Germany (Beichler, 2015 <sup>[[#fn:r917|917]]</sup> ), iv) participatory indicator development for multiple hazards in river deltas (Hagenlocher et al., 2018 <sup>[[#fn:r918|918]]</sup> ), and v) human-nature dependencies and ecosystem services for small ‐ scale fisheries in French Polynesia (Thiault et al., 2018 <sup>[[#fn:r919|919]]</sup> ). Areas, where social vulnerability prevail may be, but are not necessarily associated with hotspots of ecosystem vulnerability, highlighting the need to specifically adapt management interventions to local social-ecological settings and to adaptation goals (Hagenlocher et al., 2018 <sup>[[#fn:r920|920]]</sup> ; Thiault et al., 2018 <sup>[[#fn:r921|921]]</sup> ). The number of assessments considering both the social and the ecological part of the system are increasingly used (Sebesvari et al., 2016 <sup>[[#fn:r922|922]]</sup> ).

'''''Assessment of vulnerability to multiple hazards simultaneously'''''

Increasingly, multi-hazard risk assessments are undertaken at the coast (e.g., flooding and inundation of coastal lands in India; Kunte et al., 2014 <sup>[[#fn:r923|923]]</sup> ), to understand the inter-relationships between hazards (e.g., Gill and Malamud, 2014), and by focusing on hazard interactions where one hazard triggers another or increases the probability of others occurring. Liu et al. (2016a) provide a systematic hazard interaction classification based on the geophysical environment that allows for the consideration of all possible interactions (independent, mutex, parallel and series) between different hazards, and for the calculation of the probability and magnitude of multiple interacting natural hazards occurring together. Advances have been reported since AR5 by using, for example, modular sets of vulnerability indicators, flexibly adapting to the hazard situation (Hagenlocher et al., 2018 <sup>[[#fn:r924|924]]</sup> ).

'''''Using vulnerability functions, thresholds, innovative ways of aggregation in indicator-based assessment, improved data sources'''''

The use of vulnerability functions has been shown to be helpful in assessing the damage response of buildings to tsunamis (Tarbotton et al., 2015 <sup>[[#fn:r925|925]]</sup> ), to coastal surge and wave hazards (Hatzikyriakou and Lin, 2017 <sup>[[#fn:r926|926]]</sup> ) and accounting for non-linear relationships between mortality and temperature above a ‘comfort temperature’ (El-Zein and Tonmoy, 2017 <sup>[[#fn:r927|927]]</sup> ). Acknowledging the non-compensatory nature of different vulnerability indicators (e.g., proximity to the sea cannot always be fully compensated by being wealthy), the concepts of preference, indifference and dominance thresholds have been applied as a form of data aggregation (Tonmoy and El-Zein, 2018 <sup>[[#fn:r928|928]]</sup> ). Similar to advances in exposure assessments, freely available data and mobile technologies hold promise for enabling better input data for vulnerability assessments. Examples include using a combination of mobile phone and satellite data to determine and monitor vulnerability indicators such as poverty (Steele et al., 2017), and using data on subnational dependency ratios and high resolution gridded age/sex group datasets (Pezzulo et al., 2017 <sup>[[#fn:r930|930]]</sup> ).

<span id="dimensions-of-exposure-and-vulnerability-to-sea-level-rise"></span>
=== 4.3.2 Dimensions of Exposure and Vulnerability to Sea Level Rise ===

<div id="section-4-3-2-1point-of-departure"></div>

<span id="point-of-departure"></span>
==== 4.3.2.1 Point of Departure ====

<div id="section-4-3-2-2settlement-trends"></div>

<span id="settlement-trends"></span>
==== 4.3.2.2 Settlement Trends ====

<div id="section-4-3-2-2settlement-trends-block-1"></div>

Major changes in coastal settlement patterns have occurred in the course of the 20th century, and are continuing to take place due to various complex interacting processes (Moser et al., 2012; Bennett et al., 2016) that together configure and concentrate exposure and vulnerability to climate change and SLR along the coast (Newton et al., 2012; Bennett et al., 2016) . These processes include population growth and demographic changes (Smith, 2011; Neumann et al., 2015) , urbanisation and a rural exodus, tourism development, and displacement or (re)settlement of some indigenous communities (Ford et al., 2015) . This has resulted in a growing number of people living in the Low Elevation Coastal Zone (LECZ, coastal areas below 10 m of elevation; around 11% of the world’s population in 2010; Neumann et al., 2015; Jones and O’Neill, 2016; Merkens et al., 2016) and in significant infrastructure and assets being located in risk-prone areas ( ''high confidence'' ). High density coastal urban development is commonplace in both developed and developing countries, as documented in recent case studies, for example in Canada (Fawcett et al., 2017) , China (Yin et al., 2015; Lilai et al., 2016; Yan et al., 2016) , Fiji (Hay, 2017) , France (Genovese and Przyluski, 2013; Chadenas et al., 2014; Magnan and Duvat, 2018) , Israel (Felsenstein and Lichter, 2014) , Kiribati (Storey and Hunter, 2010; Duvat et al., 2013) , New Zealand (Hart, 2011) and the USA (Heberger, 2012; Grifman et al., 2013; Liu et al., 2016b) . This has implications for levels of SLR risk at regional and local scales ( ''medium evidence, high agreement'' ). In Latin America and the Caribbean, for example, it is estimated that 6–8% of the population live in areas that are at high or very high risk of being affected by coastal hazards (Reguero et al., 2015; Calil et al., 2017; Villamizar et al., 2017) , with higher percentages in Caribbean islands (Mycoo, 2018) . In the Pacific, ~57% of Pacific Island countries’ built infrastructure are located in risk-prone coastal areas (Kumar and Taylor, 2015) . In Kiribati, due to the flow of outer, rural populations to limited, low-elevated capital islands, together with constraints inherent in the sociocultural land tenure system, the built area located &lt;20 m from the shoreline quadrupled between 1969 and 2007–2008 (Duvat et al., 2013) . Other examples of rural exodus are reported in the recent literature, for example in the Maldives (Speelman et al., 2017) .

Population densification also affects rural areas’ exposure and vulnerability, and interacts with other factors shaping settlement patterns, such as the fact that ‘indigenous peoples in multiple geographical contexts have been pushed into marginalised territories that are more sensitive to climate impacts, in turn limiting their access to food, cultural resources, traditional livelihoods and place-based knowledge (…) [and therefore undermining] aspects of social-cultural resilience’ (Ford et al., 2016b, p. 350) . In the Pacific, for example, ‘while traditional settlements on high islands (…) were often located inland, the move to coastal locations was encouraged by colonial and religious authorities and more recently through the development of tourism’ (Ballu et al., 2011; Nurse et al., 2014, p. 1623; Duvat et al., 2017) . Although these population movements are orders of magnitude smaller than the global trends described above, they play a critical role at the very local scale in explaining the emergence of, or changes in exposure and vulnerability. In atoll contexts, for example, the growing pressure on freshwater resources together with a loss in local knowledge (e.g., how to collect water from palm trees), result in increased exposure of communities to brackish, polluted groundwater, inducing water insecurity and health problems (Storey and Hunter, 2010; Lazrus, 2015) .

<div id="section-4-3-2-4other-human-dimensions"></div>

<span id="other-human-dimensions"></span>
==== 4.3.2.4 Other Human Dimensions ====

<div id="section-4-3-2-4other-human-dimensions-block-1"></div>

The development of local scale case studies from a social science perspective, for example, in the Arctic (Ford et al., 2012 <sup>[[#fn:r1075|1075]]</sup> ; Ford et al., 2014 <sup>[[#fn:r1076|1076]]</sup> ) , small islands (Petzold, 2016 <sup>[[#fn:r1077|1077]]</sup> ; Duvat et al., 2017 <sup>[[#fn:r1078|1078]]</sup> ) and within cities (Rosenzweig and Solecki, 2014 <sup>[[#fn:r1079|1079]]</sup> ; Paterson et al., 2017 <sup>[[#fn:r1080|1080]]</sup> ; Texier-Teixeira and Edelblutte, 2017 <sup>[[#fn:r1081|1081]]</sup> ) or at the household level (Koerth et al., 2014 <sup>[[#fn:r1082|1082]]</sup> ) support a better understanding of the anthropogenic drivers of exposure and vulnerability. Four examples of drivers that were only emerging at the time of the AR5 are discussed below. Very importantly, another major emerging dimension that is not discussed here but rather in Section 4.4.4, relates to power asymmetries, politics, and the prevailing political economy, which are important drivers of exposure and vulnerability to SLR-related coastal hazards, and consequently adaptation prospects (Eriksen et al., 2015 <sup>[[#fn:r1083|1083]]</sup> ; Dolšak and Prakash, 2018 <sup>[[#fn:r1084|1084]]</sup> ) . Recent literature provides examples in coastal megacities like Jakarta, Indonesia (Shatkin, 2019 <sup>[[#fn:r1085|1085]]</sup> ) as well as in smaller cities, like Maputo, Mozambique (Broto et al., 2015 <sup>[[#fn:r1086|1086]]</sup> ) and Surat, India (Chu, 2016a; Chu, 2016b) , and many other coastal cities and settlements around the world ( ''high confidence'' ; Jones et al., 2015 <sup>[[#fn:r1087|1087]]</sup> ; Allen et al., 2018 <sup>[[#fn:r1088|1088]]</sup> ; Hughes et al., 2018 <sup>[[#fn:r1089|1089]]</sup> ; Sovacool, 2018 <sup>[[#fn:r1090|1090]]</sup> ) .

<div id="section-4-3-2-4other-human-dimensions-block-2"></div>

<span id="gender-inequality"></span>
===== 4.3.2.4.1 Gender inequality =====

Gender inequality came to prominence only recently in climate change studies (~15 years ago; see Pearse, 2017 <sup>[[#fn:r1091|1091]]</sup> ) . In light of sea-related hazards and SLR specifically, the issue is still mainly investigated in the context of developing countries, although growing attention is paid to the issue in developed countries (e.g., Lee et al., 2015; Pearse, 2017) . Recent studies in southern coastal Bangladesh, for example, show that women get less access than men to climate- and disaster-related information (both emergency information and training programmes), decision making processes at the household and community levels, economic resources including financial means such as micro-credit, land ownership, and mobility within and outside the villages (Rahman, 2013 <sup>[[#fn:r1092|1092]]</sup> ; Alam and Rahman, 2014 <sup>[[#fn:r1093|1093]]</sup> ; Garai, 2016 <sup>[[#fn:r1094|1094]]</sup> ) . Gender inequity may be inherent in unfavourable background conditions (higher illiteracy rates, deficiencies in food and calories intake and poorer health conditions) as a result of, among other things, traditions, social norms and patriarchy. Together, these barriers disadvantage women more than men in developing effective responses to anticipate gradual environmental changes such as persistent coastal erosion, flooding and soil salinisation ( ''medium evidence, high agreement'' ). Such conclusions are in line with the literature on gender inequality and climate change at large (Alston, 2013 <sup>[[#fn:r1095|1095]]</sup> ; Pearse, 2017 <sup>[[#fn:r1096|1096]]</sup> ) , thus suggesting no major SLR-inherent specificities.

<div id="section-4-3-2-4other-human-dimensions-block-3"></div>

<span id="loss-of-indigenous-knowledge-and-local-knowledge"></span>
===== 4.3.2.4.2 Loss of indigenous knowledge and local knowledge =====

Despite the identification of this issue in AR4, its treatment in AR5 remained limited. Recent literature partly focussing on SLR reaffirms that indigenous knowledge and local knowledge (IK and LK; Cross-Chapter Box 4 in Chapter 1 and Glossary) are key to determining how people recognise and respond to environmental risk (Bridges and McClatchey, 2009 <sup>[[#fn:r1097|1097]]</sup> ; Lefale, 2010 <sup>[[#fn:r1098|1098]]</sup> ; Leonard et al., 2013 <sup>[[#fn:r1099|1099]]</sup> ; Lazrus, 2015 <sup>[[#fn:r1100|1100]]</sup> ), and therefore to increasing adaptive capacity and reducing long-term vulnerability (Ignatowski and Rosales, 2013 <sup>[[#fn:r1101|1101]]</sup> ; McMillen et al., 2014 <sup>[[#fn:r1102|1102]]</sup> ; Hesed and Paolisso, 2015 <sup>[[#fn:r1103|1103]]</sup> ; Janif et al., 2016 <sup>[[#fn:r1104|1104]]</sup> ; Morrison, 2017 <sup>[[#fn:r1105|1105]]</sup> ).

IK and LK contribute both as a foundation for and an outcome of customary resource management systems aimed at regulating resource use and securing critical ecosystem protection (examples in Indonesia; Hiwasaki et al., 2015 <sup>[[#fn:r1106|1106]]</sup> ), structuring the relationship between people and authorities, and framing and maintaining a strong sense of place in the community (examples in Timor Leste; Hiwasaki et al., 2015 <sup>[[#fn:r1107|1107]]</sup> ). In turn, this allows local communities to predict and prepare for both sudden shock events that have historical precedent and, when IK and LK are embedded in day-to-day rituals and decision making processes, to also anticipate the consequences of gradual changes, as in sea level (examples in Indonesia; Hiwasaki et al., 2015 <sup>[[#fn:r1108|1108]]</sup> ). Customary resource management systems based on IK and elders’ leadership—for instance, Rahui in French Polynesia (Gharasian, 2016 <sup>[[#fn:r1109|1109]]</sup> ), or Mo in the Marshall Islands (Bridges and McClatchey, 2009)—also allow communities to diversify access to marine and terrestrial resources using seasonal calendars, to ensure collective food and water security, and to maintain ecological integrity (McMillen et al., 2014 <sup>[[#fn:r1111|1111]]</sup> ). In rural Pacific atolls, traditional food preservation and storage (e.g., storing germinated coconuts or drying fish) still play a role in anticipating disruptions in natural resource availability (Campbell, 2015 <sup>[[#fn:r1112|1112]]</sup> ; Lazrus, 2015 <sup>[[#fn:r1113|1113]]</sup> ). Such practices have enabled the survival of isolated communities from the Arctic to tropical islands in constraining sea environments for centuries to millennia (McMillen et al., 2014 <sup>[[#fn:r1114|1114]]</sup> ; Nunn et al., 2017a <sup>[[#fn:r1115|1115]]</sup> ). Morrison (2017) argues that IK and LK can also play a role in supporting internal migration in response to SLR, by avoiding social and cultural uprooting (Cross-Chapter Box 4 in Chapter 1).

In some specific contexts, climate change will also imply no-analogue changes, such as rapid ice-melt and changing conditions in the Arctic that have no precedent in the modern era, and could thus limit the relevance of IK and LK in efforts to address significantly different circumstances. Except in these specific situations, the literature suggests that the loss of IK and LK, and related social norms and mechanisms, will increase populations’ exposure and vulnerability to SLR impacts (Nakashima et al., 2012 <sup>[[#fn:r1117|1117]]</sup> ). The literature notably points out that modern, externally-driven socioeconomic dynamics, such as the introduction of imported food (noodles, rice, canned meat and fish, etc.), diminish the cultural importance of IK-based practices and diets locally, together with introducing dependency on monetisation and external markets (Hay, 2013; Campbell, 2015 <sup>[[#fn:r1118|1118]]</sup> ).

As a result, the loss of IK and LK may increase long-term vulnerability to SLR ( ''medium evidence, high agreement'' ). Given that IK and LK are largely based on observing and ‘making sense’ of the surrounding environment (moon, waves, winds, animal behaviours, topography, etc.), such a loss reflects a more general concern about the weakening of environmental connectedness in contemporary societies, which is not limited to remote, rural and developing communities ( ''medium confidence'' ). In developed contexts too, the loss of LK has played a critical role in recent coastal disasters (e.g., Katrina in 2005 in the USA, Kates et al., 2006) and increasing vulnerability to SLR (e.g., Newton and Weichselgartner, 2014; Wong et al., 2014 <sup>[[#fn:r1119|1119]]</sup> ).

<div id="section-4-3-2-4other-human-dimensions-block-4"></div>

<span id="social-capital"></span>
===== 4.3.2.4.3 Social capital =====

Coastal communities draw on social structures and capabilities that can reduce risk and increase adaptive capacity in the face of coastal hazards (Aldrich, 2017 <sup>[[#fn:r1120|1120]]</sup> ; Petzold, 2018 <sup>[[#fn:r1121|1121]]</sup> ) . Although the term is subject to debate (Meyer, 2018 <sup>[[#fn:r1122|1122]]</sup> ) , social capital—that is, the level of cohesion between individuals, between groups of individuals, and between people and institutions, within and between communities—is considered to be a key enabler for collective action to reduce risk and build adaptive capacity (Adger, 2010 <sup>[[#fn:r1123|1123]]</sup> ; Aldrich and Meyer, 2015 <sup>[[#fn:r1124|1124]]</sup> ; Petzold and Ratter, 2015 <sup>[[#fn:r1125|1125]]</sup> ) . Levels of social capital can be influenced by underlying social processes, such as socioeconomic (in)equalities, gender issues, health, social networks and social media. It applies to both developing and developed countries, for example in densely populated deltas (Jordan, 2015 <sup>[[#fn:r1126|1126]]</sup> ) , European coasts (Jones and Clark, 2014 <sup>[[#fn:r1127|1127]]</sup> ; Petzold, 2016 <sup>[[#fn:r1128|1128]]</sup> ) , Asian urban or semi-urban coastal areas (Lo et al., 2015 <sup>[[#fn:r1129|1129]]</sup> ; Triyanti et al., 2017 <sup>[[#fn:r1130|1130]]</sup> ) and Pacific islands (Neef et al., 2018 <sup>[[#fn:r1131|1131]]</sup> ) . Social capital framed as an enabler for reducing vulnerability has been studied in the context of extreme events (risk prevention mechanisms, emergency responses and post-crisis actions) and collective environmental management (e.g., replanting mangroves, beach cleaning, etc.). Social capital also enables adaptation prospects. For example, its role has been explored in public acceptability of long-term coastal adaptation policies in the UK (Jones and Clark, 2014 <sup>[[#fn:r1132|1132]]</sup> ; Jones et al., 2015 <sup>[[#fn:r1133|1133]]</sup> ) . The role of social capital in building resilience to climate stress in coastal Bangladesh was explored by Jordan (2015) , who found complex and even contradictory interactions between social capital and resilience to climate stress. Among others, Jordan (2015) also advises caution about uncritical importation of such Westernised concepts in seeking to understand and address coastal vulnerability in developing countries.

<div id="section-4-3-2-4other-human-dimensions-block-5"></div>

<span id="risk-perception"></span>
===== 4.3.2.4.4 Risk perception =====

Risk perception, which is context-specific and varies from one individual to another, may influence communities’ exposure and vulnerability as it shapes authorities’ and people’s attitudes towards sudden and slow onset hazards, as shown by Terpstra (2011) <sup>[[#fn:r1134|1134]]</sup> , Lazrus (2015), Elrick-Barr et al. (2017) <sup>[[#fn:r1135|1135]]</sup> and O’Neill et al. (2016) in the Netherlands, Tuvalu, Australia and Ireland, respectively. The progressive discounting of coastal hazard risks and subsequent loss of risk memory also played a role in coastal disasters such as Hurricane Katrina in 2005 in the USA (Burby, 2006 <sup>[[#fn:r1136|1136]]</sup> ; Kates et al., 2006 <sup>[[#fn:r1137|1137]]</sup> ) and Storm Xynthia in 2010 in France (Vinet et al., 2012 <sup>[[#fn:r1138|1138]]</sup> ; Genovese and Przyluski, 2013 <sup>[[#fn:r1139|1139]]</sup> ; Chadenas et al., 2014 <sup>[[#fn:r1140|1140]]</sup> ).

Risk perceptions stem from intertwined predictors such as ‘gender, political party identification, cause-knowledge, impact-knowledge, response-knowledge, holistic affect, personal experience with extreme weather events, [social norms] and biospheric value orientations’ (Kellens et al., 2011 <sup>[[#fn:r1141|1141]]</sup> ; Carlton and Jacobson, 2013 <sup>[[#fn:r1142|1142]]</sup> ; Lujala et al., 2015 <sup>[[#fn:r1143|1143]]</sup> ; van der Linden, 2015, p. 112; Weber, 2016 <sup>[[#fn:r1144|1144]]</sup> ; Elrick-Barr et al., 2017 <sup>[[#fn:r1145|1145]]</sup> ; Goeldner-Gianella et al., 2019 <sup>[[#fn:r1146|1146]]</sup> ). In general, there is a lack of education, training and thus knowledge and literacy on recent and projected trends in sea level, which compromises ownership of science facts and projections at all levels, from individuals and institutions to society at large.

While some studies have begun to highlight the influence of the distance from the sea on risk perceptions (Milfont et al., 2014 <sup>[[#fn:r1147|1147]]</sup> ; Lujala et al., 2015 <sup>[[#fn:r1148|1148]]</sup> ; O’Neill et al., 2016), there is still little knowledge about how risk perceptions vary across different geographical and social contexts, and how this influences exposure and vulnerability to coastal hazards (e.g., Terpstra, 2011; van der Linden, 2015 <sup>[[#fn:r1149|1149]]</sup> ). There is a critical lack of studies specifically addressing SLR. Some recent works conducted in coastal Australia suggest that while people are confident about their ability to cope with an already experienced event, when it comes to SLR, the dominant narrative is articulated around the barriers related to the ‘uncertainty in the nature and scale of the impacts as well as the response options available’ (Elrick-Barr et al., 2017, p. 1147). Similar conclusions have been highlighted in the Caribbean islands of St. Vincent (Smith, 2018 <sup>[[#fn:r1150|1150]]</sup> ) and the Bahamas (Thomas and Benjamin, 2018 <sup>[[#fn:r1151|1151]]</sup> ). SLR is rarely addressed separately from sea-related extreme events, which masks a crucial difference between already-observed and delayed impacts. Climate change is considered a “distant psychological risk” (Spence et al., 2012 <sup>[[#fn:r1152|1152]]</sup> ), making it and SLR per se ‘markedly different from the way that our ancestors have traditionally perceived threats in their local environment’ (Milfont et al., 2014 <sup>[[#fn:r1153|1153]]</sup> ; Lujala et al., 2015 <sup>[[#fn:r1154|1154]]</sup> ; van der Linden, 2015, p. 112; O’Neill et al., 2016).

<div id="section-4-3-2-5-towards-a-synthetic-understanding-of-the-drivers-of-exposure-and-vulnerability"></div>

<span id="towards-a-synthetic-understanding-of-the-drivers-of-exposure-and-vulnerability"></span>
==== 4.3.2.5 Towards a Synthetic Understanding of the Drivers of Exposure and Vulnerability ====

<div id="section-4-3-2-5-towards-a-synthetic-understanding-of-the-drivers-of-exposure-and-vulnerability-block-1"></div>

Recent literature confirms that anthropogenic drivers played an important role, over the last century, in increasing exposure and vulnerability worldwide, and indicates that they will continue to do so in the absence of adaptation (medium evidence, high agreement). Some scholars argue that ‘even with pervasive and extensive environmental change associated with ~2oC warming, it is non-climatic factors that primarily determine impacts, response options and barriers to adapting’ (Ford et al., 2015, p. 1046). Although it is the interaction of climate and non-climate factors that eventually determine the level of impacts, acknowledging the role of a range of purely anthropogenic drivers has important implications for action. It suggests that major action can be taken now to enhance long-term adaptation prospects, notwithstanding uncertainty about local RSL rise and resultant impacts in the distant future (medium evidence, high agreement; Magnan et al., 2016 <sup>[[#fn:r1155|1155]]</sup> ). Acting on the human-driven drivers and root causes of vulnerability could yield co-benefits, for example by improving the state and condition of coastal ecosystems – and hence the capacity to cope with or adapt to SLR impacts – or, in deltaic regions, lowering the rates of anthropogenic subsidence and, in turn, minimising changes in sea level.

In addition, coastal ecosystem degradation is acknowledged as another major non-climatic driver of exposure and vulnerability (high confidence). The ability of coastal ecosystems to serve as a buffer zone between the sea and human assets (settlements and infrastructure), and to provide regulating services with respect to SLR-related coastal hazards (including inundation and salinisation), is progressively being lost due to coastal squeeze, pollution, and habitat and land degradation mainly due to land-use conversion.

We now better understand the diversity and interactions of the climate and non-climate drivers of exposure and vulnerability, as well as their dynamics over time (Bennett et al., 2016 <sup>[[#fn:r1156|1156]]</sup> ; Duvat et al., 2017 <sup>[[#fn:r1157|1157]]</sup> ). As a result, it is now realised how many context-specificities interact (including geography, economic development, social inequity, power and politics, and risk perceptions) and play a critical role in shaping the direction and influence of individual drivers and of their possible combinations on the ground (medium evidence, high agreement; Eriksen et al., 2015 <sup>[[#fn:r1158|1158]]</sup> ; Hesed and Paolisso, 2015 <sup>[[#fn:r1159|1159]]</sup> ; McCubbin et al., 2015 <sup>[[#fn:r1160|1160]]</sup> ). This also provides a stronger foundation to identify the range of possible responses (Sections 1.6.1, 1.6.2 and 4.4.3) to observed impacts and projected risks, as well as critical areas of action to enhance adaptation pathways (Section 4.4.4).

Recent studies (e.g., cited in Sections 4.3.2.1.1, 4.3.2.2, 4.3.2.4.2 and 4.3.2.4.4) also confirm AR5 conclusions that both developing and developed countries are exposed and vulnerable to SLR (high confidence).

<div id="section-4-3-2-3terrestrial-processes-shaping-coastal-exposure-and-vulnerability"></div>

<span id="terrestrial-processes-shaping-coastal-exposure-and-vulnerability"></span>
==== 4.3.2.3 Terrestrial Processes Shaping Coastal Exposure and Vulnerability ====

<div id="section-4-3-2-3terrestrial-processes-shaping-coastal-exposure-and-vulnerability-block-1"></div>

Coastal areas, including deltas, are highly dynamic as they are affected by natural and/or human-induced processes locally or originating from both the land and the sea. Changes within the catchment can therefore have severe consequences for coastal areas in terms of sediment supply, pollution, and/or land subsidence. Sediment supply reaching the coast is a critical factor for delta sustainability (Tessler et al., 2018 <sup>[[#fn:r1037|1037]]</sup> ) and has declined drastically in the last few decades due to dam construction, land use changes and sand mining (Ouillon, 2018 <sup>[[#fn:r1038|1038]]</sup> ; ''high confidence'' ). For instance, Anthony et al. (2015) reported large-scale erosion affecting over 50% of the delta shoreline in the Mekong delta between 2003 and 2012, which was attributed in part to a reduction in surface-suspended sediments in the Mekong river potentially linked to dam construction within the river basin, sand mining in the river channels, and land subsidence linked to groundwater over-abstraction locally. Schmitt et al. (2017) <sup>[[#fn:r1039|1039]]</sup> demonstrated that these and other drivers in sediment budget changes can have severe effects on the very physical existence of the Mekong delta by the end of this century, with the most important single driver leading to inundation of large portions of the delta being ground-water pumping induced land subsidence. Thi Ha et al. (2018) estimated the decline in sediment supply to the Mekong delta to be around 75% between the 1970s and the period 2009–2016. In the Red River, the construction of the Hoa Binh Dam in the 1980s led to a 65% drop in sediment supply to the sea (Vinh et al., 2014 <sup>[[#fn:r1040|1040]]</sup> ). Based on projections of historical and 21st century sediment delivery to the Ganges-Brahmaputra-Meghna, Mahanadi and Volta deltas, Dunn et al. (2018) showed that these deltas fall short in sediment and may not be able to maintain their current elevation relative to sea level, suggesting increasing salinisation, erosion, flood hazards and adaptation demands.

Another rarely considered factor is the shift in TC climatology which also plays a critical role in explaining changes in fluvial suspended sediment loads to deltas as demonstrated by Darby et al. (2016) <sup>[[#fn:r1041|1041]]</sup> , again for the Mekong delta. More generally, most conventional engineering strategies that are commonly employed to reduce flood risk (including levees, sea walls, and dams) disrupt a delta’s natural mechanisms for building land. These approaches are rather short-term solutions which overall reduce the long-term resilience of deltas (Tessler et al., 2015 <sup>[[#fn:r1042|1042]]</sup> ; Welch et al., 2017 <sup>[[#fn:r1043|1043]]</sup> ). Systems particularly prone to flood risk due to anthropogenic activities include North America’s Mississippi River delta, Europe’s Rhine River delta, and deltas in East Asia (Renaud et al., 2013 <sup>[[#fn:r1044|1044]]</sup> ; Day et al., 2016 <sup>[[#fn:r1045|1045]]</sup> ). In regions where suspended sediments are still available in relatively large quantities, rates of sedimentation can vary depending on multiple factors, including the type of infrastructure present locally, as was shown by Rogers and Overeem (2017) <sup>[[#fn:r1046|1046]]</sup> for the Ganges-Brahmaputra-Meghna (Bengal) delta in Bangladesh as well as seasonal differences in sediment supply and place of deposition. For example, in meso-tidal and macro-tidal estuaries, during floods most of the sediments are depositing in the coastal zones and a large part of these sediments are brought back to the estuary during the low flow season by tidal pumping. This can lead to significantly higher deposition rates in the dry season as shown by Lefebvre et al. (2012) in the lower Red River estuary and by Gugliotta et al. (2018) <sup>[[#fn:r1048|1048]]</sup> in the Mekong delta. Enhanced sedimentation further upstream in estuaries and a silting-up of estuarine navigation channels can have high economic consequences for cities with a large estuarine harbour. In Haiphong city, in North Vietnam, the authorities decided to build a new harbour further downstream, for a cost estimated at 2 billion USD (Duy Vinh et al., 2018).

Overall, reduced freshwater and sediment inputs from the river basins are critical factors determining delta sustainability (Renaud et al., 2013 <sup>[[#fn:r1049|1049]]</sup> ; Day et al., 2016 <sup>[[#fn:r1050|1050]]</sup> ). In some contexts, this can be addressed through basin-scale management which allow more natural flows of water and sediments through the system, including methods for long-term flood mitigation such as improved river-floodplain connectivity, the controlled redirection of a river (i.e., avulsions) during times of elevated sediment loads, the removal of levees, and the redirection of future development to lands less prone to extreme flooding (Renaud et al., 2013 <sup>[[#fn:r1051|1051]]</sup> ; Day et al., 2016 <sup>[[#fn:r1052|1052]]</sup> ; Brakenridge et al., 2017 <sup>[[#fn:r1053|1053]]</sup> ). These actions could potentially increase the persistence of coastal landforms in the context of SLR. Next to decreasing sediment inputs to the coast, river bed and beach sand mining has been shown to contribute to shoreline erosion, for example, for shorelines of Crete (Foteinis and Synolakis, 2015 <sup>[[#fn:r1054|1054]]</sup> ), and several sub-Saharan countries such Kenya, Madagascar, Mozambique, South Africa and Tanzania (UNEP, 2015 <sup>[[#fn:r1055|1055]]</sup> ). At the global scale, 24% of the world’s sandy beaches are eroding at rates exceeding 0.5 m yr <sup>–1</sup> , while 28% are accreting for the period 1984–2016. The largest and longest eroding sandy coastal stretches are in North America (Texas; Luijendijk et al., 2018 <sup>[[#fn:r1056|1056]]</sup> ).

Shoreline erosion leads to coastal squeeze if the eroding coastline approaches fixed and hard built or natural structures as noted in AR5 (Pontee, 2013 <sup>[[#fn:r1057|1057]]</sup> ; Wong et al., 2014 <sup>[[#fn:r1058|1058]]</sup> ), a process to which SLR also contributes (Doody, 2013 <sup>[[#fn:r1059|1059]]</sup> ; Pontee, 2013 <sup>[[#fn:r1060|1060]]</sup> ). The AR5 further noted that coastal squeeze is expected to accelerate due to rising sea levels (Wong et al., 2014 <sup>[[#fn:r1061|1061]]</sup> ). Doody (2013) characterised coastal squeeze as coastal habitats being pushed landward through the effects of SLR and other coastal processes on the one hand and, on the other hand, the presence of static natural or artificial barriers effectively blocking this migration, thereby squeezing habitats into an ever narrowing space. Distinctions are made between coastal squeeze being limited to (1) the consequences of SLR vs. other environmental changes on the coastline and (2) the presence of only coastal defence structures vs. natural sloping land or other artificial infrastructure (Pontee, 2013 <sup>[[#fn:r1062|1062]]</sup> ). Recent publications have emphasised coastal squeeze related to SLR, although inland infrastructure blocking habitat migration is not necessarily limited to defence structures (Torio and Chmura, 2015 <sup>[[#fn:r1063|1063]]</sup> ; McDougall, 2017 <sup>[[#fn:r1064|1064]]</sup> ). Coastal ecosystem degradation by human activities leading to coastal erosion is also an important consideration (McDougall, 2017 <sup>[[#fn:r1065|1065]]</sup> ). Taking into consideration the current challenges to attribute coastal impacts to SLR (Section 4.3.3.1), it can be hypothesised here that as long as SLR impacts remain moderate, the dominant driving factor of coastal squeeze will be anthropogenic land-based development (e.g., Section 4.3.2.2). With higher SLR scenarios and in the case of no further development at the coast, SLR may become the dominant driver before the end of this century.

Preserved coastal habitats can play important roles in reducing risks related to some coastal hazards and initiatives are being put in place to reduce coastal squeeze, such as managed realignment (Sections 4.1, 4.4.3.1) which includes removing inland barriers (Doody, 2013 <sup>[[#fn:r1066|1066]]</sup> ). Coastal squeeze can lead to degradation of coastal ecosystems and species (Martínez et al., 2014 <sup>[[#fn:r1067|1067]]</sup> ), but if inland migration is unencumbered, observation data and modelling have shown that the net area of coastal ecosystems could increase under various scenarios of SLR, depending on the ecosystems considered (Torio and Chmura, 2015 <sup>[[#fn:r1068|1068]]</sup> ; Kirwan et al., 2016 <sup>[[#fn:r1069|1069]]</sup> ; Mills et al., 2016 <sup>[[#fn:r1070|1070]]</sup> ). However, recent modelling research has shown that rapid SLR in a context of coastal squeeze could be detrimental to the areal extent and functionality of coastal ecosystems (Mills et al., 2016 <sup>[[#fn:r1071|1071]]</sup> ) and, for marshes, could lead to a reduction of habitat complexity and loss of connectivity, thus affecting both aquatic and terrestrial organisms (Torio and Chmura, 2015 <sup>[[#fn:r1072|1072]]</sup> ). Contraction of marsh extent is also identified by Kirwan et al. (2016) <sup>[[#fn:r1073|1073]]</sup> when artificial barriers to landward migration are in place. Adaptation to SLR therefore needs to account for both development and conservation objectives so that trade-offs between protection and realignment that satisfy both objectives can be identified (Mills et al., 2016 <sup>[[#fn:r1074|1074]]</sup> ).

In summary, catchment-scale changes have very direct impacts on the coastline, particularly in terms of water and sediment budgets ( ''high confidence'' ). The changes can be rapid and modify coastlines over short periods of time, outpacing the effects of SLR and leading to increased exposure and vulnerability of social-ecological systems ( ''high confidence'' ). Without losing sight of this fact, management of catchment-level processes contribute to limiting rapid increases in exposure and vulnerability. Further to hinterland influences, coastal squeeze increases coastal exposure as well as vulnerability by the loss of a buffer zone between the sea and infrastructure behind the habitat undergoing coastal squeeze. The clear implication is that coastal ecosystems progressively lose their ability to provide regulating services with respect to coastal hazards, including as a defence against SLR driven inundation and salinisation ( ''high confidence'' ). Vulnerability is also increased if freshwater resources become salinised, particularly if these resources are already scarce. The exposure and vulnerability of human communities is exacerbated by the loss of other provisioning, supporting and cultural services generated by coastal ecosystems, which is especially problematic for coast-dependent communities ( ''high confidence'' ).

<span id="observed-impacts-and-current-and-future-risk-of-sea-level-rise"></span>
=== 4.3.3 Observed Impacts, and Current and Future Risk of Sea Level Rise ===

<div id="section-4-3-3observed-impacts-and-current-and-future-risk-of-sea-level-rise-block-1"></div>

SLR leads to hazards and impacts that are also partly inherent in other processes such as starvation of sediments provided by rivers (Kondolf et al., 2014); permafrost thaw and ice retreat; or the disruption of natural dynamics by land reclamation or sediment mining. Six main concerns for low-lying coasts (Figure 4.13) are: (i) permanent submergence of land by mean sea levels or mean high tides; (ii) more frequent or intense flooding; (iii) enhanced erosion; (iv) loss and change of ecosystems; (v) salinisation of soils, ground and surface water; and (vi) impeded drainage. This section discusses some of these hazards (flooding, erosion, salinisation) as well as observed and projected impacts on some critical marine ecosystems (marshes, mangroves, lagoons, coral reefs and seagrasses), ecosystem services (coastal protection) and human societies (people, assets, infrastructures, economic and subsistence activities, inequity and well-being, etc.). In many cases, the Chapter 4 assessment of impacts and responses uses results from literature based on values of SLR and ESL events prior to SROCC. However, the general findings reported here also carry forward with the new SROCC SLR and ESL values. Except in the case of submergence and flooding of coastal areas (Section 4.3.3.2), this section assumes no major additional adaptation efforts compared to today (i.e., neither significant intensification of ongoing action nor new types of action), thus reflecting the state of knowledge in the literature.

<div id="section-4-3-3observed-impacts-and-current-and-future-risk-of-sea-level-rise-block-2"></div>

<span id="figure-4.13"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 4.13'''

<span id="figure-4.13-overview-of-the-main-cascading-effects-of-sea-level-rise-slr.-styles-and-colours-of-lines-left-hand-side-lightdark-blue-right-hand-side-dottednon-dotted-and-orangegreendark-yellowpurpleturquoise-and-boxes-are-used-only-for-the-readability-of-the-figure.-sea-level-hazards-are-discussed-in-section-4.2.-the-various-impacts-listed"></span>
<!-- IMG CAPTION -->
'''Figure 4.13 | Overview of the main cascading effects of sea level rise (SLR). Styles and colours of lines (left hand side: light/dark blue; right hand side: dotted/non dotted and orange/green/dark yellow/purple/turquoise) and boxes are used only for the readability of the figure. Sea level hazards are discussed in Section 4.2. The various impacts listed […]'''
<!-- IMG FILE -->
[[File:a70ccb0c8c0059a8a89dee0a28cf0bc9 IPCC-SROCC-CH_4_13-2318x3000.jpg]]

Figure 4.13 | Overview of the main cascading effects of sea level rise (SLR). Styles and colours of lines (left hand side: light/dark blue; right hand side: dotted/non dotted and orange/green/dark yellow/purple/turquoise) and boxes are used only for the readability of the figure. Sea level hazards are discussed in Section 4.2. The various impacts listed in this figure are discussed in the sections below: Submergence of land and enhanced flooding (4.3.3.2); Erosion of land and beaches (4.3.3.3); Salinisation (4.3.3.4); Loss of and changes in ecosystems (4.3.3.5); Loss of land and land uses (4.3.3.2); Loss of ecosystems services (4.3.3.5); Damage to people and to the built environment (4.3.3.2, 4.3.3.3, 4.3.3.4 and 4.3.3.6); Damages to human activities (4.3.3.6). Non-climate anthropogenic drivers are discussed in Section 4.3.2 and other climate-related drivers are notably discussed in Section 5.2.1 and 5.2.2.

<!-- END IMG -->
<div id="section-4-3-3-1attribution-of-observed-physical-changes-to-sea-level-rise"></div>

<span id="attribution-of-observed-physical-changes-to-sea-level-rise"></span>
==== 4.3.3.1 Attribution of Observed Physical Changes to Sea Level Rise ====

<div id="section-4-3-3-1attribution-of-observed-physical-changes-to-sea-level-rise-block-1"></div>

The AR5 concludes that attribution of coastal changes to SLR is difficult because ‘the coastal sea level change signal is often small when compared to other processes’ (Wong et al., 2014: 375). New literature, however, shows that extreme water levels at the coast are rising due to mean SLR (4.2.2.4 for observations, and 4.3.5 for projections), with observable impacts on chronic flooding in some regions (Sweet and Park, 2014; Strauss et al., 2016).

On coastal morphological changes for example, contemporary SLR currently acts as a ‘background driver’, with extreme events, changes in wave patterns, tides and human intervention often described as the prevailing drivers of observed changes (Grady et al., 2013; Albert et al., 2016). Morphological changes are also interacting with other impacts of SLR, such as coastal flooding (Pollard et al., 2018). Despite the complexity of the attribution issue (Romine et al., 2013; Le Cozannet et al., 2014), recent literature suggests possibly emerging signs of the direct influence of recent SLR on shoreline behaviour, for example on small highly-sensitive reef islands in New Caledonia (Garcin et al., 2016) and in the Solomon Islands (Albert et al., 2016). Early signs of the direct influence of SLR on estuaries’ water salinity are also emerging, for example, in the Delaware, USA, where Ross et al. (2015) estimate a rate of salinity increase by as much as 4.4 psu (Practical Salinity Unit) per metre of SLR since the 1950s.

Overall, while the literature suggests that it is still too early to attribute coastal impacts to SLR in most of the world’s coastal areas, there is ''very high confidence'' that as sea level continues to rise (Sections 4.2.3.2, 4.2.3.3), the frequency, severity and duration of hazards and related impacts increases (Woodruff et al., 2013; Lilai et al., 2016; Vitousek et al., 2017; Sections 4.2.3.4, 6.3.1.3). Detectable impacts and attributable impacts on shoreline behaviour are expected as soon as the second half of the 21st century (Nicholls and Cazenave, 2010; Storlazzi et al., 2018).

<div id="section-4-3-3-2submergence-and-flooding-of-coastal-areas"></div>

<span id="submergence-and-flooding-of-coastal-areas"></span>
==== 4.3.3.2 Submergence and Flooding of Coastal Areas ====

<div id="section-4-3-3-2submergence-and-flooding-of-coastal-areas-block-1"></div>

Since AR5, a number of continental and global scale coastal exposure studies have accounted for sub-national human dynamics such as coastward migration or coastal urbanisation. These studies project a population increase in the LECZ (coastal areas below 10 m of elevation) by 2100 of 85 to 239 million people as compared to only considering national dynamics (Merkens et al., 2016 <sup>[[#fn:r1176|1176]]</sup> ; Section 4.3.2). Under the five SSPs and without SLR, the population living in the LECZ increases from 640–700 million in 2000 to over one billion in 2050 under all SSPs, and then declines to 500–900 million in 2100 under all SSPs, except for SSP3 (i.e., a world in which countries will increasingly focus on domestic issues, or at best regional ones), for which the coastal population reaches 1.1–1.2 billion (Jones and O’Neill, 2016 <sup>[[#fn:r1177|1177]]</sup> ; Merkens et al., 2016 <sup>[[#fn:r1178|1178]]</sup> ).

The population exposed to mean and ESL events will grow significantly during the 21st century (high confidence) with socioeconomic development and SLR contributing roughly equally (medium confidence). Considering an average relative SLR of 0.7–0.9 m but no population growth, the number of people living below the hundred-year ESL in Latin America and the Caribbean will increase from 7.5 million in 2011 to 9 million by the end of the century (Reguero et al., 2015 <sup>[[#fn:r1179|1179]]</sup> ). Considering population growth and urbanisation, only 21 cm of global mean SLR by 2060 would increase the global population living below the hundred-year ESL from about 189 million in 2000 to 316–411 million in 2060, with the largest absolute changes in South and Southeast Asia and the largest relative changes in Africa (Neumann et al., 2015 <sup>[[#fn:r1180|1180]]</sup> ). Considering population growth, Hauer et al. (2016) <sup>[[#fn:r1181|1181]]</sup> estimate that 4.3 and 13.1 million people in the USA would live below the levels of 0.9 and 1.8 m SLR by 2100.

New coastal flood risk studies conducted since AR4 at global, continental and city scale, reinforce AR5 findings that if coastal societies do not adapt, flood risks will increase by 2–3 orders of magnitude reaching catastrophic levels by the end of the century, even under the lower end SLR expected under RCP2.6 (high confidence; Hinkel et al., 2014 <sup>[[#fn:r1182|1182]]</sup> ; Abadie et al., 2016 <sup>[[#fn:r1183|1183]]</sup> ; Diaz, 2016 <sup>[[#fn:r1184|1184]]</sup> ; Hunter et al., 2017 <sup>[[#fn:r1185|1185]]</sup> ; Lincke and Hinkel, 2018 <sup>[[#fn:r1186|1186]]</sup> ; Abadie, 2018 <sup>[[#fn:r1187|1187]]</sup> ; Brown et al., 2018a <sup>[[#fn:r1188|1188]]</sup> ; Nicholls, 2018 <sup>[[#fn:r1189|1189]]</sup> ). In combination, these studies take into account a SLR scenario range wider than the likely range of AR5 but consistent with the range of projections assessed in this report (Section 4.2.3.2). For example, considering 25–123 cm of SLR in 2100, all SSPs and no adaptation, Hinkel et al. (2014) find that 0.2–4.6% of global population is expected to be flooded annually in 2100, with expected annual damages (EAD) amounting to 0.3–9.3% of global GDP. Assessing 120 cities globally, Abadie (2018) find that under a weighted combination of the probabilistic scenarios, New Orleans and Guangzhou Guangdong rank highest with EAD above 1 trillion USD (not discounted) in each city. For Europe, EAD are expected to rise from 1.25 billion EUR today to 93–960 billion EUR by the end of the century (Vousdoukas et al., 2018b <sup>[[#fn:r1190|1190]]</sup> ). Already today, many small islands face large flood damages relative to their GDP specifically through TCs (Cashman and Nagdee, 2017 <sup>[[#fn:r1191|1191]]</sup> ) and under SLR EAD can reach up to several percent of GDP in 2100, as highlighted in AR5 (Wong et al., 2014 <sup>[[#fn:r1192|1192]]</sup> ). Similar to the exposure studies, estimates of future flood risk without considering adaptation, as presented in this paragraph, do not provide a meaningful characterisation of coastal flood risks, because adaptation and specifically hard protection is expected to be widespread during the 21st century in urban areas and cities (high confidence; Section 4.4.3.2.2). Rather, these estimates need to be seen as illustrations of the scale of adaptation needed to offset risk.

Flood risk studies that have included adaptation find that hard coastal protection is generally very effective in reducing flood risks during the 21st century even under high SLR scenarios (high confidence; Hinkel et al., 2014 <sup>[[#fn:r1193|1193]]</sup> ; Diaz, 2016 <sup>[[#fn:r1194|1194]]</sup> ; Brown et al., 2018a <sup>[[#fn:r1195|1195]]</sup> ; Hinkel et al., 2018 <sup>[[#fn:r1196|1196]]</sup> ; Lincke and Hinkel, 2018 <sup>[[#fn:r1197|1197]]</sup> ; Tamura et al., 2019 <sup>[[#fn:r1198|1198]]</sup> ) (Section 4.4.2.2.2). For example, Hinkel et al. (2014) find that under 25–123 cm of SLR in 2100 and all SSPs, hard coastal protection reduces the annual number of people affected by coastal floods and EAD by 2–3 orders of magnitude. Under high-end SLR and beyond the 21st century, effectiveness of coastal adaptation is expected to decline rapidly, but there is a lack of studies addressing this issue. Furthermore, there is a lack of studies taking into account responses beyond hard protection such as ecosystem-based adaptation, accommodation, advance and retreat (Sections 4.4.2).

Studies also confirm AR5 findings that the relative costs and benefits of coastal adaptation are distributed unequally across countries and regions (high confidence; Wong et al., 2014 <sup>[[#fn:r1199|1199]]</sup> ; Diaz, 2016 <sup>[[#fn:r1200|1200]]</sup> ; Lincke and Hinkel, 2018 <sup>[[#fn:r1201|1201]]</sup> ; Tamura et al., 2019 <sup>[[#fn:r1202|1202]]</sup> ). For example, while the median cost of protection and retreat under RCP8.5 in 2050 has been estimated to be under 0.09% of national GDP, large relative costs are found for small island states such as the Marshall Islands (7.6%), the Maldives (7.5%), Tuvalu (4.6%) and Kiribati (4.1%; Diaz, 2016 <sup>[[#fn:r1203|1203]]</sup> ). Furthermore, on a global average and for urban and densely populated regions, hard protection is highly cost efficient with benefit-cost ratios up to 104, but for poorer and less densely populated areas benefit-cost ratios are generally smaller than one (Lincke and Hinkel, 2018 <sup>[[#fn:r1204|1204]]</sup> ). Hence, without substantial transfer payments supporting poor areas, coastal flood risks will evolve unequally during this century, with richer and densely populated areas well protected behind hard structures and poorer less densely populated areas suffering losses and damages, and eventually retreating from the coast.

While continental to global scale flood exposure and risk studies have also explored a wider range of uncertainty as compared to AR5, much remains to be done. All of these studies rely on global elevation data, but few studies have explored the underlying bias. For example, for the Po delta in Italy, it was found that elevation data based on the widely used Shuttle Radar Topography Mission (SRTM), Reuter et al. (2007) overestimates the 100-year floodplain by about 50% as compared to local Lidar data (Wolff et al., 2016 <sup>[[#fn:r1205|1205]]</sup> ), while in the Ria Formosa region in Portugal SRTM underestimates EAD by up to 50% depending on the resampled resolution of the Lidar data (Vousdoukas et al., 2018a <sup>[[#fn:r1206|1206]]</sup> ). For the USA, SRTM data systemically underestimates population exposure below 3 m by more than 60% as compared to coastal Lidar data (Kulp and Strauss, 2016 <sup>[[#fn:r1207|1207]]</sup> ). A global scale comparison of major contributors to flood risk uncertainty finds that uncertainty in digital elevation data is roughly at equal footing with uncertainties in socioeconomic development, emission scenarios, and SLR in determining the magnitude of flood risks in the 21st century (Hinkel et al., 2014 <sup>[[#fn:r1208|1208]]</sup> ). At a European level, the number of people living in the 100-year coastal floodplain can vary between 20–70% depending on the different inundation models used and the inclusion or exclusion of wave set up (Vousdoukas, 2016 <sup>[[#fn:r1209|1209]]</sup> ). Comparing damage functions attained in different studies for European cities, Prahl et al. (2018 <sup>[[#fn:r1210|1210]]</sup> ) find up to four-fold differences in damages for floods above 3 m. Another major source of uncertainty relates to uncertainties in present-day ESL events due to the application of different extreme value methods (Wahl et al., 2017 <sup>[[#fn:r1211|1211]]</sup> ; Section 4.2.3.4). While all of the uncertainties reported above affected the actual size of exposure and flood risk figures, they do not affect the overall conclusions drawn here.

<div id="section-4-3-3-3coastal-erosion-and-projected-global-impacts-of-enhanced-erosion-on-human-systems"></div>

<span id="coastal-erosion-and-projected-global-impacts-of-enhanced-erosion-on-human-systems"></span>
==== 4.3.3.3 Coastal Erosion and Projected Global Impacts of Enhanced Erosion on Human Systems ====

<div id="section-4-3-3-3coastal-erosion-and-projected-global-impacts-of-enhanced-erosion-on-human-systems-block-1"></div>

Recent global assessments of coastal erosion indicate that land losses currently dominate over land gains and that human interventions are a major driver of shoreline changes (Cazenave and Cozannet, 2014 <sup>[[#fn:r1212|1212]]</sup> ; Luijendijk et al., 2018 <sup>[[#fn:r1213|1213]]</sup> ; Mentaschi et al., 2018 <sup>[[#fn:r1214|1214]]</sup> ). Luijendijk et al. (2018) <sup>[[#fn:r1215|1215]]</sup> estimate that over the 1984–2016 period, about a quarter of the world’s sandy beaches eroded at rates exceeding 0.5m yr–1 while about 28% accreted. While such global results can be challenged due to the relatively large detection threshold used (±0.5 m yr–1), there is growing literature indicating that coastal erosion is occurring or increasing, e.g. in the Arctic (Barnhart et al., 2014a <sup>[[#fn:r1216|1216]]</sup> ; Farquharson et al., 2018 <sup>[[#fn:r1217|1217]]</sup> ; Irrgang et al., 2019 <sup>[[#fn:r1218|1218]]</sup> ), Brazil (Amaro et al., 2015 <sup>[[#fn:r1219|1219]]</sup> ), China (Yang et al., 2017 <sup>[[#fn:r1220|1220]]</sup> ), Colombia (Rangel-Buitrago et al., 2015 <sup>[[#fn:r1221|1221]]</sup> ), India (Kankara et al., 2018 <sup>[[#fn:r1222|1222]]</sup> ), and along a large number of deltaic systems worldwide (e.g., Section 4.2.2.4).

Since AR5, however, there is growing appreciation and understanding of the ability of coastal systems to respond dynamically to SLR (Passeri et al., 2015 <sup>[[#fn:r1223|1223]]</sup> ; Lentz et al., 2016 <sup>[[#fn:r1224|1224]]</sup> ; Deng et al., 2017 <sup>[[#fn:r1225|1225]]</sup> ). Most low-lying coastal systems exhibit important feedbacks between biological and physical processes (e.g., Wright and Nichols, 2018), that have allowed them to maintain a relatively stable morphology under moderate rates of SLR (&lt;0.3 cm yr–1) over the past few millennia (Woodruff et al., 2013 <sup>[[#fn:r1226|1226]]</sup> ; Cross-Chapter Box 5 in Chapter 1). In a global review on multi-decadal changes in the land area of 709 atoll islands, Duvat (2019) <sup>[[#fn:r1227|1227]]</sup> shows that in a context of more rapid SLR than the global mean (Becker et al., 2012 <sup>[[#fn:r1228|1228]]</sup> ; Palanisamy et al., 2014 <sup>[[#fn:r1229|1229]]</sup> ), 73.1% of islands were stable in area, while respectively 15.5% and 11.4% increased and decreased in size. While anthropogenic drivers played a major role, especially in urban islands (e.g., shoreline stabilisation by coastal defences, increase in island size as a result of reclamation works), this study and others (e.g., McLean and Kench, 2015) suggest that these islands have had the capacity to maintain their land area by naturally adjusting to SLR over the past decades (high confidence). However, it has been argued that this capacity could be reduced in the coming decades, due to the combination of higher rates of SLR, increased wave energy (Albert et al., 2016 <sup>[[#fn:r1230|1230]]</sup> ), changes in run-up (Shope et al., 2017 <sup>[[#fn:r1231|1231]]</sup> ) and storm wave direction (Harley et al., 2017 <sup>[[#fn:r1232|1232]]</sup> ), effects of ocean warming and acidification on critical ecosystems such as coral reefs (Section 4.3.3.5.2), and a continued increase in anthropogenic pressure.

From a global scale perspective, based on AR4 SLR scenarios and without considering the potential benefits of adaptation, Hinkel et al. (2013b) estimate that about 6000 to 17,000 km2 of land is expected to be lost during the 21st century due to enhanced coastal erosion associated with SLR, in combination with other drivers. This could lead to a displacement of 1.6–5.3 million people and associated cumulative costs of 300 to 1000 billion USD (Section 4.4.3.5). Importantly, these global figures mask the wide diversity of local situations; and some literature is emerging on the non-physical and non-quantifiable impacts of coastal erosion, for example, on the loss of recreational grounds and the induced risks to the associated social dimensions (i.e., how local communities experience coastal erosion impacts; Karlsson et al., 2015 <sup>[[#fn:r1233|1233]]</sup> ).

<div id="section-4-3-3-4salinisation"></div>

<span id="salinisation"></span>
==== 4.3.3.4 Salinisation ====

<div id="section-4-3-3-4salinisation-block-1"></div>

With rising sea levels, saline water intrusion into coastal aquifers and surface waters and soils is expected to be more frequent and enter farther landwards. Salinisation of groundwater, surface water and soil resources also increases with land-based drought events, decreasing river discharges in combination with water extraction and SLR ''(high confidence).''

<div id="section-4-3-3-4salinisation-block-2"></div>

<span id="coastal-aquifers-and-groundwater-lenses"></span>
===== 4.3.3.4.1 Coastal aquifers and groundwater lenses =====

<div id="section-4-3-3-4salinisation-block-3"></div>

Groundwater volumes will primarily be affected by variations in precipitation patterns (Taylor et al., 2013; Jiménez Cisneros et al., 2014), which are expected to increase water stress in small islands (Holding et al., 2016). While SLR will mostly impact groundwater quality (Bailey et al., 2016) and in turn exacerbate salinisation induced by marine flooding events (Gingerich et al., 2017), it will also affect the watertable height (Rotzoll and Fletcher, 2013; Jiménez Cisneros et al., 2014; Masterson et al., 2014; Werner et al., 2017). In addition, the natural migration of groundwater lenses inland in response to SLR can also be severely constrained by urbanisation, for example, in semi-arid South Texas, USA (Uddameri et al., 2014).

These changes will affect both freshwater availability (for drinking water supply and agriculture) and vegetation dynamics. At many locations, however, direct anthropogenic influences, such as groundwater pumping for agricultural or urban uses, already impact salinisation of coastal aquifers more strongly than what is expected from SLR in the 21st century (Ferguson and Gleeson, 2012; Jiménez Cisneros et al., 2014; Uddameri et al., 2014), with trade-offs in terms of groundwater depletion that may contribute to anthropogenic subsidence and thus increase coastal flood risk. Recent studies also suggest that the influence of land-surface inundation on seawater intrusion and resulting salinisation of groundwater lenses on small islands has been underestimated until now (Ataie-Ashtiani et al., 2013; Ketabchi et al., 2014). Such impacts will potentially also combine with a projected drying of most of the tropical-to-temperate islands by mid-century (Karnauskas et al., 2016).

<div id="section-4-3-3-4salinisation-block-4"></div>

<span id="surface-waters"></span>
===== 4.3.3.4.2 Surface waters =====

<div id="section-4-3-3-4salinisation-block-5"></div>

The quality of surface water resources (in estuaries, rivers, reservoirs, etc.) can be affected by the intrusion of saline water, both in a direct (increased salinity) and indirect way (altered environmental conditions which change the behaviour of pollutants and microbes). In terms of direct impacts, statistical models and long-term (1950 to present) records of salinity show significant upward trends in salinity and a positive correlation between rising sea levels and increasing residual salinity, for example in the Delaware Estuary, USA (Ross et al., 2015). Higher salinity levels, further inland, have also been reported in the Gorai river basin, southwestern Bangladesh (Bhuiyan and Dutta, 2012), and in the Mekong Delta, Vietnam. In the Mekong Delta for instance, salinity intrusion extends around 15 km inland during the rainy season and typically around 50 km during dry season (Gugliotta et al., 2017). Importantly, salinity intrusion in these deltas is caused by a variety of factors such as changes in discharge and water abstraction along with relative SLR. More broadly, the impact of salinity intrusion can be significant in river deltas or low-lying wetlands, especially during low-flow periods such as in the dry season (Dessu et al., 2018). In Bangladesh, for instance, some freshwater fish species are expected to lose their habitat with increasing salinity, with profound consequences on fish-dependent communities (Dasgupta et al., 2017). In the Florida Coastal Everglades, sea level increasingly exceeds ground surface elevation at the most downstream freshwater sites, affecting marine-to-freshwater hydrologic connectivity and transport of salinity and phosphorous upstream from the Gulf of Mexico. The impact of SLR is higher in the dry season when there is practically no freshwater inflow (Dessu et al., 2018). Salinity intrusion was shown to cause shifts in the diatom assemblages, with expected cascading effects through the food web (Mazzei and Gaiser, 2018). Salinisation of surface water may lead to limitations in drinking water supply (Wilbers et al., 2014), as well as to future fresh water shortage in reservoirs, for example in Shanghai (Li et al., 2015). Salinity changes the partitioning and mobility of some metals, and hence their concentration or speciation in the water bodies (Noh et al., 2013; Wong et al., 2015; de Souza Machado et al., 2018). Varying levels of salinity also influence the abundance and toxicity of ''Vibrio cholerae'' in the Ganges Delta (Batabyal et al., 2016).

<div id="section-4-3-3-4salinisation-block-6"></div>

<span id="soils"></span>
===== 4.3.3.4.3 Soils =====

<div id="section-4-3-3-4salinisation-block-7"></div>

Salinisation is one of the major drivers of soil degradation, with sea water intrusion being one of the common causes (Daliakopoulos et al., 2016). In a study in the Ebro Delta, Spain, for instance, soil salinity was shown to be directly related to distances to the river, to the delta inner border, and to the old river mouth (Genua-Olmedo et al., 2016). Land elevation was the most important variable in explaining soil salinity.

SLR was also shown to decrease organic carbon (C <sub>org</sub> ) concentrations and stocks in sediments of salt marshes as reworked marine particles contribute with a lower amount of C <sub>org</sub> than terrigenous sediments. C <sub>org</sub> accumulation in tropical salt marshes can be as high as in mangroves and the reduction of C <sub>org</sub> stocks by ongoing SLR might cause high CO <sub>2</sub> releases (Ruiz-Fernández et al., 2018). In many cases attribution to SLR is missing, but, independent from clear attribution, sea water intrusion leads to a salinisation of exposed soils with changes in carbon dynamics (Ruiz-Fernández et al., 2018) and microbial communities (Sánchez-Rodríguez et al., 2017), soil enzyme activity and metal toxicity (Zheng et al., 2017). Water salinity levels in the pores of coastal marsh soils can become significantly elevated in just one week of flooding by sea water, which can potentially negatively impact associated microbial communities for significantly longer time periods (McKee et al., 2016). SLR will also alter the frequency and magnitude of wet/dry periods and salinity levels in coastal ecosystems, with consequences for the formation of climate relevant GHGs (Liu et al., 2017b) and therefore feedbacks to the climate.

Soil salinisation affects agriculture directly with impacts on plant germination (Sánchez-García et al., 2017), plant biomass (rice and cotton) production (Yao et al., 2015), and yield (Genua-Olmedo et al., 2016). Impact on agriculture is especially relevant in low-lying coastal areas where agricultural production is a major land use, such as in river deltas.

<div id="section-4-3-3-5ecosystems-and-ecosystem-services"></div>

<span id="ecosystems-and-ecosystem-services"></span>
==== 4.3.3.5 Ecosystems and Ecosystem Services ====

<div id="section-4-3-3-5ecosystems-and-ecosystem-services-block-1"></div>

<span id="tidal-wetlands"></span>
===== 4.3.3.5.1 Tidal wetlands =====

<div id="section-4-3-3-5ecosystems-and-ecosystem-services-block-2"></div>

Global coastal wetlands have been reduced by a half since the pre-industrial period due to the impacts of both climatic and non-climatic drivers such as flooding, coastal urbanisation, alterations in drainage and sediment supply. (Sections 4.3.2.3, 5.3.2). Potentially one of the most important of the eco-morphodynamic feedback allowing for relatively stable morphology under SLR is the ability of marsh and mangrove systems to enhance the trapping of sediment, which in turn allows tidal wetlands to grow and increase the production and accumulaGlobal coastal wetlands have been reduced by a half since the pre-industrial period due to the impacts of both climatic and non-climatic drivers such as flooding, coastal urbanisation, alterations in drainage and sediment supply. (Sections 4.3.2.3, 5.3.2). Potentially one of the most important of the eco-morphodynamic feedbacks allowing for relatively stable morphology under SLR is the ability of marsh and mangrove systems to enhance the trapping of sediment, which in turn allows tidal wetlands to grow and increase the production and accumulation of organic material (Kirwan and Megonigal, 2013 <sup>[[#fn:r1272|1272]]</sup> ). When ecosystem health is maintained and sufficient sediment exists to support their growth, this particular feedback has generally allowed marshes and mangrove systems to build vertically at rates equal to or greater than SLR up to the present day (Kirwan et al., 2016 <sup>[[#fn:r1273|1273]]</sup> ; Woodroffe et al., 2016 <sup>[[#fn:r1274|1274]]</sup> ).

While recent reviews suggest that mangroves’ surface accretion rate will keep pace with a high SLR scenario (RCP8.5) up to years 2055 and 2070 in fringe and basin mangrove settings, respectively (Sasmito et al., 2016 <sup>[[#fn:r1275|1275]]</sup> ), process-based models of vertical marsh growth that incorporate biological and physical feedbacks support survival under rates of SLR as high as 1–5 cm yr–1 before drowning (Kirwan et al., 2016 <sup>[[#fn:r1276|1276]]</sup> ). Threshold rates of SLR before marsh drowning however vary significantly from site-to-site and can be substantially lower than 1 cm yr–1 in micro-tidal regions where the tidal trapping of sediment is reduced and/or in areas with low sediment availability (Lovelock et al., 2015 <sup>[[#fn:r1277|1277]]</sup> ; Ganju et al., 2017 <sup>[[#fn:r1278|1278]]</sup> ; Jankowski et al., 2017 <sup>[[#fn:r1279|1279]]</sup> ; Watson et al., 2017 <sup>[[#fn:r1280|1280]]</sup> ). Global environmental change may also to lead to changes in growth rates, productivity and geographic distribution of different mangrove and marsh species, including the replacement of environmentally sensitive species by those possessing greater climatic tolerance (Krauss et al., 2014 <sup>[[#fn:r1281|1281]]</sup> ; Reef and Lovelock, 2014 <sup>[[#fn:r1282|1282]]</sup> ; Coldren et al., 2019 <sup>[[#fn:r1283|1283]]</sup> ). Processes impacting lateral changes at the marsh boundary including wave erosion are just as important, if not more, than vertical accretion rates in determining coastal wetland survival (e.g., Mariotti and Carr, 2014). For most low-lying coastlines, a seaward loss of wetland area due to marsh retreat could be offset by a similar landward migration of coastal wetlands (Kirwan and Megonigal, 2013 <sup>[[#fn:r1284|1284]]</sup> ; Schile et al., 2014 <sup>[[#fn:r1285|1285]]</sup> ), this landward migration having the potential to maintain and even increase the extent of coastal wetlands globally (Morris et al., 2012 <sup>[[#fn:r1286|1286]]</sup> ; Kirwan et al., 2016 <sup>[[#fn:r1287|1287]]</sup> ; Schuerch et al., 2018 <sup>[[#fn:r1288|1288]]</sup> ). This natural process will however be constrained in areas with steep topography or hard engineering structures (i.e., coastal squeeze, Section 4.3.2.4). Seawalls, levees and dams can also prevent the fluvial and marine transport of sediment to wetland areas and reduce their resilience further (Giosan, 2014 <sup>[[#fn:r1289|1289]]</sup> ; Tessler et al., 2015 <sup>[[#fn:r1290|1290]]</sup> ; Day et al., 2016 <sup>[[#fn:r1291|1291]]</sup> ; Spencer et al., 2016 <sup>[[#fn:r1292|1292]]</sup> ). When ecosystem health is maintained and sufficient sediment exists to support their growth, this particular feedback has generally allowed marshes and mangrove systems to build vertically at rates While recent reviews suggest that mangroes’ surface accretion rate will keep pace with a high SLR scenario (RCP8.5) up to years 2055 and 2070 in fringe and basin mangrove settings, respectively (Sasmito et al., 2016 <sup>[[#fn:r1275|1275]]</sup> ), process-based models of vertical marsh growth that incorporate biological and physical feedbacks support survival under rates of SLR as high as 1–5 cm yr–1 before drowning (Kirwan et al., 2016 <sup>[[#fn:r1276|1276]]</sup> ). Threshold rates of SLR before marsh drowning however vary significantly from site-to-site and can be substantially lower than 1 cm yr–1 in micro-tidal regions where the tidal trapping of sediment is reduced and/or in areas with low sediment availability (Lovelock et al., 2015 <sup>[[#fn:r1277|1277]]</sup> ; Ganju et al., 2017 <sup>[[#fn:r1278|1278]]</sup> ; Jankowski et al., 2017 <sup>[[#fn:r1279|1279]]</sup> ; Watson et al., 2017 <sup>[[#fn:r1280|1280]]</sup> ). Global environmental change may also to lead to changes in growth rates, productivity and geographic distribution of different mangrove and marsh species, including the replacement of environmentally sensitive species by those possessing greater climatic tolerance (Krauss et al., 2014 <sup>[[#fn:r1281|1281]]</sup> ; Reef and Lovelock, 2014 <sup>[[#fn:r1282|1282]]</sup> ; Coldren et al., 2019 <sup>[[#fn:r1283|1283]]</sup> ). Processes impacting lateral changes at the marsh boundary including wave erosion are just as important, if not more, than vertical accretion rates in determining coastal wetland survival (e.g., Mariotti and Carr, 2014). For most low-lying coastlines, a seaward loss of wetland area due to marsh retreat could be offset by a similar landward migration of coastal wetlands (Kirwan and Megonigal, 2013 <sup>[[#fn:r1284|1284]]</sup> ; Schile et al., 2014 <sup>[[#fn:r1285|1285]]</sup> ), this landward migration having the potential to maintain and even increase the extent of coastal wetlands globally (Morris et al., 2012 <sup>[[#fn:r1286|1286]]</sup> ; Kirwan et al., 2016 <sup>[[#fn:r1287|1287]]</sup> ; Schuerch et al., 2018 <sup>[[#fn:r1288|1288]]</sup> ). This natural process will however be constrained in areas with steep topography or hard engineering structures (i.e., coastal squeeze, Section 4.3.2.4). Seawalls, levees and dams can also prevent the fluvial and marine transport of sediment to wetland areas and reduce their resilience further (Giosan, 2014 <sup>[[#fn:r1289|1289]]</sup> ; Tessler et al., 2015 <sup>[[#fn:r1290|1290]]</sup> ; Day et al., 2016 <sup>[[#fn:r1291|1291]]</sup> ; Spencer et al., 2016 <sup>[[#fn:r1292|1292]]</sup> ).

<div id="section-4-3-3-5ecosystems-and-ecosystem-services-block-4"></div>

<span id="coral-reefs"></span>
===== 4.3.3.5.2 Coral reefs =====

<div id="section-4-3-3-5ecosystems-and-ecosystem-services-block-5"></div>

Coral reefs are considered to be the marine ecosystem most threatened by climate-related ocean change, especially ocean warming and acidification, even under an RCP2.6 scenario (Gattuso et al., 2015 <sup>[[#fn:r1293|1293]]</sup> ; Albright et al., 2018 <sup>[[#fn:r1294|1294]]</sup> ; Hoegh-Guldberg et al., 2018 <sup>[[#fn:r1295|1295]]</sup> ; Díaz et al., 2019 <sup>[[#fn:r1296|1296]]</sup> ; Section 5.3.4). AR5 concluded that ‘a number of coral reefs could […] keep up with the maximum rate of SLR of 15.1 mm yr–1 projected for the end of the century […] (medium confidence) [but a future net accretion rate lower] than during the Holocene (Perry et al., 2013 <sup>[[#fn:r1297|1297]]</sup> ) and increased turbidity (Storlazzi et al., 2011 <sup>[[#fn:r1298|1298]]</sup> ) will weaken this capability (very high confidence)’ (Wong et al., 2014: 379 <sup>[[#fn:r1299|1299]]</sup> ). Subsequently, some studies suggested that SLR may have negligible impacts on coral reefs’ vertical growth because the projected rate and magnitude of SLR by 2100 are within the potential accretion rates of most coral reefs (van Woesik et al., 2015 <sup>[[#fn:r1300|1300]]</sup> ). Other studies, however, stressed that the overall net vertical accretion of reefs may decrease after the first 30 years of rise in a 1.2 m SLR scenario (Hamylton et al., 2014 <sup>[[#fn:r1301|1301]]</sup> ), and that most reefs will not be able to keep up with SLR under RCP4.5 and beyond (Perry et al., 2018 <sup>[[#fn:r1302|1302]]</sup> ). The SR1.5 also concludes that coral reefs ‘are projected to decline by a further 70–90% at 1.5°C (high confidence) with larger losses (&gt;99%) at 2°C (very high confidence)’ (Hoegh-Guldberg et al., 2018: 10 <sup>[[#fn:r1303|1303]]</sup> ). A key point is that SLR will not act in isolation of other drivers. Cumulative impacts, including anthropogenic drivers, are estimated to reduce the ability of coral reefs to keep pace with future SLR (Hughes et al., 2017 <sup>[[#fn:r1304|1304]]</sup> ; Yates et al., 2017 <sup>[[#fn:r1305|1305]]</sup> ) and thereby reduce the capacity of reefs to provide sediments and protection to coastal areas. For example, the combination of reef erosion due to acidification and human-induced mechanical destruction is altering seafloor topography, increasing risks from SLR in carbonate sediment dominated regions (Yates et al., 2017 <sup>[[#fn:r1306|1306]]</sup> ). Both ocean acidification (Albright et al., 2018 <sup>[[#fn:r1307|1307]]</sup> ; Eyre et al., 2018 <sup>[[#fn:r1308|1308]]</sup> ) and ocean warming (Perry and Morgan, 2017 <sup>[[#fn:r1309|1309]]</sup> ) have been considered to slow future growth rates and reef accretion (Section 5.3.4). Recent literature also shows that alterations of coral reef 3D structure from changes in growth, breakage, disease or acidification can profoundly affect their ability to buffer waves impacts (through wave breaking and wave energy damping), and therefore keep-up with SLR (Yates et al., 2017 <sup>[[#fn:r1310|1310]]</sup> ; Harris et al., 2018 <sup>[[#fn:r1311|1311]]</sup> ). Such prospects contribute to raise concerns about the future ability of atoll islands to adjust naturally to SLR and persist (Section 4.3.3.3, Cross-Chapter Box 9). Another concern is that locally, even minimal SLR can increase turbidity on fringing reefs, reducing light and, therefore, photosynthesis and calcification. SLR-induced turbidity can be caused by increased coastal erosion and the transfer of sediment to nearby reefs and enhanced sediment resuspension (Field et al., 2011 <sup>[[#fn:r1312|1312]]</sup> ).

<div id="section-4-3-3-5ecosystems-and-ecosystem-services-block-6"></div>

<span id="seagrasses"></span>
===== 4.3.3.5.3 Seagrasses =====

<div id="section-4-3-3-5ecosystems-and-ecosystem-services-block-7"></div>

Due to their natural capacity to enhance accretion and in the absence of mechanical or chemical destruction by human activities, seagrasses are not expected to be severely affected by SLR, except indirectly through the increase of the impacts of extreme weather events and waves on coastal morphology (i.e., erosion) as well as through changes in light levels and through effects on adjacent ecosystems (Saunders et al., 2013 <sup>[[#fn:r1313|1313]]</sup> ). Extreme ﬂooding events have also been shown to cause large-scale losses of seagrass habitats (Bandeira and Gell, 2003 <sup>[[#fn:r1314|1314]]</sup> ), for example seagrasses in Queensland, Australia, were lost in a disastrous ﬂooding event (Campbell and McKenzie, 2004 <sup>[[#fn:r1315|1315]]</sup> ). Changes in ocean currents can have either positive or negative effects on seagrasses, creating new space for seagrasses to grow or eroding seagrass beds (Bjork et al., 2008 <sup>[[#fn:r1316|1316]]</sup> ). But overall, seagrass will primarily be negatively affected by the direct effects of increased sea temperature on growth rates and the occurrence of disease (Marba and Duarte, 2010 <sup>[[#fn:r1317|1317]]</sup> ; Burge et al., 2013 <sup>[[#fn:r1318|1318]]</sup> ; Koch et al., 2013 <sup>[[#fn:r1319|1319]]</sup> ; Thompson et al., 2015 <sup>[[#fn:r1320|1320]]</sup> ; Chefaoui et al., 2018 <sup>[[#fn:r1321|1321]]</sup> ; Gattuso et al., 2018 <sup>[[#fn:r1322|1322]]</sup> ; Section 5.3.2) as well as by heavy rains that may dilute the seawater to a lower salinity. Such impacts will be exacerbated by major causes of seagrass decline including coastal eutrophication, siltation and coastal development (Waycott et al., 2009 <sup>[[#fn:r1323|1323]]</sup> ). Noteworthy is that some positive impacts are expected, as ocean acidification is expected to benefit photosynthesis and growth rates of seagrass (Repolho et al., 2017 <sup>[[#fn:r1324|1324]]</sup> ).

<div id="section-4-3-3-5ecosystems-and-ecosystem-services-block-8"></div>

<span id="coastal-protection-by-coastal-and-marine-ecosystems"></span>
===== 4.3.3.5.4 Coastal protection by coastal and marine ecosystems =====

<div id="section-4-3-3-5ecosystems-and-ecosystem-services-block-9"></div>

Major ‘protection’ benefits derived from the above-mentioned coastal ecosystems include wave attenuation and shoreline stabilisation, for example, by coral reefs (Elliff and Silva, 2017 <sup>[[#fn:r1325|1325]]</sup> ; Siegle and Costa, 2017 <sup>[[#fn:r1326|1326]]</sup> ), mangroves (Zhang et al., 2012 <sup>[[#fn:r1327|1327]]</sup> ; Barbier, 2016 <sup>[[#fn:r1328|1328]]</sup> ; Menéndez et al., 2018 <sup>[[#fn:r1329|1329]]</sup> ) or salt marshes (Möller et al., 2014 <sup>[[#fn:r1330|1330]]</sup> ; Hu et al., 2015 <sup>[[#fn:r1331|1331]]</sup> ). Recently, a global meta-analysis of 69 studies demonstrated that, on average, these ecosystems together reduced wave heights between 35–71% at the limited locations considered (Narayan et al., 2016 <sup>[[#fn:r1332|1332]]</sup> ), with coral reefs, salt marshes, mangroves and seagrass/kelp beds reducing wave heights by 54–81%, 62–79%, 25–37% and 25–45% respectively (see Narayan et al., 2016 for map of locations considered). Additional studies suggest greater wave attenuation in mangrove systems (Horstman et al., 2014 <sup>[[#fn:r1333|1333]]</sup> ), and highlight broader complexities in wave attenuation related to total tidal wetland extent, water depth, and species. Global analyses show that natural and artificial seagrasses can attenuate wave height and energy by as much as 40% and 50%, respectively (Fonseca and Cahalan, 1992 <sup>[[#fn:r1334|1334]]</sup> ; John et al., 2015 <sup>[[#fn:r1335|1335]]</sup> ), while coral reefs have been observed to reduce total wave energy by 94–98% (n = 13; Ferrario et al., 2014 <sup>[[#fn:r1336|1336]]</sup> ) and wave driven flooding volume by 72% (Beetham et al., 2017 <sup>[[#fn:r1337|1337]]</sup> ). In addition, storm surge attenuation based on a recent literature review by Stark et al. (2015) <sup>[[#fn:r1338|1338]]</sup> range from -2–25 cm km <sup>–1</sup> length of marsh, where the negative value denotes actual amplification. Other ecosystems provide coastal protection, including macroalgae, oyster and mussel beds, and also beaches, dunes and barrier islands, but there is less understanding of the level of protection conferred by these other organisms and habitats (Spalding et al., 2014 <sup>[[#fn:r1339|1339]]</sup> ).

While there is little literature on the extent to which SLR specifically will affect coastal protection by coastal and marine ecosystems, it is estimated that SLR may reduce this ecosystem service ( ''limited evidence, high agreement'' ) through the above-described impacts on the ecosystems themselves, and in combination with the impacts of other climate-related changes to the ocean (e.g., ocean warming and acidification; Sections 5.3.1 to 5.3.6, 5.4.1). Wave attenuation by coral reefs, for example, is estimated to be negatively affected in the near future by changes in coral reefs’ structural complexity more than by SLR (Harris et al., 2018 <sup>[[#fn:r1340|1340]]</sup> ); changes in mean and ESL events will rather add a layer of stress. Beck et al. (2018) estimate that under RCP8.5 by 2100, a 1 m loss in coral reefs’ height will increase the global area flooded under a 100-year storm event by 116% compared to today, against +66% with no reef loss.

<div id="section-4-3-3-6human-activities"></div>

<span id="human-activities"></span>
==== 4.3.3.6 Human Activities ====

<div id="section-4-3-3-6human-activities-block-1"></div>

<span id="coastal-agriculture"></span>
===== 4.3.3.6.1 Coastal agriculture =====

SLR will affect agriculture mainly through land submergence, soil and fresh groundwater resources salinisation, and land loss due to permanent coastal erosion, with consequences on production, livelihood diversification and food security, especially in heavily coastal agriculture-dependent countries such as Bangladesh (Khanom, 2016 <sup>[[#fn:r13|13]]</sup> 41). Recent literature confirms that salinisation is already a major problem for traditional agriculture in deltas (Wong et al., 2014 <sup>[[#fn:r1342|1342]]</sup> ; Khai et al., 2018 <sup>[[#fn:r1343|1343]]</sup> ) and low-lying island nations where some edible cultivated plants such as taro patches are threatened (Nunn et al., 2017b <sup>[[#fn:r1344|1344]]</sup> ). Taking the case of rice cultivation, recent works emphasise the prevailing role of combined surface elevation and soil salinity, such as in the Mekong delta (Vietnam; Smajgl et al., 2015 <sup>[[#fn:r1345|1345]]</sup> ) and in the Ebro delta (Spain; Genua-Olmedo et al., 2016 <sup>[[#fn:r1346|1346]]</sup> ), estimating for the latter a decrease in the rice production index from 61.2% in 2010 to 33.8% by 2100 in a 1.8 m SLR scenario. For seven wetland species occurring in coastal freshwater marshes in central Veracruz on the Gulf of Mexico, an increase in salinity was shown to affect the germination process under wetland salt intrusion (Sánchez-García et al., 2017 <sup>[[#fn:r1347|1347]]</sup> ). In coastal Bangladesh, oilseed, sugarcane and jute cultivation was reported to be already discontinued due to challenges to cope with current salinity levels (Khanom, 2016 <sup>[[#fn:r1348|1348]]</sup> ), and salinity is projected to have an unambiguously negative influence on all dry-season crops over the next 15–45 years (especially in the southwest; Clarke et al., 2018 <sup>[[#fn:r1349|1349]]</sup> ; Kabir et al., 2018 <sup>[[#fn:r1350|1350]]</sup> ). Salinity intrusion and salinisation can trigger land use changes towards brackish or saline aquaculture such as shrimp or rice-shrimp systems with impacts on environment, livelihoods and income stability (Renaud et al., 2015 <sup>[[#fn:r1351|1351]]</sup> ). However, increasing salinity is only one of the land use change drivers along with, for example, policy changes and market prices at the household level (Renaud et al., 2015 <sup>[[#fn:r1352|1352]]</sup> ).

<div id="section-4-3-3-6human-activities-block-2"></div>

<span id="coastal-tourism-and-recreation"></span>
===== 4.3.3.6.2 Coastal tourism and recreation =====

SLR may significantly affect tourism and recreation through impacts on landscapes (e.g., beaches), cultural features (e.g., Marzeion and Levermann, 2014; Fang et al., 2016 <sup>[[#fn:r1353|1353]]</sup> ), and critical transportation infrastructures such as harbours and airports (Monioudi et al., 2018 <sup>[[#fn:r1354|1354]]</sup> ). Coastal areas’ future tourism and recreation attractiveness will however also depend on changes in air temperature, seasonality and sea surface temperature (including induced effects such as invasive species, e.g., jellyfishes, and disease spreading; Burge et al., 2014 <sup>[[#fn:r1355|1355]]</sup> ; Weatherdon et al., 2016 <sup>[[#fn:r1356|1356]]</sup> ; Hoegh-Guldberg et al., 2018 <sup>[[#fn:r1357|1357]]</sup> ; Section 5.4.2). Future changes in climatic conditions in tourists’ areas of origin will also play a role in reshaping tourism flows (Bujosa and Rosselló, 2013 <sup>[[#fn:r1358|1358]]</sup> ; Amelung and Nicholls, 2014 <sup>[[#fn:r1359|1359]]</sup> ), in addition to mitigation policies on air transportation, non-climatic features (e.g., accommodation and travel prices) and tourists’ and tourism developers’ perceptions of climate-related changes (Shakeela and Becken, 2015 <sup>[[#fn:r1360|1360]]</sup> ). Since AR5, forecasting the consequences of climate change effects on global-to-local tourism flows has remained challenging (Rosselló-Nadal, 2014 <sup>[[#fn:r1361|1361]]</sup> ; Wong et al., 2014 <sup>[[#fn:r1362|1362]]</sup> ; Hoegh-Guldberg et al., 2018 <sup>[[#fn:r1363|1363]]</sup> ). There are also concerns about the effect of SLR on tourism facilities, for example hotels in Ghana (Sagoe-Addy and Addo, 2013 <sup>[[#fn:r1364|1364]]</sup> ), in a context where tourism infrastructure often contributes to the degradation of natural buffering environments through, for example, coastal squeeze (e.g., Section 4.3.2.4) and human-driven coastal erosion. Again, forecasting is constrained by the lack of scientific studies on tourism stakeholders’ long-term strategies and adaptive capacity (Hoogendoorn and Fitchett, 2018 <sup>[[#fn:r1365|1365]]</sup> ).

<div id="section-4-3-3-6human-activities-block-3"></div>

<span id="coastal-fisheries-and-aquaculture"></span>
===== 4.3.3.6.3 Coastal fisheries and aquaculture =====

Recent studies support the AR5 conclusion that ocean warming and acidification are considered more influential drivers of change in fisheries and aquaculture than SLR (Larsen et al., 2014; Nurse et al., 2014 <sup>[[#fn:r1366|1366]]</sup> ; Wong et al., 2014 <sup>[[#fn:r1367|1367]]</sup> ). The negative effects of SLR on fisheries and aquaculture are indirect, through adverse impacts on habitats (e.g., coral reef degradation, reduced water quality in deltas and estuarine environments, soil salinisation, etc.), as well as on facilities (e.g., damage to small and large harbours). This makes future projections on SLR implications for coastal and marine fisheries and aquaculture an understudied field of research. Conclusions only state that future impacts will be highly context-specific due to local manifestations of SLR and local fishery-dependent communities’ ability to adapt to alterations in fish and aquaculture conditions and productivity (Hollowed et al., 2013 <sup>[[#fn:r1368|1368]]</sup> ; Weatherdon et al., 2016 <sup>[[#fn:r1369|1369]]</sup> ). Salinity intrusion also contributes to conversion of land or freshwater ponds to brackish or saline aquaculture in many low-lying coastal areas of Southeast Asia such as in the Mekong Delta in Vietnam (Renaud et al., 2015 <sup>[[#fn:r1370|1370]]</sup> ).

<div id="section-4-3-3-6human-activities-block-4"></div>

<span id="social-values"></span>
===== 4.3.3.6.4 Social values =====

Social values refer to what people consider of critical importance about the places in which they live, and range from material to immaterial things (assets, beliefs, etc.; Hurlimann et al., 2014 <sup>[[#fn:r1371|1371]]</sup> ; Rouse et al., 2017 <sup>[[#fn:r1372|1372]]</sup> ). Consideration of social values offers an opportunity to address a wider perspective on impacts on human systems, for example, complementary to quantitative assessments of health impacts (e.g., loss of source of calories and food insecurity; Keim, 2010 <sup>[[#fn:r1374|1374]]</sup> ). This also encompasses immaterial dimensions, such as threats to cultural heritage (Marzeion and Levermann, 2014 <sup>[[#fn:r1374|1374]]</sup> ; Fatorić and Seekamp, 2017a <sup>[[#fn:r1375|1375]]</sup> ), socialising activities (Karlsson et al., 2015 <sup>[[#fn:r1376|1376]]</sup> ), integration of marginalised groups (Maldonado, 2015 <sup>[[#fn:r1377|1377]]</sup> ) and cultural ecosystem services (Fish et al., 2016), and provides an opportunity to better reflect context-specificities in valuing the physical/ecological/human/cultural impacts’ importance for and distribution within a given society (Fatorić and Seekamp, 2017b <sup>[[#fn:r1379|1379]]</sup> ). This field of research (no detailed mention found in AR5) is just emerging due to the transdisciplinary and qualitative nature of the topic. Graham et al. (2013) <sup>[[#fn:r1380|1380]]</sup> advance a 5-category framing of social values specifically at risk from SLR: health (i.e., the social determinants of survival such as environmental and housing quality and healthy lifestyles), feeling of safety (e.g., financial and job security), belongingness (i.e., attachment to places and people), self-esteem (e.g., social status or pride that can be affected by coastal retreat), and self-actualisation (i.e., people’s efforts to define their own identity). Another emerging issue relates to social values at risk due to land submergence in low-lying islands (Yamamoto and Esteban, 2014 <sup>[[#fn:r1381|1381]]</sup> ) and parts of countries and individual properties (Marino, 2012 <sup>[[#fn:r1382|1382]]</sup> ; Maldonado et al., 2013 <sup>[[#fn:r1383|1383]]</sup> ; Aerts, 2017 <sup>[[#fn:r1384|1384]]</sup> ; Allgood and McNamara, 2017 <sup>[[#fn:r1385|1385]]</sup> ). Recent studies also highlight the potential additional risks to social values in areas where displaced people relocate (Davis et al., 2018 <sup>[[#fn:r1386|1386]]</sup> ).

<span id="conclusion-on-coastal-risk-reasons-for-concern-and-future-risks"></span>
=== 4.3.4 Conclusion on Coastal Risk: Reasons for Concern and Future Risks ===

<div id="section-4-3-4-conclusion-on-coastal-risk-reasons-for-concern-and-future-risks-block-1"></div>

SLR projections for the 21st century, together with other ocean related changes (e.g., acidification and warming) and the possible increase in human-driven pressures at the coast (e.g., demographic and settlement patterns), make low-lying islands, coasts and communities relevant illustrations of some of the five Reasons for Concern (RFCs) developed by the IPCC since the Third Assessment Report (McCarthy et al., 2001; Smith et al., 2001) to assess risks from a global perspective. The AR5 Synthesis Report (IPCC, 2014) as well as the more recent SR1.5 (Hoegh-Guldberg et al., 2018) refined the RFC approach. The AR5 Synthesis Report (IPCC, 2014) developed two additional RFCs related to the coasts, subsequently updated along with the other RFCs (O’Neill et al., 2017) . One refers to risks to marine species arising from ocean acidification, and the other one refers to risks to human and natural systems from SLR. Despite the difficulty in attributing observed impacts to SLR per se (Section 4.3.3.1), O’Neill et al. (2017) estimate that risks related to SLR are already detectable globally and will increase rapidly, so that high risk may occur before a 1m rise level is reached. O’Neill et al. (2017) also suggest that limits to coastal protection and EbA by 2100 could occur in a 1 m SLR rise scenario. P revious assessments however left gaps, including quantifying the benefits from adaptation in terms of risk reduction.

<div id="section-4-3-4-1-methodological-advances"></div>

<span id="methodological-advances"></span>
==== 4.3.4.1 Methodological Advances ====

<div id="section-4-3-4-1-methodological-advances-block-1"></div>

Rather than revisiting the AR5 and O’Neill et al. (2017) assessments from the particular perspective of risk related to SLR and for the global scale, this section provides a complementary perspective by assessing risks for specific geographies ( resource-rich coastal cities , urban atoll islands, large tropical agricultural deltas and selected Arctic communities), based on the methodological advances below.

'''''Scale of analysis and geographical scope''''' – To date, the RFCs and associated burning embers have been developed at a global scale (Oppenheimer et al., 2014; Gattuso et al., 2015; O’Neill et al., 2017) and do not address the spatial variability of risk highlighted in this report (Sections 4.3.2.7, 4.3.4, 5.3.7, Cross-Chapter Box 9, Box 4.1). In addition, assessments usually identify risks either for global human dimensions (e.g., to people, livelihood, breakdown of infrastructures, biodiversity, global economy, etc.; IPCC, 2014; Oppenheimer et al., 2014; O’Neill et al., 2017) or for ecosystems and ecosystem services (Gattuso et al., 2015; Hoegh-Guldberg et al., 2018) (Section 5.3.7). This section moves the focus from the global to more local scales by considering four generic categories of low-lying coastal areas (Figure 4.3, Panel B): selected Arctic communities remote from regions of rapid GIA, large tropical agricultural deltas, urban atoll islands, and resource-rich coastal cities . Each of these categories is informed by several real-world case studies.

'''''Risks considered''''' – In line with the AR5 (IPCC, 2014) , current and future risks result from the interaction of SLR-related hazards with the vulnerability of exposed ecosystems and societies. According to the specific scope of the chapter, this assessment focusses on the additional risks due to SLR and does not account for changes in extreme event climatology. Hazards considered are coastal flooding (Section 4.3.4.2), erosion (Section 4.3.4.3) and salinisation (Section 4.3.4.4). The proxies used to describe exposure and vulnerability are the density of assets at the coast (Section 4.3.2.2) and the level of degradation of natural buffering by marine and terrestrial ecosystems (Sections 4.3.2.3, 4.3.3.5.4, and 5.3.2 to 5.3.4). The assessment especially addresses risks to human assets at the coast, including populations, infrastructures and livelihoods. Specific metrics were developed (see SM4.3 for details), and their contribution to present-day observed impacts and to end-century risk have been assessed based on the authors’ expert judgment and a methodological grid presented in SM4.3 (SM4.3.1 to SM4.3.6). The author’s expert judgment draws on Sections 4.3.3.2 to 4.3.3.5 as well as additional literature for local scale perspectives (SM4.3.9).

'''''Sea level rise scenarios''''' – Based on the updates for ranges and mean values developed in this chapter (Section 4.2, Table 4.3), this assessment considers the end-century GMSL (2100) relative to 1986–2005 levels for two scenarios, SROCC RCP2.6 and SROCC RCP8.5. Both mean values and the SROCC RCP8.5 upper end of the  ''likely '' range are used to assess risk transitions (Figure 4.3, Panel A). For the sake of readability, the following values were used: 43 cm (mean SROCC RCP2.6), 84 cm (mean SROCC RCP8.5) and 110 cm (SROCC RCP8.5 upper end of  ''likely '' range). While GMSL serves as a representation of different possible climate change scenarios (see Panel A in Figure 4.3, Section 4.1.2), the assessment of additional risks due to SLR on specific geographies is developed against end-century relative SLR (RSL) in order to allow a geographically accurate approach (Panel B, Figure 4.3). Accordingly, risk was assessed to illustrative geographies based on RSLs for each of the two SROCC RCP scenarios and each of the real-world case studies to (SM4.3.6 and Table SM4.3.2; see dotted lines in Panel B of Figure 4.3). RSL observations include some or all of the following VLMs: both uplift (e.g., due to tectonics) and subsidence due to natural (e.g., tectonics, sediment compaction) and human (e.g., oil/gas/water extraction, mining activities) factors, as well as to GIA. However, in SROCC, numerical RSL projections only include GIA and the regional gravitational, rotational, and deformational responses (GRD, see Section 4.2.1.5) to ice mass loss. The main reason is the difficulty   of  project ing   the influence o n   some factors such as human interventions to the end of the century.

'''''Adaptation scenarios''''' – Risk will also depend on the effectiveness of coastal societies’ responses to both extreme events and slow onset changes. To capture the response dimension, four metrics have been considered that refer to the implementation of adequately calibrated hard, engineered coastal defences (Section 4.4.2.2), the restoration of the degraded ecosystems or the creation of new natural buffers areas (Section 4.4.2.2 and 4.4.2.3), planned and local-scale relocation (Section 4.4.2.6), and measures to limit human-induced subsidence (Sections 4.4.2.2, 4.4.2.5). On these bases, two contrasting adaptation scenarios were considered. The first one is called ‘No-to-moderate response’ (see (A) bars in Panel B, Figure 4.3) and represents a business-as-usual scenario where no major additional adaptation efforts compared to today are implemented. That is, neither substantial intensification of current actions nor new types of actions, e.g., only moderate raising of existing protections in high-density areas or sporadic episodes of relocation or beach nourishment where largescale efforts are not already underway. The second one, called ‘Maximum potential response’ (bars (B) in Figure 4.3), refers to an ambitious combination of both incremental and transformational adaptation (i.e., significantly upscaled effort); for example, relocation of entire districts or raised protections in some cities, or creation/restoration at a significant scale of beach-dune systems including indigenous vegetation.

<div id="section-4-3-4-2-key-findings-on-future-risks-and-adaptation-benefits"></div>

<span id="key-findings-on-future-risks-and-adaptation-benefits"></span>
==== 4.3.4.2 Key Findings on Future Risks and Adaptation Benefits ====

<div id="section-4-3-4-2-key-findings-on-future-risks-and-adaptation-benefits-block-1"></div>

<span id="future-risks"></span>
===== 4.3.4.2.1 Future risks =====

The findings suggest that risks from SLR are already detectable for all of the geographies considered (Panel B in Figure 4.3), and that risk is expected to increase over this century in virtually all low-lying coastal areas whatever their context-specificities or nature (island/continental, developed/developing county) (Cross-Chapter Box 9). In the absence of high adaptation (bars (A)), risk is expected to significantly increase in urban atoll islands and the selected Arctic coastal communities even in a SROCC RCP2.6 scenario, and all geographies are expected to experience almost high to very high risks at the upper ''likely'' range of SROCC RCP8.5. These results allow refining AR5 conclusions by showing, first, that high risk can indeed occur before the 1m rise benchmark (Oppenheimer et al., 2014 <sup>[[#fn:r1399|1399]]</sup> ; O’Neill et al., 2017) and, second, that risk as a function of SLR is highly variable from one geography to another. Some rationale is provided below for our assessment of illustrative geographies, summarising the more detailed description provided in SM4.3 (SM4.3.6 to SM4.3.8). Note however that the text below is not intended to be fully comprehensive and does not necessarily include all elements for which there is a substantive body of literature, nor does it necessarily include all elements which are of particular interest to decision makers.

'''''Resource-rich coastal cities''''' (SM4.3.8.1, Panel B in Figure 4.3) – Resource-rich coastal cities considered in this analysis are Shanghai, New York (see Box 4.1 for further details and references on Shanghai and NYC), and Rotterdam (Brinke et al., 2010 <sup>[[#fn:r1400|1400]]</sup> ; Hinkel et al., 2018 <sup>[[#fn:r1401|1401]]</sup> ). High, and in many cases, growing population density and total population, and high exposure of people and infrastructure to GMSL rise and ESL events characterise coastal megacities (Hanson et al., 2011 <sup>[[#fn:r1402|1402]]</sup> ). These are high concentrations of income and wealth in geographic terms but within relatively small area exhibit large distributional differences of both with important implications for emergency response and adaptation. Concentration translates into high exposure of monetary value to coastal hazards and the cities noted here have both historical and recent experience with damaging ESL events, such as Typhoon Winnie which struck Shanghai in 1997 (Xian et al., 2018 <sup>[[#fn:r1403|1403]]</sup> ), Hurricane Sandy in New York in 2012 (Rosenzweig and Solecki, 2014 <sup>[[#fn:r1404|1404]]</sup> ), and the North Sea storm of 1953 which impacted the Rotterdam area (Gerritsen, 2005 <sup>[[#fn:r1405|1405]]</sup> ; Jonkman et al., 2008 <sup>[[#fn:r1406|1406]]</sup> ). However, high density, limited space and high cost of land leads to development of below-ground space for transportation (e.g., subways, road tunnels; MTA, 2017) and storage, and even habitation, creating vulnerabilities not seen in low-density areas. Natural ecosystems within the megacity boundaries and nearby have been exploited for centuries and in some cases decimated or even extirpated (Hartig et al., 2002 <sup>[[#fn:r1407|1407]]</sup> ). Accordingly, they provide limited benefits in terms of coastal protection for the densest part of these cities but can be critically important for protection of lower-density areas, for example, wetlands and sandy beaches in the Jamaica Bay/Rockaway sector of New York that protect nearby residential communities (Hartig et al., 2002 <sup>[[#fn:r1408|1408]]</sup> ). Space limitations also constrain the potential benefits of EbA measures. Instead, resource-rich coastal cities depend largely on hard defences like sea walls and surge barriers for coastal protection (Section 4.4.2.2). Such defences are costly but generally cost effective due to the aforementioned concentration of population and value. However, barriers to planning and implementing adaptation include governance challenges (Section 4.4.2) such as limited control over finances and the intermittent nature of ESLs which inhibit focused attention over the long time scales needed to plan and implement hard defences (Section 4.4.2.2). As a result, coastal adaptation for resource-rich cities is uneven and the three presented here were selected with a view toward exhibiting a range of current and potential future effectiveness.

'''''Urban atoll islands''''' (SM4.3.8.2, Panel B in Figure 4.3) – The capital islands (or groups of islands) of three atoll nations in the Pacific and Indian Oceans are considered here: Fongafale (Funafuti Atoll, Tuvalu), the South Tarawa Urban District (Tarawa Atoll, Kiribati) and Male’ (North Kaafu Atoll, Maldives). Urban atoll islands have low elevation (&lt;4 m above mean sea level; in South Tarawa, e.g., lagoon sides where settlement concentrates are &lt;1.80 m in elevation) (Duvat, 2013) and are mainly composed of reef-derived unconsolidated material. Their future is of nation-wide importance as they concentrate populations, economic activities and critical infrastructure (airports, main harbours). They illustrate the prominence of anthropogenic-driven disturbances to marine and terrestrial ecosystems (e.g., mangrove clearing in South Tarawa or human-induced coral reef degradation through land reclamation in Male’; Duvat et al., 2013 <sup>[[#fn:r1409|1409]]</sup> ; Naylor, 2015 <sup>[[#fn:r1411|1411]]</sup> ) and therefore to services such as coastal protection delivered by the coral reef (i.e., wave energy attenuation that reduces flooding and erosion, and sediment provision that contributes to island persistence over time) (McLean and Kench, 2015 <sup>[[#fn:r1412|1412]]</sup> ; Quataert et al., 2015 <sup>[[#fn:r1413|1413]]</sup> ; Elliff and Silva, 2017 <sup>[[#fn:r1414|1414]]</sup> ; Storlazzi et al., 2018 <sup>[[#fn:r1415|1415]]</sup> ).

The controlling factors of urban atoll islands’ future habitability are the density of assets exposed to marine flooding and coastal erosion (SM4.3.8.2), future trends in these hazards, and ecosystem response to both ocean-climate related pressures and human activities. Urban atoll islands already experience coastal flooding, for example, in Male’ (Wadey et al., 2017 <sup>[[#fn:r1416|1416]]</sup> ) and Funafuti (Yamano et al., 2007 <sup>[[#fn:r1417|1417]]</sup> ; McCubbin et al., 2015 <sup>[[#fn:r1418|1418]]</sup> ). Coastal erosion is also a major concern along non-armoured shoreline in South Tarawa (Duvat et al., 2013 <sup>[[#fn:r1419|1419]]</sup> ) and Fongafale (Onaka et al., 2017 <sup>[[#fn:r1420|1420]]</sup> ), but not in Male’ where surrounding fortifications have extended along almost the entire shoreline from several decades (Naylor, 2015). Salinisation already affects groundwater lenses, but its contribution to risk varies from one case to another, from low in Male’ (relying on desalinised seawater) to important for human consumption and agriculture in South Tarawa (Bailey et al., 2014 <sup>[[#fn:r1422|1422]]</sup> ; Post et al., 2018 <sup>[[#fn:r1423|1423]]</sup> ).

Together, high population densities (from ~3,200 people per km 2 in South Tarawa to ~65,700 people per km 2 in Male’) (Government of the Maldives, 2014 <sup>[[#fn:r1424|1424]]</sup> ; McIver et al., 2015 <sup>[[#fn:r1425|1425]]</sup> ) and the concentration of critical infrastructure and settlements in naturally low-lying flood-prone areas already substantially contribute to coastal risk (Duvat et al., 2013 <sup>[[#fn:r1426|1426]]</sup> ; Field et al., 2017 <sup>[[#fn:r1427|1427]]</sup> ). Even stabilised densities in the future would translate into a substantial increase of risk under a 43cm GMSL rise. Risk will also be exacerbated by the negative effects of ocean warming and acidification, especially on coral reef and mangrove capacity to cope with SLR (Pendleton et al., 2016 <sup>[[#fn:r1428|1428]]</sup> ; Van Hooidonk et al., 2016 <sup>[[#fn:r1429|1429]]</sup> ; Perry and Morgan, 2017 <sup>[[#fn:r1430|1430]]</sup> ; Perry et al., 2018 <sup>[[#fn:r1431|1431]]</sup> ) (Sections 4.3.3.5, 5.3). In addition, even small values of SLR will significantly increase risk to atoll islands’ aquifers (Bailey et al., 2016; Storlazzi et al., 2018). Finally, land scarcity in atoll environments will exacerbate the importance of SLR induced damages (on housing, agriculture and infrastructure especially) and cascading impacts (on livelihoods, for example, as a result of groundwater and soil salinisation).

'''''Large tropical agricultural deltas''''' (SM4.3.8.3, Panel B in Figure 4.3) – River deltas considered in this analysis are the Mekong Delta and the Ganges-Brahmaputra-Meghna Delta. Both deltas are large, low-lying and dominated by agricultural production. The risk assessment to SLR considered the entire delta area (not only the coastal fringe; see SM4.3.6 for explanation). High population densities (1280 people per km 2  and 433 people per km 2  in the Ganges-Brahmaputra-Meghna and Mekong deltas, respectively) (Ericson et al., 2006 <sup>[[#fn:r1433|1433]]</sup> ; Government of the Maldives, 2014 <sup>[[#fn:r1434|1434]]</sup> ) and the removal of natural vegetation buffers contribute to high exposure rates to coastal flooding, erosion, and salinisation. Agricultural production contributes to GDP strongly (Smajgl et al., 2015 <sup>[[#fn:r1435|1435]]</sup> ; Hossain et al., 2018 <sup>[[#fn:r1436|1436]]</sup> ), making agricultural fields important assets. In both deltas, mangroves are partially degraded (Ghosh et al., 2018 <sup>[[#fn:r1437|1437]]</sup> ; Veettil et al., 2018 <sup>[[#fn:r1438|1438]]</sup> ) as well as other wetlands at the coast and further inland (Quan et al., 2018a <sup>[[#fn:r1439|1439]]</sup> ; Rahman et al., 2018 <sup>[[#fn:r1440|1440]]</sup> ). Currently, riverine flooding dominates in both deltas (Auerbach et al., 2015 <sup>[[#fn:r1441|1441]]</sup> ; Rahman and Rahman, 2015 <sup>[[#fn:r1442|1442]]</sup> ; Ngan et al., 2018 <sup>[[#fn:r1443|1443]]</sup> ). However, high tides and cyclones can generate large coastal flooding events, especially in the Ganges-Brahmaputra-Meghna Delta (Auerbach et al., 2015 <sup>[[#fn:r1444|1444]]</sup> ; Rahman and Rahman, 2015 <sup>[[#fn:r1445|1445]]</sup> ). Human-induced subsidence increases the likelihood of flooding in both deltas (Brown et al., 2018b <sup>[[#fn:r1446|1446]]</sup> ). Coastal and river bank erosion is already a problem in both delta (Anthony et al., 2015 <sup>[[#fn:r1447|1447]]</sup> ; Brown and Nicholls, 2015 <sup>[[#fn:r1448|1448]]</sup> ; Li et al., 2017 <sup>[[#fn:r1449|1449]]</sup> ) as well as salinity intrusion, which is impacting coastal aquifers, soils and surface waters (Anthony et al., 2015 <sup>[[#fn:r1450|1450]]</sup> ; Brown and Nicholls, 2015 <sup>[[#fn:r1451|1451]]</sup> ; Li et al., 2017 <sup>[[#fn:r1452|1452]]</sup> ). Salinisation of water and soil resources remains a coastal phenomenon (Smajgl et al., 2015 <sup>[[#fn:r1453|1453]]</sup> ), but salinity intrusion can reach far inland in some extreme years and significantly contribute to risk at the delta scale (Section 4.3.3.4.2). Both deltas are partly protected with hard engineered defences such as dikes and sluice gates to prevent riverine flooding, and polders and dikes in some coastal stretches to prevent salinity intrusion and storm surges (Smajgl et al., 2015 <sup>[[#fn:r1454|1454]]</sup> ; Rogers and Overeem, 2017 <sup>[[#fn:r1455|1455]]</sup> ; Warner et al., 2018a <sup>[[#fn:r1456|1456]]</sup> ). Today, in both deltas, the measures implemented to restore natural buffers are still limited to mangroves ecosystems (Quan et al., 2018a <sup>[[#fn:r1457|1457]]</sup> ; Rahman et al., 2018 <sup>[[#fn:r1458|1458]]</sup> ), and the measures aiming at reducing subsidence are underdeveloped (Schmidt, 2015 <sup>[[#fn:r1459|1459]]</sup> ; Schmitt et al., 2017 <sup>[[#fn:r1460|1460]]</sup> ). Assuming stable population densities in the future, coastal flooding will contribute increasingly to risk at the delta level (Brown and Nicholls, 2015 <sup>[[#fn:r1461|1461]]</sup> ; Brown et al., 2018a <sup>[[#fn:r1462|1462]]</sup> ; Dang et al., 2018 <sup>[[#fn:r1463|1463]]</sup> ). Coastal erosion will increase (Anthony et al., 2015 <sup>[[#fn:r1464|1464]]</sup> ; Liu et al., 2017a <sup>[[#fn:r1465|1465]]</sup> ; Uddin et al., 2019 <sup>[[#fn:r1466|1466]]</sup> ) and salinisation of coastal waters and soils will be more significant (Tran Anh et al., 2018 <sup>[[#fn:r1467|1467]]</sup> ; Vu et al., 2018 <sup>[[#fn:r1468|1468]]</sup> ; Rakib et al., 2019 <sup>[[#fn:r1469|1469]]</sup> ) and will strongly impact agriculture and water supply for the entire delta (Jiang et al., 2018 <sup>[[#fn:r1470|1470]]</sup> ; Timsina et al., 2018 <sup>[[#fn:r1471|1471]]</sup> ; Nhung et al., 2019 <sup>[[#fn:r1472|1472]]</sup> ). Without increased adaptation, coastal ecosystems will be largely destroyed at 110 cm of SLR (Schmitt et al., 2017 <sup>[[#fn:r1473|1473]]</sup> ; Mehvar et al., 2019 <sup>[[#fn:r1474|1474]]</sup> ; Mukul et al., 2019 <sup>[[#fn:r1475|1475]]</sup> ). Given the size of these deltas, it is only under high emission scenarios, that flooding, erosion and salinisation lead to high risk at the entire delta scale.

'''''Arctic communities''''' (SM4.3.8.4, Panel B in Figure 4.3) – Five small indigenous settlements located on the Arctic Coastal Plain are considered in this analysis: Bykovsky (Lena Delta, Russian Federation), Shishmaref and Kivalina (Alaska, USA), and Shingle Point and Tuktoyaktuk (Mackenzie Delta, Canada). They lie on exposed coasts composed of unlithified ice-rich sediments in permafrost, in areas with seasonal sea ice and slow to moderate SLR. These communities have populations ranging from 380 to 900 (fewer and seasonal at Shingle Point) that are heavily dependent on marine subsistence resources (Forbes, 2011 <sup>[[#fn:r1476|1476]]</sup> ; Ford et al., 2016a <sup>[[#fn:r1477|1477]]</sup> ). Shishmaref and Kivalina are located on low-lying barrier islands highly susceptible to rising sea level (Marino, 2012 <sup>[[#fn:r1478|1478]]</sup> ; Bronen and Chapin, 2013 <sup>[[#fn:r1479|1479]]</sup> ; Fang et al., 2018 <sup>[[#fn:r1480|1480]]</sup> ; Rolph et al., 2018 <sup>[[#fn:r1481|1481]]</sup> ). Shingle Point is situated on an active gravel spit; Tuktoyaktuk is built on low ground with high concentrations of massive ice; and Bykovsky is mostly situated on an ice-rich eroding terrace about 20 m above sea level. All the selected communities are remote from regions of rapid positive GIA; many other areas in the Arctic experience rapid GIA uplift (James et al., 2015 <sup>[[#fn:r1482|1482]]</sup> ; Forbes et al., 2018 <sup>[[#fn:r1483|1483]]</sup> ) and have very low sensitivity to SLR, which may in fact help to reduce shoaling.

Especially in the Arctic, anthropogenic drivers in recent decades resulted in the induced settlement of indigenous peoples in marginalised climate-sensitive communities (Ford et al., 2016b) and the construction of infrastructure in nearshore areas, with the assumption of stable coastlines. This resulted in increased exposure to coastal hazards. Coastal erosion is already a major problem in all of the case studies, where space for building is usually limited. Accelerating permafrost thaw is promoting rapid erosion of ice-rich sediments, e.g., at Bykovsky (Myers, 2005 <sup>[[#fn:r1485|1485]]</sup> ; Lantuit et al., 2011 <sup>[[#fn:r1486|1486]]</sup> ; Vanderlinden et al., 2018 <sup>[[#fn:r1487|1487]]</sup> ) and Tuktoyaktuk (Lamoureux et al., 2015 <sup>[[#fn:r1488|1488]]</sup> ; Ford et al., 2016a <sup>[[#fn:r1489|1489]]</sup> ). Related to this, Kivalina, Shishmaref, Shingle Point, Tuktoyaktuk, and parts of the Lena delta (less so for Bykovsky) are already facing high risk of flooding. Shishmaref, for example, experienced 10 flooding events between 1973 and 2015 that resulted in emergency declarations (Bronen and Chapin, 2013 <sup>[[#fn:r1490|1490]]</sup> ; Lamoureux et al., 2015 <sup>[[#fn:r1491|1491]]</sup> ; Irrgang et al., 2019 <sup>[[#fn:r1492|1492]]</sup> ). There is however no evidence of salinisation in the selected communities, but brackish water flooding of the outer Mackenzie Delta caused by a 1999 storm surge (a rare event due to upwelling ahead of the storm) led to widespread die-off of vegetation with negative ecosystem impacts (Pisaric et al., 2011 <sup>[[#fn:r1493|1493]]</sup> ; Kokelj et al., 2012 <sup>[[#fn:r1494|1494]]</sup> ).

Permafrost thaw is already accelerating due to increasing ground temperatures that weaken the mechanical stability of frozen ground (Section 3.4.2.2). Arctic SLR and sea surface warming have the potential to substantially contribute to this thawing (Forbes, 2011 <sup>[[#fn:r1495|1495]]</sup> ; Barnhart et al., 2014b <sup>[[#fn:r1496|1496]]</sup> ; Lamoureux et al., 2015 <sup>[[#fn:r1497|1497]]</sup> ; Fritz et al., 2017 <sup>[[#fn:r1498|1498]]</sup> ). An additional factor unique to the polar regions is the decrease in seasonal sea ice extent in the Arctic (Sections 3.2.1 and 3.2.2), which together with a lengthening open water season, provides less protection from storm impacts, particularly later in the year when storms are prevalent (Forbes, 2011 <sup>[[#fn:r1499|1499]]</sup> ; Lantuit et al., 2011 <sup>[[#fn:r1500|1500]]</sup> ; Barnhart et al., 2014a <sup>[[#fn:r1501|1501]]</sup> ; Melvin et al., 2017 <sup>[[#fn:r1502|1502]]</sup> ; Fang et al., 2018 <sup>[[#fn:r1503|1503]]</sup> ; Forbes, 2019 <sup>[[#fn:r1504|1504]]</sup> ) and therefore reduces the physical protection of the land (Section 6.3.1.3).

<div id="section-4-3-4-2-key-findings-on-future-risks-and-adaptation-benefits-block-2"></div>

B in Figure 4.3), and that risk is expected to increase over this century in virtually all low-lying coastal areas whatever their context-specificities or nature (island/continental, developed/developing county) (Cross-Chapter Box 9). In the absence of high adaptation (bars (A)), risk is expected to significantly increase in urban atoll islands and the selected Arctic coastal communities even in a SROCC RCP2.6 scenario, and all geographies are expected to experience almost high to very high risks at the upper ''likely'' range of SROCC RCP8.5. These results allow refining AR5 conclusions by showing, first, that high risk can indeed occur before the 1m rise benchmark (Oppenheimer et al., 2014; O’Neill et al., 2017) and, second, that risk as a function of SLR is highly variable from one geography to another. Some rationale is provided below for our assessment of illustrative geographies, summarising the more detailed description provided in SM4.3 (SM4.3.6 to SM4.3.8). Note however that the text below is not intended to be fully comprehensive and does not necessarily include all elements for which there is a substantive body of literature, nor does it necessarily include all elements which are of particular interest to decision makers.

'''''Resource-rich coastal cities''''' (SM4.3.8.1, Panel B in Figure 4.3) – Resource-rich coastal cities considered in this analysis are Shanghai, New York (see Box 4.1 for further details and references on Shanghai and NYC), and Rotterdam (Brinke et al., 2010 <sup>[[#fn:r1400|1400]]</sup> ; Hinkel et al., 2018 <sup>[[#fn:r1401|1401]]</sup> ). High, and in many cases, growing population density and total population, and high exposure of people and infrastructure to GMSL rise and ESL events characterise coastal megacities (Hanson et al., 2011 <sup>[[#fn:r1402|1402]]</sup> ). These are high concentrations of income and wealth in geographic terms but within relatively small area exhibit large distributional differences of both with important implications for emergency response and adaptation. Concentration translates into high exposure of monetary value to coastal hazards and the cities noted here have both historical and recent experience with damaging ESL events, such as Typhoon Winnie which struck Shanghai in 1997 (Xian et al., 2018 <sup>[[#fn:r1403|1403]]</sup> ), Hurricane Sandy in New York in 2012 (Rosenzweig and Solecki, 2014 <sup>[[#fn:r1404|1404]]</sup> ), and the North Sea storm of 1953 which impacted the Rotterdam area (Gerritsen, 2005 <sup>[[#fn:r1405|1405]]</sup> ; Jonkman et al., 2008 <sup>[[#fn:r1406|1406]]</sup> ). However, high density, limited space and high cost of land leads to development of below-ground space for transportation (e.g., subways, road tunnels; MTA, 2017) and storage, and even habitation, creating vulnerabilities not seen in low-density areas. Natural ecosystems within the megacity boundaries and nearby have been exploited for centuries and in some cases decimated or even extirpated (Hartig et al., 2002 <sup>[[#fn:r1407|1407]]</sup> ). Accordingly, they provide limited benefits in terms of coastal protection for the densest part of these cities but can be critically important for protection of lower-density areas, for example, wetlands and sandy beaches in the Jamaica Bay/Rockaway sector of New York that protect nearby residential communities (Hartig et al., 2002 <sup>[[#fn:r1408|1408]]</sup> ). Space limitations also constrain the potential benefits of EbA measures. Instead, resource-rich coastal cities depend largely on hard defences like sea walls and surge barriers for coastal protection (Section 4.4.2.2). Such defences are costly but generally cost effective due to the aforementioned concentration of population and value. However, barriers to planning and implementing adaptation include governance challenges (Section 4.4.2) such as limited control over finances and the intermittent nature of ESLs which inhibit focused attention over the long time scales needed to plan and implement hard defences (Section 4.4.2.2). As a result, coastal adaptation for resource-rich cities is uneven and the three presented here were selected with a view toward exhibiting a range of current and potential future effectiveness.

'''''Urban atoll islands''''' (SM4.3.8.2, Panel B in Figure 4.3) – The capital islands (or groups of islands) of three atoll nations in the Pacific and Indian Oceans are considered here: Fongafale (Funafuti Atoll, Tuvalu), the South Tarawa Urban District (Tarawa Atoll, Kiribati) and Male’ (North Kaafu Atoll, Maldives). Urban atoll islands have low elevation (&lt;4 m above mean sea level; in South Tarawa, e.g., lagoon sides where settlement concentrates are &lt;1.80 m in elevation) (Duvat, 2013 <sup>[[#fn:r1409|1409]]</sup> ) and are mainly composed of reef-derived unconsolidated material. Their future is of nation-wide importance as they concentrate populations, economic activities and critical infrastructure (airports, main harbours).They illustrate the prominence of anthropogenic-driven disturbances to marine and terrestrial ecosystems (e.g., mangrove clearing in South Tarawa or human-induced coral reef degradation through land reclamation in Male’; Duvat et al., 2013 <sup>[[#fn:r1410|1410]]</sup> ; Naylor, 2015 <sup>[[#fn:r1411|1411]]</sup> ) and therefore to services such as coastal protection delivered by the coral reef (i.e., wave energy attenuation that reduces flooding and erosion, and sediment provision that contributes to island persistence over time) (McLean and Kench, 2015 <sup>[[#fn:r1412|1412]]</sup> ; Quataert et al., 2015 <sup>[[#fn:r1413|1413]]</sup> ; Elliff and Silva, 2017 <sup>[[#fn:r1414|1414]]</sup> ; Storlazzi et al., 2018 <sup>[[#fn:r1415|1415]]</sup> ).

The controlling factors of urban atoll islands’ future habitability are the density of assets exposed to marine flooding and coastal erosion (SM4.3.8.2), future trends in these hazards, and ecosystem response to both ocean-climate related pressures and human activities. Urban atoll islands already experience coastal flooding, for example, in Male’ (Wadey et al., 2017 <sup>[[#fn:r1416|1416]]</sup> ) and Funafuti (Yamano et al., 2007 <sup>[[#fn:r1417|1417]]</sup> ; McCubbin et al., 2015 <sup>[[#fn:r1418|1418]]</sup> ). Coastal erosion is also a major concern along non-armoured shoreline in South Tarawa (Duvat et al., 2013 <sup>[[#fn:r1419|1419]]</sup> ) and Fongafale (Onaka et al., 2017 <sup>[[#fn:r1420|1420]]</sup> ), but not in Male’ where surrounding fortifications have extended along almost the entire shoreline from several decades (Naylor, 2015 <sup>[[#fn:r1421|1421]]</sup> ). Salinisation already affects groundwater lenses, but its contribution to risk varies from one case to another, from low in Male’ (relying on desalinised seawater) to important for human consumption and agriculture in South Tarawa (Bailey et al., 2014 <sup>[[#fn:r1422|1422]]</sup> ; Post et al., 2018 <sup>[[#fn:r1423|1423]]</sup> ).

Together, high population densities (from ~3,200 people per km <sup>2</sup> in South Tarawa to ~65,700 people per km <sup>2</sup> in Male’) (Government of the Maldives, 2014 <sup>[[#fn:r1424|1424]]</sup> ; McIver et al., 2015 <sup>[[#fn:r1425|1425]]</sup> ) and the concentration of critical infrastructure and settlements in naturally low-lying flood-prone areas already substantially contribute to coastal risk (Duvat et al., 2013 <sup>[[#fn:r1426|1426]]</sup> ; Field et al., 2017 <sup>[[#fn:r1427|1427]]</sup> ). Even stabilised densities in the future would translate into a substantial increase of risk under a 43cm GMSL rise. Risk will also be exacerbated by the negative effects of ocean warming and acidification, especially on coral reef and mangrove capacity to cope with SLR (Pendleton et al., 2016 <sup>[[#fn:r1428|1428]]</sup> ; Van Hooidonk et al., 2016 <sup>[[#fn:r1429|1429]]</sup> ; Perry and Morgan, 2017 <sup>[[#fn:r1430|1430]]</sup> ; Perry et al., 2018 <sup>[[#fn:r1431|1431]]</sup> ) (Sections 4.3.3.5, 5.3). In addition, even small values of SLR will significantly increase risk to atoll islands’ aquifers (Bailey et al., 2016 <sup>[[#fn:r1431|1431]]</sup> ; Storlazzi et al., 2018 <sup>[[#fn:r1432|1432]]</sup> ). Finally, land scarcity in atoll environments will exacerbate the importance of SLR induced damages (on housing, agriculture and infrastructure especially) and cascading impacts (on livelihoods, for example, as a result of groundwater and soil salinisation).

'''''Large tropical agricultural deltas''''' (SM4.3.8.3, Panel B in Figure 4.3) – River deltas considered in this analysis are the Mekong Delta and the Ganges-Brahmaputra-Meghna Delta. Both deltas are large, low-lying and dominated by agricultural production. The risk assessment to SLR considered the entire delta area (not only the coastal fringe; see SM4.3.6 for explanation). High population densities (1280 people per km <sup>2</sup>  and 433 people per km <sup>2</sup>  in the Ganges-Brahmaputra-Meghna and Mekong deltas, respectively) (Ericson et al., 2006 <sup>[[#fn:r1433|1433]]</sup> ; Government of the Maldives, 2014 <sup>[[#fn:r1434|1434]]</sup> ) and the removal of natural vegetation buffers contribute to high exposure rates to coastal flooding, erosion, and salinisation. Agricultural production contributes to GDP strongly (Smajgl et al., 2015 <sup>[[#fn:r1435|1435]]</sup> ; Hossain et al., 2018 <sup>[[#fn:r1436|1436]]</sup> ), making agricultural fields important assets. In both deltas, mangroves are partially degraded (Ghosh et al., 2018 <sup>[[#fn:r1437|1437]]</sup> ; Veettil et al., 2018 <sup>[[#fn:r1438|1438]]</sup> ) as well as other wetlands at the coast and further inland (Quan et al., 2018a <sup>[[#fn:r1439|1439]]</sup> ; Rahman et al., 2018 <sup>[[#fn:r1440|1440]]</sup> ). Currently, riverine flooding dominates in both deltas (Auerbach et al., 2015 <sup>[[#fn:r1411|1411]]</sup> ; Rahman and Rahman, 2015 <sup>[[#fn:r1442|1442]]</sup> ; Ngan et al., 2018 <sup>[[#fn:r1443|1443]]</sup> ). However, high tides and cyclones can generate large coastal flooding events, especially in the Ganges-Brahmaputra-Meghna Delta (Auerbach et al., 2015 <sup>[[#fn:r1444|1444]]</sup> ; Rahman and Rahman, 2015 <sup>[[#fn:r1445|1445]]</sup> ). Human-induced subsidence increases the likelihood of flooding in both deltas (Brown et al., 2018b <sup>[[#fn:r1446|1446]]</sup> ). Coastal and river bank erosion is already a problem in both delta (Anthony et al., 2015 <sup>[[#fn:r1447|1447]]</sup> ; Brown and Nicholls, 2015 <sup>[[#fn:r1448|1448]]</sup> ; Li et al., 2017 <sup>[[#fn:r1449|1449]]</sup> ) as well as salinity intrusion, which is impacting coastal aquifers, soils and surface waters (Anthony et al., 2015 <sup>[[#fn:r1450|1450]]</sup> ; Brown and Nicholls, 2015 <sup>[[#fn:r1451|1451]]</sup> ; Li et al., 2017 <sup>[[#fn:r1452|1452]]</sup> ). Salinisation of water and soil resources remains a coastal phenomenon (Smajgl et al., 2015 <sup>[[#fn:r1453|1453]]</sup> ), but salinity intrusion can reach far inland in some extreme years and significantly contribute to risk at the delta scale (Section 4.3.3.4.2). Both deltas are partly protected with hard engineered defences such as dikes and sluice gates to prevent riverine flooding, and polders and dikes in some coastal stretches to prevent salinity intrusion and storm surges (Smajgl et al., 2015 <sup>[[#fn:r1454|1454]]</sup> ; Rogers and Overeem, 2017 <sup>[[#fn:r1455|1455]]</sup> ; Warner et al., 2018a <sup>[[#fn:r1456|1456]]</sup> ). Today, in both deltas, the measures implemented to restore natural buffers are still limited to mangroves ecosystems (Quan et al., 2018a <sup>[[#fn:r1457|1457]]</sup> ; Rahman et al., 2018 <sup>[[#fn:r1458|1458]]</sup> ), and the measures aiming at reducing subsidence are underdeveloped (Schmidt, 2015 <sup>[[#fn:r1459|1459]]</sup> ; Schmitt et al., 2017 <sup>[[#fn:r1460|1460]]</sup> ). Assuming stable population densities in the future, coastal flooding will contribute increasingly to risk at the delta level (Brown and Nicholls, 2015 <sup>[[#fn:r1461|1461]]</sup> ; Brown et al., 2018a <sup>[[#fn:r1462|1462]]</sup> ; Dang et al., 2018 <sup>[[#fn:r1463|1463]]</sup> ). Coastal erosion will increase (Anthony et al., 2015 <sup>[[#fn:r1464|1464]]</sup> ; Liu et al., 2017a <sup>[[#fn:r1465|1465]]</sup> ; Uddin et al., 2019 <sup>[[#fn:r1466|1466]]</sup> ) and salinisation of coastal waters and soils will be more significant (Tran Anh et al., 2018 <sup>[[#fn:r1467|1467]]</sup> ; Vu et al., 2018 <sup>[[#fn:r1468|1468]]</sup> ; Rakib et al., 2019 <sup>[[#fn:r1469|1469]]</sup> ) and will strongly impact agriculture and water supply for the entire delta (Jiang et al., 2018 <sup>[[#fn:r1470|1470]]</sup> ; Timsina et al., 2018 <sup>[[#fn:r1471|1471]]</sup> ; Nhung et al., 2019 <sup>[[#fn:r1472|1472]]</sup> ). Without increased adaptation, coastal ecosystems will be largely destroyed at 110 cm of SLR (Schmitt et al., 2017 <sup>[[#fn:r1473|1473]]</sup> ; Mehvar et al., 2019 <sup>[[#fn:r1474|1474]]</sup> ; Mukul et al., 2019 <sup>[[#fn:r1475|1475]]</sup> ).Given the size of these deltas, it is only under high emission scenarios, that flooding, erosion and salinisation lead to high risk at the entire delta scale.

'''''Arctic communities''''' (SM4.3.8.4, Panel B in Figure 4.3) – Five small indigenous settlements located on the Arctic Coastal Plain are considered in this analysis: Bykovsky (Lena Delta, Russian Federation), Shishmaref and Kivalina (Alaska, USA), and Shingle Point and Tuktoyaktuk (Mackenzie Delta, Canada). They lie on exposed coasts composed of unlithified ice-rich sediments in permafrost, in areas with seasonal sea ice and slow to moderate SLR. These communities have populations ranging from 380 to 900 (fewer and seasonal at Shingle Point) that are heavily dependent on marine subsistence resources (Forbes, 2011 <sup>[[#fn:r1476|1476]]</sup> ; Ford et al., 2016a <sup>[[#fn:r1477|1477]]</sup> ). Shishmaref and Kivalina are located on low-lying barrier islands highly susceptible to rising sea level (Marino, 2012 <sup>[[#fn:r1478|1478]]</sup> ; Bronen and Chapin, 2013 <sup>[[#fn:r1479|1479]]</sup> ; Fang et al., 2018 <sup>[[#fn:r1480|1480]]</sup> ; Rolph et al., 2018 <sup>[[#fn:r1481|1481]]</sup> ). Shingle Point is situated on an active gravel spit; Tuktoyaktuk is built on low ground with high concentrations of massive ice; and Bykovsky is mostly situated on an ice-rich eroding terrace about 20 m above sea level. All the selected communities are remote from regions of rapid positive GIA; many other areas in the Arctic experience rapid GIA uplift (James et al., 2015 <sup>[[#fn:r1482|1482]]</sup> ; Forbes et al., 2018 <sup>[[#fn:r1483|1483]]</sup> ) and have very low sensitivity to SLR, which may in fact help to reduce shoaling.

Especially in the Arctic, anthropogenic drivers in recent decades resulted in the induced settlement of indigenous peoples in marginalised climate-sensitive communities (Ford et al., 2016b) and the construction of infrastructure in nearshore areas, with the assumption of stable coastlines. This resulted in increased exposure to coastal hazards. Coastal erosion is already a major problem in all of the case studies, where space for building is usually limited. Accelerating permafrost thaw is promoting rapid erosion of ice-rich sediments, e.g., at Bykovsky (Myers, 2005 <sup>[[#fn:r1485|1485]]</sup> ; Lantuit et al., 2011 <sup>[[#fn:r1486|1486]]</sup> ; Vanderlinden et al., 2018 <sup>[[#fn:r1487|1487]]</sup> ) and Tuktoyaktuk (Lamoureux et al., 2015 <sup>[[#fn:r1488|1488]]</sup> ; Ford et al., 2016a <sup>[[#fn:r1489|1489]]</sup> ). Related to this, Kivalina, Shishmaref, Shingle Point, Tuktoyaktuk, and parts of the Lena delta (less so for Bykovsky) are already facing high risk of flooding. Shishmaref, for example, experienced 10 flooding events between 1973 and 2015 that resulted in emergency declarations (Bronen and Chapin, 2013 <sup>[[#fn:r1490|1490]]</sup> ; Lamoureux et al., 2015 <sup>[[#fn:r1491|1491]]</sup> ; Irrgang et al., 2019 <sup>[[#fn:r1492|1492]]</sup> ). There is however no evidence of salinisation in the selected communities, but brackish water flooding of the outer Mackenzie Delta caused by a 1999 storm surge (a rare event due to upwelling ahead of the storm) led to widespread die-off of vegetation with negative ecosystem impacts (Pisaric et al., 2011 <sup>[[#fn:r1493|1493]]</sup> ; Kokelj et al., 2012 <sup>[[#fn:r1494|1494]]</sup> ).

Permafrost thaw is already accelerating due to increasing ground temperatures that weaken the mechanical stability of frozen ground (Section 3.4.2.2). Arctic SLR and sea surface warming have the potential to substantially contribute to this thawing (Forbes, 2011 <sup>[[#fn:r1495|1495]]</sup> ; Barnhart et al., 2014b <sup>[[#fn:r1496|1496]]</sup> ; Lamoureux et al., 2015 <sup>[[#fn:r1497|1497]]</sup> ; Fritz et al., 2017 <sup>[[#fn:r1498|1498]]</sup> ). An additional factor unique to the polar regions is the decrease in seasonal sea ice extent in the Arctic (Sections 3.2.1 and 3.2.2), which together with a lengthening open water season, provides less protection from storm impacts, particularly later in the year when storms are prevalent (Forbes, 2011 <sup>[[#fn:r1499|1499]]</sup> ; Lantuit et al., 2011 <sup>[[#fn:r1500|1500]]</sup> ; Barnhart et al., 2014a <sup>[[#fn:r1501|1501]]</sup> ; Melvin et al., 2017 <sup>[[#fn:r1502|1502]]</sup> ; Fang et al., 2018 <sup>[[#fn:r1503|1503]]</sup> ; Forbes, 2019 <sup>[[#fn:r1504|1504]]</sup> ) and therefore reduces the physical protection of the land (Section 6.3.1.3).

<div id="section-4-3-4-2-key-findings-on-future-risks-and-adaptation-benefits-block-3"></div>

<span id="adaptation-benefits"></span>
===== 4.3.4.2.2 Adaptation benefits =====

<div id="section-4-3-4-2-key-findings-on-future-risks-and-adaptation-benefits-block-4"></div>

The assessment also shows that benefits in terms of risk reduction over this century are to be expected from ambitious adaptation efforts (bars (B), Sections 4.4.2, 4.4.3 and 4.4.3). In the case of resource-rich coastal cities especially, adequately engineered coastal defences can play a decisive role in reducing risk (Section 4.4.2.2, Box 4.1), for example from high to moderate at the SROCC RCP8.5 upper ''likely'' range. In other contexts, such as atoll islands for example, while engineered protection structures will reduce risk of flooding, they will not necessarily prevent seawater infiltration due to the permeable nature of the island substratum. So even adequate coastal protection would not eliminate risk (SM4.3.8.3). In urban atoll islands, large tropical agricultural deltas and the selected Arctic communities, ambitious adaptation efforts mixing adequate coastal defences, the restoration and creation of buffering ecosystems (e.g., coral reefs), and a moderate amount of relocation are expected to reduce risk. For resource-rich coastal cities , adequately engineered hard protection can virtually eliminate risk of flooding up to 84 cm except for residual risk of structural failure (Sections 4.4.2 to 4.4.5). Benefits are relatively important in a 84 cm SLR scenario, as they reduce risk from high-to-very-high to moderate-to-high (atolls, Arctic) and from moderate-to-high to moderate (deltas). These benefits become more modest when approaching the upper ''likely'' range of SROCC RCP8.5, and risk tends to return to high-to-very-high (atolls, Arctic) levels once the 110 cm rise in sea level is reached. Noteworthy in urban atoll islands, intensified proactive coastal relocation (e.g., relocation of buildings and infrastructures that are very close to the shoreline) is expected to play a substantial role in risk reduction under all SLR scenarios. Proactive relocation can indeed compensate for the increasing extent of coastal flooding and associated damages (SM4.3.8.3). When taken to the extreme, relocation could lead to the elimination of risk in situ, for example in the case of the relocation of the full population of urban atoll islands either elsewhere in the country (e.g., on another island) or abroad (i.e., international migration). This is an extreme situation where it is hard to distinguish whether the measure is an impact of SLR (and ocean change more broadly), for example, displacement, or an adaptation solution. In addition, relocation of people displaces pressure to destination areas, with a potential increase of risk for the latter. In other words, the broader ‘coastal retreat’ category (Section 4.4.2.6) raises the issue of the ‘limits to adaptation’, which is not represented in Figure 4.3.

These conclusions must be nuanced, first, by the fact that our assessment does not consider either financial or social aspects that can act as limiting factors to the development of adaptation options (Sections 4.4.3 and 4.4.5), for instance, hard engineering coastal defences (Hurlimann et al., 2014 <sup>[[#fn:r1505|1505]]</sup> ; Jones et al., 2014 <sup>[[#fn:r1506|1506]]</sup> ; Elrick-Barr et al., 2017 <sup>[[#fn:r1507|1507]]</sup> ; Hinkel et al., 2018 <sup>[[#fn:r1508|1508]]</sup> ) . However, from a general perspective, these findings suggest that although ambitious adaptation will not necessarily eradicate end-century risk from SLR across all low-lying coastal areas around the world, it will help to buy time in many locations and therefore contribute to developing a robust foundation for adaptation beyond 2100. Second, the future of other climate-related drivers of risk (such as ESL, waves and cyclones; Sections 4.2.3.4.1 to 4.2.3.4.3, 6.3.1.1 to 6.3.1.3) is not fully and systematically included in each risk assessment above, so that much larger risks than assessed here are to be expected.

<span id="responding-to-sea-level-rise"></span>