A BACKGROUND PAPER >> SEA LEVEL RISE This work is a product of the staff of The World Bank and the Global Facility for Disaster Reduction and Recovery (GFDRR) with external contributions. The sole responsibility of this publications lies with the authors. The findings, analysis and conclusions expressed in this document do not necessarily reflect the views of any individual partner organization of The World Bank (including the European Union), its Board of Directors, or the governments they represent, and therefore they are not responsible for any use that may be made of the information contained therein. Although the World Bank and GFDRR make reasonable efforts to ensure all the information presented in this document is correct, its accuracy and integrity cannot be guaranteed. Use of any data or information from this document is at the user’s own risk and under no circumstances shall the World Bank, GFDRR or any of its partners be liable for any loss, damage, liability or expense incurred or suffered which is claimed to result from reliance on the data contained in this document. The boundaries, colors, denomination, and other information shown in any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Assessing the impact of sea level rise and resilience potential in the Caribbean Revisiting Resilience in the Caribbean: a 360-degree approach c Confidential Assessing the impact of sea level rise and resilience potential in the Caribbean Revisiting Resilience in the Caribbean: a 360-degree approach Author(s) Alessio Giardino Tim Leijnse Luisa Torres Duenas Panos Athanasiou Marjolijn Haasnoot 2 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Assessing the impact of sea level rise and resilience potential in the Caribbean Revisiting Resilience in the Caribbean: a 360-degree approach Client The World Bank Contact Julie Rozenberg Keywords Caribbean, SIDS, Sea level rise, Coastal erosion, Flood risk, Resilience, Adaptation, Delft-FIAT, SFINCS Document control Version 0.1 Date 22-09-2020 Project nr. 11205356-002 Document ID 20-03-0028 Pages 67 Status final Authors Alessio Giardino Tim Leijnse Luisa Torres Dueñas Panos Athanasiou Yingrong Wen Marjolijn Haasnoot Doc. version Author Reviewer Approver Publish 0.1 Alessio Giardino Ap van Dongeren Toon van Segeren Acknowledgements The authors would like to thank Lorenzo Mentaschi and Michalis Vousdoukas from JRC for providing the global wave data and giving their feedback on the methodological approach. 3 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Contents 1 Summary 6 2 Introduction 7 2.1 Scope of the study 8 3 Study area 9 4 Material and methods 11 4.1 Data 11 4.1.1 Topography 11 4.1.2 Water levels 12 4.1.3 Offshore waves 12 4.1.4 Sea level rise scenarios 13 4.1.5 Socio-economic data and projections 14 4.1.6 Sandy beaches and nearshore slopes 17 4.2 Modelling framework 18 4.2.1 Introduction 18 4.2.2 Flood hazard module 18 4.2.3 Flood impact module 20 4.2.4 Beach erosion modelling 22 4.3 Impact indicators 24 5 Results 25 5.1 Introduction 25 5.2 Flood hazard assessment 25 5.2.1 General patterns flooding 26 5.2.2 Sensitivity local DEM 30 5.3 Flood risk assessment 32 5.4 Sandy beach erosion assessment 38 5.5 Strategies for adaptation: the St. Lucia case study 41 5.5.1 Introduction 41 5.5.2 Sea level rise impact at St. Lucia 41 5.5.3 Strategies for adaptation 44 5.5.4 Identification of areas suitable for future development 47 6 Discussion 50 6.1 Perspective results to other studies 50 6.2 Assumptions 51 6.2.1 Sea level rise and climate change scenarios 51 6.2.2 Flood hazard 51 6.2.3 Flood risk 52 6.2.4 Sandy beach erosion and associated damages 52 6.2.5 Coastal adaptation 53 6.3 Recommendations 53 7 Conclusions 55 4 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 8 References 57 Appendix 63 5 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 1 Summary The Caribbean region suffers major economic losses from natural hazards such as flooding due to storms, cyclones, extreme waves, winds and precipitation, coastal erosion, volcanic eruptions and landslides. Consequently, as typical at most small coastal states, when a disaster strikes, a large part of the population, infrastructure and businesses, generally concentrated in the coastal areas, are directly or indirectly affected. Climate change and sea level rise (SLR), in combination with socio-economic growth, are likely to exacerbate this situation, which is already critical for many of these countries. In particular, the effect of SLR will lead to more frequent and intense flooding events and chronical coastal erosion, with a direct effect on the local and regional economies. In this study, a regional estimation of the effects of SLR in terms of coastal flooding and erosion of sandy beaches was carried out for 18 countries in the Caribbean with the aim of deriving proxies to evaluate the resilient potential of each country and their potential to adaptation. The (change in) risk resulting from SLR was estimated until 2100 under different SLR scenarios and socio-economic pathways. Other types of hazards such as from rainfall, river flooding and extreme winds are not included in the study. The results show that the effect of SLR alone could lead to an increase in Expected Average Annual Damages (EAAD) at the regional level by almost 30% (40%)1 by 2050 and 80% (145%) by 2100 under RCP 4.5 (RCP 8.5), only as result of coastal flooding, following regional SLR estimates of Vousdoukas et al. (2018). The expected average annual people affected could increase by almost 25% (35%) by 2050 and 70% (120%) by 2100 under RCP 4.5 (RCP 8.5). A total of 192 km2 (542 km2) of sandy beaches is expected to be lost over all the considered countries by 2050 (2100). A reduction in carbon emissions (e.g. following a RCP 4.5 scenario) can play an important role in reducing projected land loss. The reduction can be almost 20% by 2050 and almost 40% by 2100, relative to the high-emission RCP 8.5 scenario. The countries more vulnerable to the effects of SLR on coastal flooding are Guyana, Suriname and Saint Martin. The countries where we expect the largest losses of land at sandy beaches are the Bahamas, Belize, Turks and Caicos Islands, Dominican Republic and Haiti. Potential strategies for adaptation were explored for a pilot case at St. Lucia, with focus on when or where a “protect”, “accommodate” and “planned retreat” strategy can be impleme nted at different locations around the country. —————————————— 1 Values before brackets refer to RCP 4.5 scenario, value in brackets refer to RCP 8.5 scenario. 6 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 2 Introduction Presently, the Caribbean suffers major economic losses from natural hazards including flooding due to storms, cyclones, extreme waves, winds and precipitation, coastal erosion, volcanic eruptions and landslides. The Caribbean’s vulnerability is typical of small island states, but somehow accentuated by the combination of all these different hazards. Average disaster damage as a ratio to GDP was estimated to be 4.5 times greater for small states than for larger ones, but six times higher for countries in the Caribbean (Trevor et al., 2017). Since 1950, small states worldwide, defined as states with population lower than 1.5 million (IMF, 2016), have suffered 511 disasters. Among these disasters, 324 were in the Caribbean, killing 250,000 people and affecting more than 24 million. The lack of resilience to disasters of most of these countries is a barrier to sustainable growth. For instance, it was estimated that without tropical cyclones, Jamaica’s economy could have grown by as much as 4% per year; instead, over the past 40 years, it has grown only 0.8% annually (Hsiang and Jina, 2014). Sometimes, island economies were dealt devastating blows: when Hurricane Maria struck Dominica in 2017, it caused damages and losses equivalent to over 220% of the country’s GDP (UNDRR, 2019). Climate change and Sea Level Rise (SLR) may exacerbate the frequency and intensity of storms causing flooding, erosion, extreme winds and landslides with a direct or indirect effect on people living at the coast, critical infrastructures, tourism and coastal ecosystems. In case of no adaptation against those hazards, losses are likely to increase due to current and future population growth combined with changes in climate and SLR. The global impact that Sea Level Rise (SLR) will have worldwide, over the course of the 21st century, has been subject of several studies (e.g. Hallegatte et al., 2013; Hinkel et al., 2014; Tiggeloven et al., 2019; Vousdoukas et al., 2016; Vousdoukas et al., 2018). The Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC) reported that global sea levels will rise by a likely range (representing at least 66% probability) of 0.26 – 0.98 m (compared to 1986 – 2005) by the end of the 21st century (2100), with unprecedented impacts on the society (IPCC, 2013). Available information suggests that SLR trends in the Caribbean have been broadly similar to global trends over the last 60 years (Nicholls, R. and Cazenave, A., 2010; Palanisamy et al., 2012). Similarly, regional future projections show comparable future trends to the global projections (Church et al., 2013). Moreover, there is a low but not negligible probability that SLR may be higher than the above mentioned predictions due to accelerated ice sheet melting (De Conto and Pollard, 2016). To adapt to these possible rapid and highly uncertain changes it will be necessary to take decisions on how to adapt sooner and/or to implement solutions (Haasnoot et al., 2019). Simpson et al. (2010) estimated that over 110,000 people may be displaced in the CARICOM (CARibbean COMmunity) nations in case of 1 m SLR and no adaptation. The countries which will be mostly affected economically will be the ones that are the most dependent on the touristic sector. The numerical modelling of coastal hazards and resulting impacts can be very useful to assess the possible consequences that SLR will have, evaluate the resilience potential of a country and, finally, plan the most suitable adaption strategies and options (see e.g. Giardino et al., 2018). 7 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 2.1 Scope of the study The scope of the study is to assess the effects that SLR will have on (increase in) coastal flood hazard and risk as well as on erosion of sandy beaches across 18 countries that are part of the Caribbean region (Section 3), and to derive proxies to estimate their resilient potential. The assessment was carried out based on of data analysis (Section 4.1), complemented by the use of a numerical modelling framework to provide a consistent assessment throughout the countries (Section 4.2). Using different impact indicators (Section 4.3), the results from the flood risk and coastal erosion assessment were intercompared between countries (Sections 5.3 and 5.4). The effect of both SLR and extreme storm events were taken into account in the flood risk assessment. For the case of St. Lucia, the assessment was used as a basis to discuss possible options for adaptation (Section 5.5). As described in Section 6.2, the study strictly focusses on assessing impacts due to SLR only, therefore excluding other forms of hazards such as from rainfall, river flooding and extreme winds. Only the effect of SLR of sandy beach retreat was assessed, while the possible effect of SLR on other coastal types (e.g. muddy, gravel or rocky coast) was not considered. For consistency reason in the assessment between countries, the study makes use of global datasets of topographic data, population and land use maps and vulnerability curves. When available, results from global data (i.e. DEM) were compared with results obtained by using local data, indicating the importance of reliable data for an accurate assessment. 8 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 3 Study area The study focuses on 18 countries in the Caribbean region (Figure 1), most of which are officially recognized as Small Islands Development States (SIDS) by the United Nations. In total, 3,350 islands were identified as part of the 18 countries. However, only the 207 islands where buildings could be identified from global datasets and/or islands with a surface area of at least 50 km2 were finally included in the assessment. Among these islands, the Dutch and French part of the island of Sint Maarten were considered separately, named respectively Sint Maarten (NL) and Saint Martin (FR). Figure 1 Countries considered in this study. For an overview of the assessed number of islands per country as well as the Gross Domestic Product (GDP) and total population see Table 1. 9 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Table 1 Overview of selected countries with number of islands, total area, nominal GDP and total population. The years of data and source are also provided. nominal GDP Total area nominal GDP estimate Source Population Source Country Islands [km2] [Million USD] year GDP in 2018 Population Antigua and World World 4 481 1,611 2018 96,286 Barbuda Bank Bank World World Bahamas 147 16,195 12,425 2018 385,640 Bank Bank World World Barbados 1 461 5,145 2018 286,641 Bank Bank World World Belize 7 24,236 1,871 2018 383,071 Bank Bank World World Dominica 1 816 551 2018 71,625 Bank Bank Dominican World World 3 54,017 85,555 2018 10,627,165 Republic Bank Bank World World Grenada 4 379 1,186 2018 111,454 Bank Bank World World Guyana 6 212,841 3,879 2018 779,004 Bank Bank World World Haiti 6 30,462 9,659 2018 11,123,176 Bank Bank World World Jamaica 1 12,449 15,714 2018 2,934,855 Bank Bank St. Kitts and World World 2 295 1,011 2018 52,441 Nevis Bank Bank World World St. Lucia 1 656 1,922 2018 181,899 Bank Bank St. Vincent and World World 8 422 811 2018 110,210 the Grenadines Bank Bank World Saint Martin (FR) 3 62 562 2005 CIA 37,264 Bank United World Sint Maarten (NL) 1 42 1,059 2017 40,654 Nations Bank World World Suriname 1 148,220 3,591 2018 575,991 Bank Bank Trinidad and World World 3 5,369 23,808 2018 1,389,858 Tobago Bank Bank Turks and Caicos World World 8 1,155 1,022 2018 37,665 Islands Bank Bank 10 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 4 Material and methods 4.1 Data The assessment required the following input data: topography, water levels, wave forcing, sea level rise scenarios, socio-economic projections, location of sandy beaches and average slopes. The simple approach used for wave transformation did not require explicitly the use of bathymetric data since the offshore wave conditions were directly used as basis to estimate the wave setup around the islands (Section 4.1.3 and 4.2.2). 4.1.1 Topography Global topographic (elevation) data were used at most of the islands, as input for the flood hazard assessment. Locally-measured DEM (Digital Elevation Model) data, provided by the Worldbank, were only available at the mainland of Belize and the islands of Dominica, Grenada and St. Lucia, with a spatial resolution of 30 m, 0.5 m, 1 m and 30 m respectively. At countries at which local data was available, this was used instead of the global DEM. Since several global products were available, at first an investigation was performed to investigate which global data source would perform better compared to locally available data. The island of Grenada was used as test case and with specific focus on the coastal area (i.e. elevation below 10 meters). In particular, a comparison was made between the local DEM and global products as CoastalDEM (Kulp and Strauss, 2018), MERIT (Yamazaki et al., 2017), SRTM (Farr et al., 2007), ALOS (Rosenqvist et al., 2007), ASTER (Fujisada et al., 2005) and TanDEM-X (Krieger et al., 2007). The considered global DEMs are a mix of data based on optical and radar derived satellite products, while CoastalDEM and MERIT are error corrected (i.e. for vegetation) products of SRTM. The different global products all overestimated the elevation of the relatively steep island of Grenada with several meters. After deriving error statistics, the dataset of MERIT showed to give the lowest combination of bias and RMSE (Root Mean Squared Error, Table 2). Therefore, MERIT was used as the global topographic dataset within this study. While the bias for this steep island is high (5.6 m), it is generally expected that for flatter islands (i.e. also more prone to the impact of SLR) the global DEM will be more accurate in the coastal zone. However, no local DEM was available for this type of islands so the error could not be quantified. On the other hand, for mainland countries, a local DEM was available for Belize, and a validation was performed for the global MERIT and CoastalDEM products. The bias was -1.8 m and -2.9 m for MERIT and CoastalDEM respectively, while the RMSE was 4.4 m and 5 m. These results indicate that for the flatter coastal areas of Belize the global DEM products perform better than at the steep islands. Table 2 Error statistics for elevations up to 10 meters for Grenada as derived from global DEM sources compared to local DEM data. DEM source Bias [m] RMSE [m] CoastalDEM 6.10 8.87 MERIT 5.62 7.63 SRTM30 9.00 10.02 ALOS 8.63 9.41 ASTER 9.18 10.34 TanDEMX 5.44 11.27 11 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Based on the topography, different island/country typologies can be distinguished: mild steep islands, steep islands and mainland countries (Section 5.2.1). 4.1.2 Water levels The values of the extreme water levels were based on Muis et al. (2020) which provides extreme sea levels along the world’s coastlines for specific segments as used in Hinkel and Klein (2009). The data contains extreme water levels for different return periods (i.e. the once per year, once per 2, 5, 10, 50 and 100year levels). The data includes combined tidal and surge components. The coverage of the dataset is high enough that all countries has one or more data point. The data in Muis et al. (2020) is referenced to the Mean Sea Level (MSL), while the applied topography is referenced with respect to the EGM96 geoid. To convert the extreme water level data to the same reference level, the Mean Dynamic Topography (MDT) of the oceans was used following Muis et al. (2017). The MDT is the difference between the mean sea surface and the geoid. To this end, the global MDT model MDT-CNES-CLS18 (AVISO, 2018) was used, which is based on observations of satellite sea surface heights from TOPEX/Poseidon over the 1993-2012 period. It was assumed that the extreme water levels and tide estimated for the current situation are also representative for the end of the century climate. This assumption can be justified by the fact that the largest changes in extreme water levels and flooding levels are expected to be the result of SLR, while changes in storm surges, tides and the interaction between those two are of relative lower importance (see e.g. Muis et al., 2020). 4.1.3 Offshore waves Wave data was retrieved from a global dataset but downscaled to local conditions for the considered countries. The values of the wave heights were based on a 36-year long time-series of offshore wave heights as determined by JRC (Mentaschi et al., 2017). To translate these global offshore data, which do not resolve all the geomorphological details of the coastlines, to local conditions, wave sheltering effects were introduced. This was done by first defining transects approximately every 5 kilometers along the entire coastline (see e.g. Figure 2a). Then, for all transects, a reduction factor on the incoming wave direction was used, accounting for the angle between the incoming wave direction and the coastal orientation. With this correction factor, local wave conditions are different at each transect. For example, at the lee side of an island wave conditions are lower than at the side facing the incoming waves. The sheltering effect was applied for every time step of the 36-year long time-series and for all transects of all countries. This simple wave downscaling method does not solve the full wave transformation and the local effects that local bathymetry may have on this transformation. Based on the corrected wave data, an Extreme Value Analysis was performed using a peak-over- threshold method providing downscaled offshore significant wave heights at all countries for different return periods (i.e. 1, 2, 5, 10, 50, 100 years return period). An example is shown for Dominica in Figure 2b where the wave heights on the east coast are higher than on the west coast, due to the wave exposure to the open ocean. It is assumed that the extreme wave heights do not vary in time as a result of climate change. 12 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 (a) (b) Figure 2 Example transects for Dominica (a) and significant wave heights for a return period of 50 years (b). 4.1.4 Sea level rise scenarios To incorporate regional SLR, RCP (Representative Concentration Pathways) 4.5 and RCP 8.5 were used based on Vousdoukas et al. (2018). RCP 4.5 represent a scenario according to which emissions will decline after 2040, while RCP 8.5 is a high emissions scenario, according to which emissions will continue to rise throughout the 21st century. This dataset includes a high-end RCP 8.5 scenario, based on IPCC AR5, but uses Antarctic and Greenland ice sheet contributions from Bamber and Aspinall (2013) resulting in higher median global SLR (i.e., 0.84 m versus 0.74 m of AR5). The projections from Vousdoukas et al. (2018) were probabilistic and regional (i.e. include the regional footprint of SLR). Herein, only the median values were used for multiple time horizons (2030/2050/2070/2100). Note the large confidence interval around these median values. For an example of the projections for Dominica see Figure 3. For all countries the closest data points were taken, so there could be multiple points around an island which were all included. The flooding results including SLR were compared with the baseline situation of the current climate sea level, wherefore 2010 is taken as baseline. Figure 3 RCP4.5 and RCP8.5 sea level rise projections (left) for Dominica (right) according to Vousdoukas et al. (2018). 13 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 4.1.5 Socio-economic data and projections Socio-economic data and projections were required to derive current and future risk estimates. For the gross domestic products (GDP) of countries, the latest estimated nominal data per country was used. The nominal GDP data was retrieved from the databases of the World Bank, the World Fact Book, and the United Nations Statistics division (see World Bank Group, 2020; Central Intelligence Agency, 2020; United Nations, 2020). The values are presented in Table 1. For the total population of countries, the latest estimates of the World Bank from 2018 were used, see Table 1 (World Bank Group, 2020b). In order to account for future socio-economic growth (including projections for population and GDP), the Shared Socioeconomic Pathways (SSP’s) were used (O’Neill et al., 2017). The SSP’s are part of a framework that the climate change community has adopted to facilitate the integrated analysis of future climate impacts, vulnerabilities, adaptation, and mitigation (Riahi et al., 2017). The narratives (or scenarios) described by the framework (SSP1, SSP2, SSP3, SSP4 and SSP5) establish possible future development pathways based on factors like population, economic development, land and energy use. For the purpose of this study, only two socio-economic scenarios were chosen, SSP3 and SSP5. Here SSP3 stands for a scenario with high population growth and slow economic growth in developing countries (Regional rivalry), while SSP5 shows a rapid economic growth induced by intensive fossil fuel exploitation and low population growth. These two scenarios represent the extreme (i.e. lowest and highest) projections in terms of population growth for the countries analyzed in this study. The SSP’s database of Gidden et al.(2019) was used to retrieve projections for GDP and population growth for twelve countries2 for the different time horizons considered in the analysis (2010, 2030, 2050, 2070, 2100). The rate of increase/decrease from the SSP database (through 2010 till 2100 at intervals of 5 years) was used to derive corrected projections based on the latest GDP and population estimates (values of Table 1 taken as baseline) of each country. Table 3 shows the corrected projected GDP values per country for the two SSP’s selected, at different time horizons, while Table 4 shows the same for the population growth. —————————————— 2 Out of the 18 countries analysed in this study, the SSP database does not include projections for the following countries: Antigua and Barbuda, Dominica, Saint Kitts and Nevis, Saint Martin, Sint Martin and Turks and Caicos. For those countries, the growth rate in GDP of Grenada was used (Similar country in terms of population and GDP) and adjusted with the value of the nominal GDP of each country (see Table 3) 14 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 15 of 67 GDP 2010 2010 2030 2030 2050 2050 2070 2070 2100 2100 Country [million SSP3 SSP5 SSP3 SSP5 SSP3 SSP5 SSP3 SSP5 SSP3 SSP5 USD] Antigua and Barbuda 1,611 1,582 1,630 1,922 1,816 2,699 2,640 3,356 3,355 4,865 4,876 Bahamas 12,425 10,14 9,189 14,555 18,105 21,204 31,151 28,927 44,468 44,382 56,930 3 Barbados 5,145 6,062 5,992 4,501 5,025 5,113 7,676 5,478 10,721 6,841 14,275 Belize 1,871 1,465 1,369 2,663 2,970 4,050 4,580 5,628 6,237 9,043 11,046 20-03-0028, 22 September 2020 Dominica 551 541 558 657 621 923 903 1,148 1,148 1,664 1,668 Dominican Republic 85,555 61,11 60,211 107,871 122,079 161,206 205,039 217,284 281,894 318,207 436,05 Agency and UN is United Nations. 0 5 Grenada 1,186 1,165 1,201 1,415 1,337 1,988 1,944 2,471 2,471 3,582 3,590 Guyana 3,879 3,245 3,237 4,820 5,265 7,187 8,885 9,485 10,774 13,860 14,074 Haiti 9,659 7,454 7,014 14,701 19,299 23,631 53,335 30,997 113,136 44,920 201,28 4 Jamaica 15,714 12,76 12,746 26,917 27,045 45,476 37,878 60,000 41,567 93,586 51,975 1 Saint Kitts and Nevis 1,011 993 1,023 1,206 1,140 1,694 1,657 2,106 2,106 3,053 3,060 St. Lucia 1,922 1,470 1,442 2,365 2,606 3,371 4,464 4,251 6,669 5,767 9,490 St. Vincent and the Grenadines 811 493 488 1,134 1,154 1,535 1,522 1,884 2,086 2,576 2,722 Saint Martin (FR) 562 561 553 682 616 958 896 1,190 1,139 1,726 1,655 Assessing the impact of sea level rise and resilience potential in the Caribbean Sint Maarten (NL) 1,059 1,047 1,073 1,272 1,195 1,786 1,738 2,221 2,208 3,220 3,209 Suriname 3,591 2,142 2,120 5,511 5,796 8,246 9,755 10,736 14,058 15,102 19,589 Trinidad and Tobago 23,808 16,80 21,439 23,186 16,685 30,266 18,674 36,054 21,260 47,048 25,519 7 Turks and Caicos 1,022 1,004 1,035 1,220 1,153 1,713 1,676 2,130 2,130 3,088 3,095 2010/2030/2050/2070/2100 using SSP scenarios 3 and 5. WB is WorldBank, CIA is Central Intelligence Table 3 Nominal GDP values in million USD per country including projections of GDP values for the years 16 of 67 Population 2010 2010 2030 2030 2050 2050 2070 2070 2100 2100 Country 2018 [WB] SSP3 SSP5 SSP3 SSP5 SSP3 SSP5 SSP3 SSP5 SSP3 SSP5 Antigua and Barbuda 96,286 92,175 97,989 106,063 87,850 120,948 72,293 133,226 57,083 153,028 39,969 Bahamas 385,640 349,682 350,077 433,342 433,283 489,197 481,836 535,434 478,766 600,117 380,949 Barbados 286,641 281,892 283,236 291,903 286,947 280,013 268,066 263,516 230,776 257,254 155,047 20-03-0028, 22 September 2020 Belize 383,071 330,682 345,416 467,964 404,951 597,047 385,686 716,104 332,321 901,516 252,484 Dominica 71,625 68,567 72,892 78,898 65,350 89,970 53,777 99,104 42,463 113,834 29,732 Dominican Republic 10,627,165 9,567,881 9,908,710 12,401,637 11,212,033 15,131,056 11,081,192 17,473,590 10,022,901 20,515,545 7,842,164 Grenada 111,454 106,695 113,425 122,771 101,689 140,001 83,681 154,213 66,076 177,134 46,265 Guyana 779,004 742,865 785,782 866,140 719,558 947,722 551,524 1,001,502 379,834 1082,396 230,582 Haiti 11,123,176 10,039,731 10,432,020 12,908,506 11,538,134 15,421,882 11,111,802 17,053,714 97,82,832 18,479,281 7,506,369 2010/2030/2050/2070/2100 using SSP scenarios 3 and 5. Jamaica 2,934,855 2,791,876 2,931,014 3,232,568 2,762,050 3,629,146 2,244,748 3,959,976 1652,851 4,530,298 1,077,539 St. Kitts and Nevis 52,441 50,202 53,368 57,766 47,846 65,873 39,373 72,560 31,090 83,345 21,769 St. Lucia 181,889 167,459 171,527 203,094 190,434 227,835 190,494 242,125 174,700 263,671 130,455 Assessing the impact of sea level rise and resilience potential in the Caribbean St. Vincent and the 110,210 106,863 112,772 118,656 101,414 128,769 83,498 136,776 64,838 152,324 43,324 Grenadines Saint Martin (FR) 37,665 35,673 37,923 41,048 33,999 46,808 27,978 51,560 22,092 59,224 15,468 Sint Maarten (NL) 40,654 92,175 97,989 106,063 87,850 120,948 72,293 133,226 57,083 153,028 39,969 Suriname 575,991 349,682 350,077 433,342 433,283 489,197 481,836 535,434 478,766 600,117 380,949 Trinidad and Tobago 1,389,858 281,892 283,236 291,903 286,947 280,013 268,066 263,516 230,776 257,254 155,047 Turks and Caicos 37,665 330,682 345,416 467,964 404,951 597,047 385,686 716,104 332,321 901,516 252,484 Islands Table 4 Population data in thousands (in 2018) per country as well as projections of population for the years 4.1.6 Sandy beaches and nearshore slopes The current study focusses on the assessment of the effects that SLR may have on sandy beaches only. Although it is acknowledged that, at most countries, the majority of the coastline is not sandy (see for example Table 7), the erosion of sandy beaches has a direct effect on the touristic sector (see e.g. Thinh et al., 2019), which is crucial for many countries in the region. Therefore, we find it meaningful to have this explicitly estimated as part of this assessment. To identify the location and extents of sandy erodible coasts in the Caribbean, data from Deltares’ Shoreline Monitor were used (Luijendijk et al., 2018). The dataset provides globally the location of the sandy beaches as derived from a supervised classification based on cloud-free Sentinel-2 images. Additionally, nearshore slopes as calculated using global topobathymetric data were considered (Athanasiou et al., 2019). Slopes were estimated as the cross-shore inclination between one offshore location (i.e. the so-called “depth of closure") and the shoreline, computed at predefined transects around the global coastline. Combined, these data provided input for the assessment of the beach erosion modeling (Section 4.2.4). (a) (b) Figure 4 a) Spatial distribution of sandy beaches identified in Luijendijk et al. (2018). b) Nearshore slopes of all coastal profiles (Athanasiou et al., 2019). 17 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 4.2 Modelling framework 4.2.1 Introduction One modelling framework was set-up with the objective of quantifying the expected impact of SLR on marine flooding and coastal erosion in a consistent way throughout the region. Based on this framework, a number of impact indicators were computed for each country both as absolute value (e.g. expected annual damages) and relative values (e.g. expected annual damages relative to yearly GDP) (Section 4.3). The modelling framework included the following modules: a) Flood Hazard module (Section 4.2.2), including the quantification of: ▪ Coastal flooding b) Flood Impact module (Section 4.2.3), including the quantification of: ▪ Expected annual damages ▪ Population affected c) Coastal Erosion module (Section 4.2.4), including the quantification of: ▪ Erosion of sandy beaches In view of the regional scale of the assessment, the following characteristics were taking into consideration while designing the modelling framework: ▪ Computational time to be able to run the entire modelling chain for all 17 countries for several scenario combinations (SLR, time windows, return periods and socio-economics). ▪ Use of global data, which would be consistent throughout the region, to avoid differences between countries. ▪ Highest level of accuracy, which would still allow to keep a reasonable computational effort. For example, dynamic flooding simulations based on state-of-the art flooding models were used, while often studies at this scale are carried out based on static (“bathtub”) flooding approach. The effect of waves was also included, although in a simplistic way, why similar regional assessments tend to neglect this component. The coastal erosion estimates were carried out accounting for actual nearshore slopes, while similar past studies have been based on the assumption of one uniform slope of 1:100 (see e.g. Hinkel et al., 2013) 4.2.2 Flood hazard module Marine flooding, i.e. the flooding caused by wind-driven surges, tides and waves was computed for each country using the SFINCS model (Super-Fast Inundation of CoastS (Leijnse, 2018; Leijnse et al., 2020). This model is a reduced-physics solver especially designed for modelling compound flooding in coastal areas with high computational efficiency and sufficient accuracy, which makes it suitable to carry out a flood assessment at a regional scale, such as in the present study. SFINCS model domains were created based on the topographic data available at each island (Section 4.1.1). The model extents were based on land polygon shapefiles in combination with water occurrence data of Pekel et al. (2016) to distinguish between water and land. Within these model extents, the topographic data were interpolated on a 90 m resolution for countries with a global DEM, and 30 m for countries with a local DEM. All cells that were marked as water within the water occurrence data (deep water) as well as cells with elevation higher than 25 m above the geoid (as they are not affected by coastal flooding) were considered as inactive. For an example of the setup for St. Lucia see Figure 5. 18 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 The combination of extreme sea water levels (Section 4.1.2), waves (Section 4.1.3) and long-term SLR (Section 4.1.4) was used to provide the boundary conditions to force the flood hazard models at each country. The extent of the SFINCS model, including boundary conditions points, is shown in Figure 5c (i.e. SFINCS mask). Water level, wave heights and SLR values, originally available at different locations and with different resolution, were all assigned to the nearest boundary condition points and then interpolated linearly on the SFINCS model grid. Figure 5 Example of SFINCS setup for St. Lucia. Elevation map below the 25-meter contour (a), water occurrence (b) and mask-file indicating the extent of the SFINCS model and boundary condition points (c). Water level and wave boundary condition points coincide in the figure. No detailed wave model was set up for each country to transform offshore wave conditions to nearshore wave conditions. This was partly to maintain a reasonable computational effort in view of the regional scale of the assessment, and partly because detailed bathymetric data was not available around all islands, which would be required to properly simulate the wave transformation processes. Nevertheless, part of the effect of waves was included by adding a steady nearshore wave setup component to the total water level, computed as 0.2 times the offshore wave height, following the approach by Vousdoukas et al. (2018). Therefore, we neglect the dynamic wave- induced runup component which would require considerable computational effort. Extreme storm events, for different return periods, were schematized by using synthetic hydrographs with water levels varying following a simplified triangular shape. The duration of the storm (in hours) was assumed equal to: = 8 ∙ + 10 with surge representing the peak of the storm surge height in meters, following the approach as described in the LISCOAST methodology (Vousdoukas et al., 2016). SLR and tide were assumed to be constant during the duration of the storm, as the tidal component is relatively small in this region. An example of such a hydrograph is given in Figure 6. 19 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Figure 6 Example of an hydrograph for RCP 8.5 (Vousdoukas et al., 2018) for time horizon 2100, with a 100- year return period event. The contribution by the different sources is also shown in the figure: SLR, tide, surge and wave setup. SLR and tide are assumed constant during the duration of the storm, while surge and wave setup vary linearly in time following a triangular shape. 4.2.3 Flood impact module Natural hazards such as flooding can cause large economic losses as well as substantial human suffering. Those impacts are expected to increase with socio-economic development and climate change, especially if no adaptation measures are taken. Flood risk is the product of the probability that a given flooding occurs, and the generated damages to assets located within the hazardous zone. This means a combination of three components defined as follows: 1. Hazard: The frequency and intensity of natural hazards, in this case in the form of flood maps showing extent and maximum water depths for various return periods at different time horizons. 2. Exposure: The population and economic assets located in hazard-prone areas. Assets can be buildings, livestock, infrastructure, people, or any type of resources valuable for the economy and well-being of the country. 3. Vulnerability: The susceptibility of the exposed assets to a certain hazard. In other words, how much of the value of an asset is affected by a certain hazard intensity, to be combined with the maximum value of that asset. Figure 7 Definition of risk (Kron, 2005). 20 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 The risk analysis was carried out in a quantitative way by using the modelling tool Delft-FIAT (Flood Impact Assessment Tool; Slager et al., 2016), which has been developed by Deltares and applied at several flood risk studies around the world (e.g. Sao Tome and Principe, Vietnam, Sri Lanka). The tool estimates economic damages, linking hazards (Section 4.2.2) to exposure data by means of vulnerability information. Results are available as damage estimates per flood return period, which are then aggregated to compute the Expected Average Annual Damages (EAAD). Figure 8 provides an overview of the damage and risk calculation scheme in FIAT. Figure 8 Overview of damage and risk calculations in Delft-FIAT. Exposure information (“object maps”) are combined with hazard maps (“water depth maps”) via vulnerability functions (“damage function” and “maximum damages”), resulting in “damage maps” and “risk values”. Global datasets of exposure and vulnerability information were used in order to derive consistent results among countries. In particular: 1. Exposure dataset: a high definition population map was constructed based on a combination of two different sources: (a) The World Settlement Footprint (WSF; Marconcini et al., 2019), including building footprints, and (b) the global population data from the Global Human Settlement (GHS) dataset, available at a coarser resolution (250 m) (Pesaresi et al., 2016; Freire et al., 2016). The two datasets were combined to create an improved 10m resolution dataset of population which allocates people at locations where buildings are present (Palacios-Lopez et al., 2019). 2. Household size per country derived from the United Nations – Department of Economic and Social Affairs website (https://population.un.org/Household). 3. Vulnerability dataset: The vulnerability of residential homes to flooding was derived from a global database including depth-damage factor functions and maximum residential damage estimate per building, both at country level (Huizinga et al., 2017). The depth-damage factor function used was based on a weighted averaged between curves as reported in Huizinga et al. (2017), for different type of assets (residential, industrial, commercial, etc.). The weight of the different depth-damage factor functions depends on the importance of the damage class (e.g. residential, commercial, industrial, etc.) in the region. 21 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Figure 9 Overview of the depth-damage factor function used for the risk analysis presented in this study (based on Huizinga et al., 2017). The curve shows the damage factor of an asset resulting from different flood depths in meters. A factor of 0 indicates no damage, a value of 1 indicates maximum damage. The calculation was carried out according to the following steps: 1) The high-resolution population raster map was created by combining the building footprint map (World Settlement Footprint - WSF), with the global population dataset (Global Human Settlement - GHS). The population was evenly distributed where building are present, which resulted in raster data, with number of persons per cell. The cell size was 10 m x 10 m. 2) Based on the information of household size (per country), the number of houses per cell was estimated (i.e. number of houses per cell = number of persons per cell/household size). 3) By multiplying the flood maps by the vulnerability functions and the number of houses per cell, an estimation of the damage per cell was obtained. 4) To account for damages to other items which are non-residentials and for other indirect damages (e.g. business interruption) the damage values were multiplied by two additional factors − , and . − was estimated to be equal to 2 based on previous studies (see e.g. Wagenaar et al. (2019) and De Bruijn et al. (2014)). was estimated to be equal to 1.2, based on observations at several past flood events including the German Elbe floods (Gauderis and Kind, 2012), the UK floods of 2007 (Chatterton et al., 2010), hurricane Katrina (Gauderis and Kind, 2012), and several events mentioned in Vilier (2013). To estimate flood risk induced by SLR, three scenarios were considered in order to estimate future risks: 1. Without socio-economic growth but considering changes due to SLR: Exposure and vulnerability datasets were kept constant (with 2010 values) while accounting for SLR according to 2 RCP scenarios (RCP 4.5 and RCP 8.5) 2. With socio-economic growth but assuming that the hazard does not change over time (only SSP): Exposure and vulnerability datasets are changed according to the SSP scenario chosen (SSP3 or SSP5) but mean sea level remains constant (2010 baseline conditions). 3. With both considering socio-economic growth and SLR: Exposure and vulnerability datasets change according to the SSP scenario chosen (SSP3 or SSP5) and hazard information varies according to the climate change scenario (RCP 4.5 and RCP 8.5). 4.2.4 Beach erosion modelling A global assessment method of coastal erosion at sandy beaches was used to identify the erosion hazards of the Caribbean countries (Vousdoukas et al., 2020). Although for simplicity the 22 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 assessment presented in this study focuses on vulnerability of sandy beaches only, we do want to emphasize that, in most countries, the majority of the coastline is not sandy as shown for example in Table 7. The assessment included two different components of coastal erosion; long-term ambient changes (i.e. related to alongshore transport, sediment availability etc.) and SLR induced shoreline retreat. Since the former were based on historical satellite data which were extrapolated in the future, while this study focused on the effects of climate change and more specifically SLR, only the latter was used. The estimation of the SLR induced shoreline retreat R of sandy coasts was based on the Bruun rule (Bruun, 1962). This approach assumes that the beach morphology tends to adapt to an equilibrium profile and is given by: 1 = tan where is the nearshore slope, which are only used for the locations with a sandy erodible coast. SLR is the amount of SLR in meters considering the chosen time window and RCP scenario. Recent studies (e.g. Ranasinghe et al., 2012; Li, 2014) have shown that the Bruun rule tends to overestimate coastal retreat when compared with physics-based models. This overestimation may depend on site-specific factors, and to this end, a correction factor was applied to the equation, which varies randomly as described by a triangular distribution ranging between 0.1 and 1.0, with a highest probability at 0.75. This correction factor is used to reduce the shoreline retreat estimates accounting for some of the pitfalls of Bruun rule. For more details see Vousdoukas et al. (2020). Shoreline retreat was estimated in a probabilistic manner through Monte Carlo simulations, resulting in probability density functions of shoreline retreat R for each RCP scenario by 2050 and 2100, at all sandy points with respect to the baseline shoreline position at 2010. In order to translate these to land loss, the points from Vousdoukas et al. (2020) were interpolated in an alongshore grid with a spacing of 1 km, based on the Open Street Maps coastline shapefile (zoom level 8). Land loss estimates were derived by simply multiplying the shoreline retreat R with the length of each segment and aggregating per country. 23 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 4.3 Impact indicators Different definitions exist to describe the concept of resilience (see e.g. Hallegatte et al., 2017; De Bruijn, 2005). In the present report, the following indicators were used to define the impact and resilience potentials of the different countries: • EAAD with respect to GDP [%] • Expected Averaged Annual People (EAAP) affected [number of people] • EAAP with respect to the total population [%] While EAAD expresses the average annual damages in a country, EAAP describes the average number of people affected by flooding (i.e. estimated as averaged number of people inundated by at least 10 cm of water). Absolute values provide an indication of the actual impact to a country, while relative values (as a %) can be considered a proxy of the resilience potential. For the erosion module, the following indicators were used: • Average shoreline retreat of the sandy beaches due to coastal erosion [m] • Land loss as a result of coastal erosion of the sandy beaches in the country [km2] • Economic damage estimated as the nourishment cost required to bring back the coastline to its original position [USD] Erosion of sandy beaches can also be used as a proxy of the impact that SLR may have on other activities connected to the use of sandy beaches (e.g. tourism). 24 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 5 Results 5.1 Introduction In this chapter the main results of the study are presented. This includes the flood hazard and risk assessments in Sections 5.2 and 5.3. Thereafter, the results of the coastal erosion assessment is discussed (Section 5.4). Finally, these results are used to show strategies for adaptation for the island of St. Lucia (Section 5.5). 5.2 Flood hazard assessment Flood maps were created for each of the island (207 features in total) and for the different scenarios considered (Chapter 3). A total of 54 different scenarios were simulated, including: 5-time windows (2010/2030/2050/2070/2100), 2 SLR scenarios (RCP4.5 and 8.5) and 6 storm return periods (1/2/5/10/50/100 years). Note: SLR scenarios were only simulated for future scenarios and not for the current situation. The maximum observed water depth per cell throughout the simulation was used to create a unique 2D flooding map per simulation. Only a limited number of flood maps are shown in this report, as a total 10,000+ flood maps were produced. In general, as expected, flood maps show an increase of flood area and depth for storm events with a larger return periods and for SLR scenarios further away in the future. However, it depends per location whether static SLR or dynamic extreme events dominate the flooding for lower SLR scenarios. A 100-year extreme event in the current climate can possibly result in more flooding than low return period events (e.g. annual flooding) in 2100 including SLR. The use of a high emission SLR scenario (i.e. RCP8.5), amplifies the flooding more than a milder SLR scenario (i.e. RCP 4.5). In Section 5.2.1, these general patterns are shown respectively for three different categories of country: a) a mildly steep island, b) a steep island, c) a mainland country. The sensitivity to using local DEM data over using the global MERIT DEM data is shown in Section 5.2.2. It must be noted that the presented flood maps are the result of coastal flooding only, considering extreme events including surge, tides and wave setup and the contribution due to SLR. Therefore, the results are not representative for the total flood risk of the countries, since pluvial and fluvial processes are not included (see Section 6.2). 25 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 5.2.1 General patterns flooding Mildly steep island – Antigua and Barbuda: Relatively flat islands, like Antigua and Barbuda, are more susceptible to the effect of static SLR than steeper islands (Figure 10). The flooded area can increase significantly moving from the current situation (a, b) to the 2100 scenario (c, d). The coastal zone is relatively flat, therefore causing a higher MSL to permanently flood a much larger area or increase the probability of flooding of extreme events. (a) (b) (c) (d) Figure 10 Flood depth maps as a result of coastal flooding for the island of Antigua for current situation (a) and (b) and 2100 scenario under RCP 8.5 scenario, based on Vousdoukas et al. (2018) (c) and (d). Figure (a) and (c) refer to flood event with 1-year return period, figure (b) and (d) refer to an event with 100-year return period. 26 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Steep island - Dominica: For islands with a relatively steep coast, like Dominica, the flooding extent during storm events is generally small as the coastal zone is narrow and steep. In addition, also the effect of static SLR (on top of the storm effect) is limited. (a) (b) (c) (d) 27 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 (e) (f) (g) (h) Figure 11 Flood maps as a result of coastal flooding for the island of Dominica for current situation (a,b) and (c,d) and 2100 scenario under RCP 8.5 scenario, based on Vousdoukas et al. (2018) (e,f) and (g,h). Figure (a) and (c) refer to flood event of 1-year return period, figure (e) and (g) refer to an event of 100-year return period. The box in the figure shows a zoom on the capital (Roseau, figures (b), (d), (f) and (h)). 28 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Mainland country – Suriname: Mainland countries such as Belize, Guyana and Suriname are characterized by a relatively mild slope near the coastline. This make them prone to flooding and to the effect of SLR. An example is shown for Suriname in Figure 12. The extent of flooding when SLR is taken into account (c and d) is considerably larger than for the baseline case (2010) (a, b). (a) (b) (c) (d) Figure 12 Flood maps for Suriname as a result of coastal flooding for current situation (a) and (b) and 2100 scenario under RCP 8.5 scenario, based on Vousdoukas et al. (2018) (c) and (d). Figure (a) and (c) refer to flood event with 1-year return period, figure (b) and (d) refer to an event with 100-year return period. 29 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 5.2.2 Sensitivity local DEM The global MERIT DEM data has a horizontal resolution of about 90 meters, and it is therefore expected that not all local details in the topography of the islands can be resolved. This is especially important for (relatively) steep islands, where the elevation quickly rises from the shoreline to elevation above 10 meters, where coastal flooding will not occur. Additionally, the dataset is characterized by a considerable vertical bias (Section 4.1.1). The influence of using a local DEM versus a global DEM in the estimated flood maps is shown for the island of Grenada in Figure 12. The figure clearly indicates much more flooding taking place for the case in which a local DEM (supposed to be more accurate) is used. This is due to the fact that the global DEM overestimates the elevation for this island with steep slopes, leading to a prediction of (almost) no flooding at a number of coastal locations, as shown in the Figure 13. An example for a mainland country (i.e. Belize) is shown in Figure 14. For other islands, characterized by milder slopes the difference is expected to be smaller, however, here no local DEM data is available. However, vertical bias may still play a role. Figure 13 Comparison between flood maps as a result of coastal flooding for Grenada produced based on a global DEM and a local DEM. The simulation refer to the 2100 situation under RCP 8.5, following prediction by Vousdoukas et al. (2018), during a storm event with 100-year return period event. 30 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Figure 14 Comparison between flood maps as a result of coastal flooding for Belize produced based on a global DEM and a local DEM. The simulation refer to the 2100 situation under RCP 8.5, following prediction by Vousdoukas et al. (2018), during a storm event with 100-year return period event. 31 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 5.3 Flood risk assessment As described in Section 4.3, three different impact indicators were calculated for each country, being the relative EAAD, and the absolute and relative EAAP. These were calculated for 8 scenarios; with only SLR effects (RCP 4.5 and 8.5), only socio-economic growth effects (SSP3 and SSP5) and their combinations. An example of these projections is presented in Figure 15 for St. Lucia. The figure shows the development of the 3 impact indicators for different time windows for the 8 different combinations. The left column of Figure 15 shows the increase in impact due to the effect of SLR only (Figure 15 a,d,g). When socio-economic growth is taken into account but not SLR (middle column; Figure 15 b,e,h) the changes become more evident with a larger increase in EAAD under SSP5 scenario than under SSP3 scenario. The trend is reversed when considering the impact on population affected as SSP5 predicts a decrease in total population (Figure 15e). The changes become even more evident when considering a combination of SLR and SSP scenarios (right column; Figure 15 c,f,i).The figure shows how the combination of the two effects is not linear but exponential, if socio-economic growth will continue at areas currently affected by flooding and without adaptation, as assumed in the present study (Section 6.2). (a) (b) (c) (d) (e) (f) (g) (h) (i) Figure 15: Temporal change of coastal flood impact indicators for St. Lucia under 8 different scenarios. Graphs on the left (a, d, g) describe the flood risk projections with SLR only, under RCP 4.5 and RCP 8.5 (following SLR projections by Vousdoukas et al., 2018) using the present socio-economic data. Graphs in the middle (b, e, h) describe the flood risk projections with only socio-economic growth under SSP3 and SSP5 (Gidden et al., 2019), using the present flooding hazard. Graphs in the right (c, f, i) describe the flood risk projections under the four different combinations of the SLR and socio-economic growth scenarios. 32 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 When looking into the relative EAAD as a percentage of the total GDP of each country, there is quite some variability from country to country (Figure 16 and Table 5). The 3 countries with the highest values are Guyana, Suriname and Saint Martin (FR) with 2100 values of about 10%, 9% and 3.5% respectively, while the countries with the lowest values are Turks and Caicos Islands, Trinidad and Tobago and Barbados with values of about 0.1%, 0.07% and 0.02% respectively. The low impact on Turks and Caicos Islands is remarkable given the low- lying topography of the islands. This is mainly because the flooded areas in the current situation do not coincide with areas at which most of the assets are present. Additionally, the percentage increase of EAAD at 2100 under RCP 8.5 relative to the current situation is calculated for each of the 18 countries (Figure 17). The largest increase is observed for Turks and Caicos Islands, Guyana and Grenada with values of about 290%, 230% and 220% respectively. The largest relative increase at Turks and Caicos Islands is indeed the result of the low-lying topography, which makes this country the most vulnerable to the effects of SLR. With respect to the population affected by flooding, the 3 countries with the highest EAAP as a percentage of their total population are again Guyana, Suriname and Saint Martin (FR) with values of 4.7%, 2.9% and 2.8% (Figure 18 and Table 5). On the other hand, the countries with the lowest values are Saint Vincent and the Grenadines, Trinidad and Tobago and Barbados with values of about 0.08%, 0.05% and 0.01% respectively. Results per country for all different scenarios with only socioeconomic change and both climate change and socioeconomic combined, are presented respectively in the Appendix tables A.1 and A.2. Figure 16: Expected annual damages as % of total GDP per country by 2100 under RCP 8.5 (following SLR projections by Vousdoukas et al., 2018) and the present socio-economic exposure. EAAD refers to expected annual damages resulting from coastal flooding only. 33 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Figure 17 Percentage increase of expected annual damages at 2100 under the RCP 8.5 scenario (following SLR projections by Vousdoukas et al., 2018) with respect to the baseline value (2010). EAAD refers to expected annual damages resulting from coastal flooding only. 34 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Figure 18: Expected annual population affected as % of total population per country by 2100 under RCP 8.5 (following SLR projections by Vousdoukas et al., 2018 ) and the present socio-economic exposure. EAAD refers to expected annual damages resulting from coastal flooding only. Table 5: Relative coastal flood impact indicators (in %) for each country under RCP 4.5 and RCP 8.5 (following SLR projections by Vousdoukas et al., 2018) for 2010, 2050 and 2100 time horizons using baseline socio-economic exposure (2010 values). Baseline Baseline EAAD as % of GDP EAAP as % of population Country name GDP pop. RCP 4.5 RCP 8.5 RCP 4.5 RCP 8.5 (M USD) (x1,000) 2010 2050 2100 2050 2100 2010 2050 2100 2050 2100 Antigua and Barbuda 1,611 96 0.31 0.37 0.47 0.39 0.59 0.23 0.28 0.34 0.29 0.41 Bahamas 12,425 386 0.32 0.39 0.58 0.43 0.78 0.38 0.46 0.74 0.61 0.87 Barbados 5,145 287 0.02 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0.01 Belize 1,871 383 0.47 0.50 0.54 0.51 0.59 0.25 0.26 0.27 0.26 0.30 Dominica 551 72 0.76 0.85 0.96 0.87 1.13 1.46 1.51 1.58 1.52 1.79 Dominican Republic 85,555 10,627 0.15 0.19 0.24 0.20 0.31 0.05 0.06 0.07 0.06 0.09 Grenada 1,186 111 0.24 0.33 0.49 0.36 0.77 0.15 0.21 0.30 0.22 0.45 Guyana 3,879 779 3.11 4.54 6.94 4.95 10.25 1.61 2.32 3.34 2.51 4.73 Haiti 9,659 11,123 0.40 0.46 0.53 0.47 0.61 0.05 0.06 0.07 0.06 0.08 Jamaica 15,714 2,935 0.52 0.68 0.89 0.72 1.12 0.24 0.30 0.39 0.32 0.46 St. Kitts and Nevis 1,011 52 0.16 0.19 0.23 0.19 0.28 0.30 0.34 0.49 0.38 0.59 St. Lucia 1,922 182 0.30 0.33 0.39 0.34 0.46 0.13 0.14 0.17 0.15 0.20 St. Vincent and the Grenadines 811 110 0.28 0.30 0.32 0.30 0.35 0.07 0.07 0.08 0.07 0.08 Saint Martin (FR) 562 37 1.42 1.92 2.68 2.07 3.57 1.37 1.77 2.36 1.89 2.84 Sint Maarten (NL) 1,059 41 0.39 0.47 0.62 0.49 0.82 0.45 0.49 0.70 0.55 0.88 Suriname 3,591 576 3.37 4.55 6.54 4.88 8.84 1.37 1.72 2.33 1.82 2.93 Trinidad and Tobago 23,808 1,390 0.04 0.04 0.05 0.05 0.07 0.03 0.03 0.04 0.03 0.05 Turks and Caicos 1,022 38 0.03 0.06 0.08 0.06 0.10 0.04 0.08 0.09 0.09 0.12 35 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Besides comparing the results between countries, the risk indicators were also determined at aggregated level over all the 18 Caribbean countries (Figure 19 and Table 6). The current (2010) EAAD for the whole region is estimated to be 0.37% of the total GDP of the Caribbean region. When considering only the effects of SLR under RCP 4.5 (RCP 8.5) this value is expected to increase by almost 35% (45%) by 2050 and 85% (150%) by 2100 relative to the 2010 values (Figure 19 a). With respect to people affected, the current EAAP as % of the total population over all the countries is estimated to be almost 0.15%, a value which is expected to increase by almost 25% (40%) by 2050 and 75% (120%) by 2100 under RCP 4.5 (RCP 8.5) (Figure 19 g). When considering socio-economic growth only, both scenarios (SSP3 and SSP5) show a relative increase in EAAD as % of GDP by 30% (60%) by 2050 and 60% (85%) by 2100 for SSP3 (SSP5) with respect to 2010 (Figure 19 b). Conversely, while EAAP as % of the total population is expected to stay the same as in 2010 under SSP3 for both 2050 and 2100, we estimated a decrease in EAAP as % of the total population under SSP5, as a result of a declining population trend under this scenario, by 6% in 2050 and 13% in 2100 (Figure 19 h). When considering both SLR and socioeconomic growth, the increase of the total risk is projected to be significantly larger (Figure 19 c). The EAAD as % of GDP under the combination of RCP 4.5 and SSP3 (RCP 8.5 and SSP5) is expected to increase exponentially by almost 75% (120%) by 2050 and by 185% (300%) by 2100 relative to the 2010 values. Nonetheless, the EAAP as % of the total population is following quite different trends with a projected increase of 25% (25%) by 2050 and 65% (75%) by 2100 under the combination of RCP 4.5 and SSP3 (RCP 8.5 and SSP5) (Figure 19 i). This is related to different trends of population growth versus decrease respectively for scenarios SSP3 and SSP5. 36 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 (a) (b) (c) (d) (e) (f) (g) (h) (i) Figure 19: Temporal change of coastal flood impact indicators for all the Caribbean countries of the study under 8 different scenarios. Graphs on the left (a, d, g) describe the flood risk projections only with SLR under RCP 4.5 and RCP 8.5 (following SLR predictions by Vousdoukas et al., 2018) using the present socio- economic data. Graphs in the middle (b, e, h) describe the flood risk projections with only socio-economic growth under SSP3 and SSP5 (Gidden et al., 2019), using the present flooding hazard. Graphs in the right (c, f, i) describe the flood risk projections under the four different combinations of the SLR and socio-economic growth scenarios. 37 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Table 6: Coastal flood impact indicators for all the Caribbean countries of the study under 8 different scenarios and for 3 different time horizons; baseline (2010), mid-century (2050) and end-century (2100). Scenarios with only RCPs (Vousdoukas et al., 2018) consider only the future effect of SLR, using present day socio-economic exposure. Scenarios with only SSPs (Gidden et al., 2019) consider only the future effect of socio-economic growth, using present day flood hazard. Scenarios with both RCPs and SSPs take into account the combined effect of SLR and socio-economic growth. Increase Increase EAAD EAAD as % EAAP EAAP EAAP as % As % of of GDP wrt (number of as % of of total pop. Scenario Year GDP 2010 (%) people) total pop. wrt 2010 (%) RCP45 2010 0.37 - 44,490 0.15 - 2050 0.50 34 56,298 0.19 27 2100 0.69 86 74,825 0.26 73 RCP85 2010 0.37 - 44,490 0.15 - 2050 0.53 43 60,082 0.21 40 2100 0.94 152 95,561 0.33 120 SSP3 2010 0.47 - 43,641 0.16 - 2050 0.62 30 62,723 0.16 0 2100 0.75 58 78,285 0.16 0 SSP5 2010 0.48 - 44,490 0.16 - 2050 0.76 59 42,941 0.15 -6 2100 0.89 86 27,109 0.14 -13 RCP45/SSP3 2010 0.47 - 43,641 0.16 - 2050 0.82 73 78,493 0.2 25 2100 1.35 185 127,990 0.26 63 RCP45/SSP5 2010 0.48 - 44,490 0.16 - 2050 0.99 109 53,245 0.19 19 2100 1.48 211 43,103 0.23 44 RCP85/SSP3 2010 0.47 - 43,641 0.16 - 2050 0.87 84 83,600 0.21 31 2100 1.80 279 161,355 0.32 100 RCP85/SSP5 2010 0.48 - 44,490 0.16 - 2050 1.06 122 56,830 0.2 25 2100 1.91 301 53,258 0.28 75 5.4 Sandy beach erosion assessment Potential shoreline retreat estimates of sandy beaches were calculated under RCP 4.5 and 8.5 at 2050 and 2100 based on Vousdoukas et al. (2020). These values were then aggregated per country to estimate erosion indexes like the average shoreline retreat and total land loss. It should be noted that since there is no information on the actual beach width due to the scale of the study, the calculated values represent potential shoreline retreat, assuming that there is enough sand to be eroded. This assumption likely results in an overestimation of the actual situation as in reality some of the sandy beaches are backed by hard structures, and/or non-erodible rocky formations. For the results presented herein, only the median projections are used. All shoreline retreat projections are relative to the baseline year of 2010. The average shoreline retreat indicates how much the sandy beaches of a country will retreat landwards on average (Figure 20). Suriname, Guyana and Trinidad and Tobago are the three countries with the largest average shoreline retreat, which is projected to be almost 200, 185 and 150 m by 2100 under RCP 8.5. These values are connected to the relative mild nearshore slopes encountered at these coastlines. Nevertheless, while the average retreat is large, the relative part of the coastline that is sandy, as identified by the satellite imagery, is not that high for these three countries (2%, 10% and 14% respectively, Figure 20). The countries that are identified as having the largest percentage of sandy coastlines relatively to their total coastline length are Sint Maarten (NL), Barbados, Turks and Caicos Islands, Saint Martin (FR), the Bahamas and Antigua and 38 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Barbuda with 96, 83, 80, 74, 62 and 59% of sandy coastlines respectively. Hereby both the Dutch (Sint Maarten) the French (Saint Martin) parts of the island are projected to also have a relatively high average shoreline retreat of 111 and 115 m respectively by 2100 under RCP 8.5. On the other hand, under the same scenario, Dominica is an example of a country with a quite low shoreline retreat (i.e. ~3.5 m) as the nearshore slopes are quite steep, while additionally only a small part of the island’s coastline (i.e. 1.6%) was identified as sandy by the satellite imagery. On average, over all the 18 Caribbean countries of the study, a shoreline retreat of 27 m (35 m) and 61 m (98 m) is projected under RCP 4.5 (8.5) by 2050 and 2100 respectively (Table 7). Combining the average shoreline retreat of each country with the sandy coastline length provides an indication of the potential sandy area land loss for each country (Figure 21 and Table 7). The country with the highest land loss is the Bahamas, where by 2100 almost 375 km 2 of sandy land is projected to be lost due to SLR under RCP 8.5. This is the result of a combination of mild nearshore slopes, a quite long coastline due to many islands (5,550 km) and a quite high percentage of sandy coastline (~62%). Other countries that are projected to lose a large amount of sandy area are Belize, the Turks and Caicos Islands, the Dominican Republic and Haiti with 35, 27, 23 and 20 km2 projected to be lost by 2100 under RCP 8.5. Under the same scenario, a total of 192 km2 (542 km2) is expected to be lost over all the 18 Caribbean countries combined by 2050 (2100). Global climate change mitigation (i.e. RCP 4.5) can play an important role in reducing projected land loss by almost 20% by 2050 and almost 40% by 2100, relative to the high-emission RCP 8.5 scenario. Figure 20 Average shoreline retreat for each country at 2100 under RCP 8.5 (Vousdoukas et al., 2018b). This value is calculated as the average value of the shoreline retreat of all the sandy points of one country. The percentages on the top of each bar indicate the relative length of the sandy coastlines to the total length of the coastline of each country. 39 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Table 7: Averaged shoreline retreat and land loss using under RCP 4.5 and RCP 8.5 (Vousdoukas et al., 2018b) by 2050 and 2100 for each country. The country’s total coastal length, sandy coastal length and average nearshore slope is also shown. Average shoreline Land loss (km2) Sandy coastal Total coastal retreat (m) length (km) length (km) % of sandy nearshore Average RCP 4.5 RCP 8.5 RCP 4.5 RCP 8.5 length slope Country name 2050 2100 2050 2100 2050 2100 2050 2100 Antigua and 154 91 59.09 0.007 31 68 38 108 2.78 6.22 3.45 9.78 Barbuda Bahamas 5,550 3,436 61.91 0.084 30 66 39 109 103.95 227.88 132.98 374.33 Barbados 87 72 82.76 0.029 8 18 10 28 0.59 1.32 0.72 2.03 Belize 944 266 28.18 3.298 37 83 46 131 9.90 22.13 12.23 34.72 Dominica 123 2 1.63 0.406 1 2 1 3 0.00 0.00 0.00 0.01 Dominican 1,312 423 32.24 0.008 15 34 19 54 6.40 14.29 8.00 22.70 Republic Grenada 123 18 14.63 0.002 28 63 36 102 0.51 1.14 0.65 1.83 Guyana 815 78 9.57 0.357 53 118 65 185 4.11 9.18 5.07 14.40 Haiti 1,507 445 29.53 0.210 13 29 16 46 5.77 12.89 7.18 20.39 Jamaica 650 109 16.77 0.017 27 61 33 95 2.99 6.67 3.65 10.36 Saint Kitts 105 33 31.43 0.383 19 43 24 68 0.63 1.42 0.79 2.23 and Nevis Saint Lucia 112 28 25.00 0.007 23 52 29 83 0.65 1.46 0.82 2.33 Saint Vincent 156 36 23.08 0.018 13 30 17 49 0.48 1.08 0.62 1.76 and the Grenadines Saint 38 28 73.68 0.011 32 71 39 111 0.89 1.98 1.09 3.11 Martin (FR) Sint Maarten 25 24 96.00 0.002 33 74 41 115 0.80 1.78 0.97 2.77 (NL) Suriname 448 9 2.01 0.111 58 129 71 203 0.52 1.16 0.64 1.83 Trinidad and 533 74 13.88 0.004 43 96 53 151 3.18 7.12 3.94 11.17 Tobago Turks and Caicos 439 353 80.41 0.088 21 47 27 76 7.54 16.53 9.54 26.85 Islands Total 13121 5525 42 0.037 27 61 35 98 151.69 334.26 192.33 542.60 It is not trivial to estimate the damage connected with the projected shoreline retreat and coastal land loss of sandy beaches, due to the large special scale of the study and the inter and intra-country variability in land value and connected activities. Nevertheless, in order to produce an estimate of potential impacts due to the shoreline retreat, we assumed a land value based on the costs of land reclamation that would be needed to reset to the situation of 2010 (Deltares, 2017). We assume two values for the costs of sand nourishment which would be required; a lower value of 10 USD/m3 and a higher value of 50 USD/m3. Additionally, we assume an active height for the sandy profiles of 4 m, which we assume to be affected (retreating) as a consequence of SLR. This means that the damage of losing one square meter of land can be 40 or 200 USD respectively. SLR induced erosion was assumed to be linear over the entire 21st century, so that a yearly cost of required nourishments could be estimated. The results indicate that over all the Caribbean countries included in the present study the annual relative damages due to SLR induced shoreline retreat can be 0.09% (0.14%) of the GDP under RCP 4.5 (RCP 8.5), when assuming a 10 USD/m3 sand nourishment cost (Appendix Table A3). On the other hand, when a assuming a higher nourishment cost equal to 50 USD/m3 the yearly damages 40 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 could be 0.43% (0.70%) of GDP, under RCP 4.5 (RCP 8.5) (Appendix Table A4). The country with the highest annual damages would be Bahamas in view of the largest expected land loss. These values represent lower and upper bounds, based on the assumption that all the land that is lost due to SLR would be reclaimed back (i.e., all sandy beaches have the same value), which is in reality not the case. For example, The Bahamas have an extended coastal length of sandy beaches but not all of these sandy beaches have a high economic value as most of the islands are not inhabited. Therefore, this is an overestimation of the actual damages associated to coastal erosion. At the same time, indirect costs resulting from the loss of these beaches have not been included in this study. Figure 21: Potential land loss at sandy erodible coasts per country at 2100 under RCP 8.5 (following SLR prediction by Vousdoukas et al., 2018b). 5.5 Strategies for adaptation: the St. Lucia case study 5.5.1 Introduction In this section, different possible options for adaptations are described for one case study: the country of St. Lucia. The country is characterized by a combination of both steep coasts and low- lying beaches and town, therefore representative of different type of coastline typologies. We would like to emphasize that this example is purely exploratory, to show how the information derived from this study can be used to distinguish areas more susceptible to the effects of flooding and SLR and define possible options for adaptation. The study is therefore not meant as a feasibility or detailed design assessment of different adaptation measures, for which a more detailed study accompanied by the use of higher accuracy local data would be required. 5.5.2 Sea level rise impact at St. Lucia The maps in Figure 23 and Figure 24 show the flood hazard at St Lucia for the 2010 and 2100 situation under RCP 8.5 scenario, for a flood event with a return period of 100 years. The figure 41 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 shows a zoom at five sites on the island where coastal towns and the presence of critical infrastructures can be found. In particular: George F.L. Charles Airport, Anse La Raye, Laborie (on the left) and Dennery and Vieux Fort (on the right). All these towns are located in low lying areas and, for the case of Anse La Raye, Laborie and Dennery, they are characterized by sandy beaches. Dennery beaches are also protected by a series of breakwater fronting the sandy beaches (Figure 22). The flood maps are representative of a worse-case scenario excluding any adaptation, which is not part of the model. A few general highlights can be depicted from the flood maps: ▪ All locations are prone to flooding already in the current situation during extreme events (100-year return period event) (Figure 23). High flood hazards characterize in particular the town of Anse La Raye and the two ports located at Dennery and Vieux Fort. ▪ The flood hazard is increasing in the 2100 situation (Figure 24). Assuming that no adaptation is taking place, flooding will start affecting the taxiway and runway of George F.L. Charles Airport and most of the assets at Anse La Raye. The results are in line with other recent studies on flood hazard in the country (e.g. Monioudi et al., 2018). ▪ As a result of the topography of the island, mainly characterized by hills and mountains up to 950 m high, only the areas adjacent to the coast will be flooded as a result of SLR. ▪ Socio-economic growth without proper planning will increase the number of properties subject to flood risk. Additionally, as discussed in Section 5.4, SLR is likely to have a considerable impact on the sandy beaches of the island, which accounts for 25% of the entire coastline length. According to estimates presented in this study, the average retreat of the sandy beaches induced by SLR, may be ranging roughly between 50 and 80 m according to RCP 4.5 and RCP 8.5 scenarios. The resulting total land loss will be ranging between ~1.5 km2 and 2.3 km2. George F.L. Charles Airport Anse La Raye Laborie Dennery Vieux Fort 42 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Figure 22 From top to bottom: George F.L. Charles Airport; Anse La Raye, Laborie, Dennery and Vieux Fort. George F.L. Charles Airport Anse La Raye Dennery Vieux Fort Laborie Figure 23 Flood hazard assessment at St. Lucia for the 2010 situation, during a flood event with a return period of 100 years. The flooded area is shaded in light blue, the building assets in orange. The five insets show areas potentially prone to flooding: George F.L. Charles Airport, Anse La Raye, Laborie (on the left) and Dennery and Vieux Fort (on the right). 43 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 George F.L. Charles Airport Anse La Raye Dennery Vieux Fort Laborie Figure 24 Flood hazard assessment at St. Lucia for the 2100 situation (RCP 8.5 according to Vousdoukas et al. 2018a), during a flood event with a return period of 100 years. The five insets show areas potentially prone to flooding: George F.L. Charles Airport; Anse La Raye, Laborie (on the left) and Dennery and Vieux Fort (on the right). 5.5.3 Strategies for adaptation In general, three different options exist to adapt to coastal hazards and SLR (see e.g. IPCC, 1990). Figure 25 Different coastal adaptation strategies: protect, accommodate and retreat. ▪ Protect: defend vulnerable areas, especially population centers, economic activities and natural resources. ▪ Accommodate: continue to occupy vulnerable areas but accept the greater degree of the coastal hazard (e.g. flooding, coastal erosion), by changing land use, construction methods and/or improving preparedness. ▪ (Planned) retreat: abandon structures in currently developed areas, resettle inhabitants and require that new developments be set back from the shore, as appropriate. For the case of St. Lucia, the (planned) retreat option should be explored for houses and (some of) the economic activities located in the low-lying areas close to the coastline, in view of the topography of the island and the fact that only area adjacent to the coast may be prone to the effects of SLR (Section 5.5.2). This will imply also adapting the road network of the island to provide accessibility at areas located more inland. Potential areas for future development have been explored in Section 44 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 5.5.4 and should be further analyzed in combination with a coastal zone management plan of the island. The plan should take into account the expected changes resulting from SLR and possible other socio-economic changes. This will however not be possible for some major infrastructures on the island (e.g. airport and ports) and for which the viable strategies will be either to protect (e.g. using seawalls) or accommodate (e.g. by increasing the height of the runway of the airport). Anse La Raye, Laborie and Dennery are fronted by sandy beaches. At those location, coastal protection strategies using sand (e.g. sand nourishments) may be explored, provided that sufficient sand is available. Sand nourishments are considered a nature-based type of solution, with the advantage of being flexible, and easy to adapt over time to compensate for a faster or lower rate in SLR. Additionally, sand nourishments could preserve the touristic value of the beaches, which is very important for the country. When the use of nature-based type of solution is not possible (e.g. if not sufficient sand is available), the use of hard measures (possibly in combination with beach nourishments (i.e. hybrid solution)) could also be explored as for the case of Dennery. In this case, it is important to realize that hard defenses may reduce the effect of flooding on the island but will not compensate for the lack of sediment induced by the rise in MSL. One important aspect to consider are the coastal ecosystems presents around the island. In particular: coral reefs and mangrove systems, as shown in Figure 26 which seem to characterize especially the eastern and southern part of the island. Coastal protection interventions should minimize the impact on those ecosystems and possibly making use of these ecosystems as nature- based solution. Both systems, if well-preserved and functioning can in fact stimulate wave breaking and therefore reduce the effect of SLR. Additionally, the effect of SLR combined with the effect of episodic storm events, could be mitigated by the use of early warning system, to anticipate the impact of these events. Uncertainties in the rate and magnitude of SLR complicate the decision making on coastal adaptation (Haasnoot et al., 2020). In this study, the upper scenarios which could arise from potential ice mass-loss from Antarctica and that could rapidly increase SLR in the second half of the century were not taken into consideration. This could have important implications not only on the measures to be applied (e.g. sand availability, performance of solutions, etc. to compensate possible extreme SLR) but also on the lead time available to plan and implement the selected measures. With an increase in rate of SLR, this lead time will reduce, implying that measures will have to be planned and implemented quicker and/or planned even before the first signals of the effects of SLR effects will manifest. The information has been summarized in Table 8 where different alternatives for the St Lucia case study are listed under each of the options: accommodate, protect and retreat. In the same table, SLR scenarios are shown describing the range between which each option could become useful (lower limit) and until when they may be effective (upper limit or “tipping point”) (see e.g. Haasnoot et al., 2019). The table shows how most of solutions could be effective for standard SLR scenario (“low”, “medium”, “high”; e.g. below  1 m). For extreme SLR conditions (i.e. several meters), most of the options under “adapt” or “protect” may no longer be feasible/effective, and therefore setback zones and relocation zones should be explored for larger portions of the coastal towns. 45 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Figure 26 Maps with coral reef (left) and mangrove (right) location at St. Lucia (https://data.unep- wcmc.org/datasets/1). 46 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Table 8 Possible adaptation options for St Lucia and range of applicability for different sea level rise scenarios. The rectangle (continues line) indicate the SLR range/limits under which each option could become useful (lower limit) and until when they may be effective (upper limit or “tipping point”). The dashed line indicates uncertainty in the timing of a tipping point, after which the option may no longer be effective. Adaptation SLR magnitude Strategy Accommodate Low Medium High Extreme (i.e. melting Antarctica) Adapt building code houses & other infrastructures Adapt drainage systems & pumps Increase height airport Early warning systems Protect Beach nourishments Structures (e.g. revetments, seawalls, groynes) Hybrid Coral/mangrove restoration Retreat Planned no build zones (setback) Relocation Adaptive ICZM plan 5.5.4 Identification of areas suitable for future development Hazard maps can also be used as a planning tool for future development. We showcase this concept using the St. Lucia example and the flood map presented in Figure 24 for the five coastal towns: George F.L Charles Airport, Anse la Raye, Laborie, View Port and Dennery. Figure 27 show maps for the different coastal towns constructed by overlaying the 100-year return period hazard in 2100 (RCP 8.5) with the building footprint, the road infrastructure and the slope of the terrain (derived from the local available DEM). Areas marked in green show areas in a relatively 47 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 close proximity to the flood hazard zone (~1.5 km), with enough space to increase current urbanization, close to existing road infrastructure and with slopes between 0 and 10%3. These areas could be suitable for potential future development and possibly relocation from areas subject to flooding. Areas at (or close to) airports were not marked as suitable for relocation since nuisance, safety and future airport development have to be considered. In the maps the most extreme hazard situation (100-year return period event in 2100 using RCP 8.5) is portrayed (see blue regions in maps). Buildings and assets in locations with flood depths between 0.5 and 1 m or higher (dark blue regions), should be given priority in possible retreat initiatives, followed by locations with flood depths lower than 0.5 m (light blue regions). The effect of possible coastal erosion is also not included in these maps. A first inspection using Google Earth was done to exclude immediate spatial use conflicts. However, the regions identified in the maps do not consider current detailed information on land use, ownership of land and/or zones susceptible to other type of natural hazards (like riverine flooding, landslides or earthquakes). All these factors are important for the identification of suitable relocation areas and should be part of a more detailed future planning study. —————————————— 3 Slopes between 6% and 9% are the maximum values indicated for wheelchair accessibility and comfortable pedestrian walk (Vernon et al., 2013). These values are low enough to reduce landslide risk and reduce the required earthworks to level up a region for construction. 48 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 1 2 3 4 5 Figure 27 Overview of areas (in green) potentially suitable for future development at George F.F. Charles Airport region (figure 1), Anse la Raye (figure 2), Laborie (figure 3), View Port (figure 4), Dennery (figure 5). 49 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 6 Discussion 6.1 Perspective results to other studies In this section we perform a brief literature review to compare our results to other similar studies. Giardino et al. (2018) performed a coastal flooding risk assessment at Ebeye (Marshall Islands), highlighting that the EAAD of the island could increase by a factor of 3-4 by the end of the century due to the effect of climate change. This value is comparable to our findings, which indicate an increase of the EAAD in the 18 countries analyzed by 145% under RCP 8.5 by 2100 (Table 6), i.e. an almost 2.5 increase in EAAD, with individual countries having values of up to 4 times. In a recent report of the World Bank (Building the Resilience of the Poor in the Face of Natural Disasters; Hallegatte et al., 2017) it was highlighted how low-income countries are usually the ones that are dealing with the highest losses relatively to their GDP. The study assessed the combined impact by various historic hazards (river floods, coastal floods due to storm surge, windstorms, earthquakes, and tsunamis) for more than 100 countries around the world. This is in line with our findings, where Suriname, Guyana, Jamaica, Dominica and Haiti are ranked high in the estimated losses with respect to their GDP. In the same study from the World Bank, the annual risk to assets was estimated as 1.18 % of the GDP for Dominican Republic (including all the aforementioned hazards), while in our study we calculated a value of 0.15% for the baseline year and only for the coastal flooding. A study focusing on the CARICOM member states (Modelling the Transformational Impacts and Costs of Sea Level Rise in the Caribbean; Simpson et al., 2010) reported that the combined flooding and erosion damages can reach a value of 0.9%-1.2% of the GDP by 2050. This is in the same order of magnitude of the results presented in our study for our flooding estimate, in which we estimated EAAD equal to 0.50%-0.53% of the GDP by 2050 over the 18 countries. In the same study, nearly 1,300 km2 is projected to be lost with a 1 m SLR, while in our study we project a loss of 542 km2 under RCP 8.5 by 2100 (with an average SLR close to 0.8 m). Differences can be attributed to the updated datasets and scenarios that we have used in the current study. A study by Hinkel et al. (2010) focusing on the European Union assessed the risks from SLR and found the highest damage costs relative to national GDP to be 0.3% at 2100 for the Netherlands, while all other countries had values that did not exceed 0.1%. These values are quite lower in comparison to our estimations, but this could be potentially explained by the methodological differences of our study to Hinkel et al. (2010) as for example: a) the previously lower estimates of SLR scenarios used, b) the exclusion of important processes like the wave setup in the flooding calculation or the c) use of lower quality topography data. Nevertheless, Hinkel et al. (2010) reports 5-fold increases in damage costs by the end of the century which is in the same order of magnitude with values reported herein. 50 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 6.2 Assumptions A number of assumptions have been used to carry out this study. These assumptions are discussed below and grouped under different items. 6.2.1 Sea level rise and climate change scenarios The study was based on median values derived from one set of RCP scenarios (RCP 4.5 and RCP 8.5) as published by the Joint Research Centre (Vousdoukas et al., 2018b).The uncertainty range within this set of median values has not been investigated. Additionally, different climate models can lead to different SLR projections. This study did not look into the variability which would arise from the use of different climate models. For example, the uncertainties in the rate and magnitude of SLR which could follow from potential ice mass-loss from Antarctica has not explicitly been considered in this study. As a consequence, extreme SLR have not been considered (higher percentiles than median). Similarly, SLR resulting from very stringent pathways (e.g. RCP 2.6) were not modelled. The study assumes that SLR will be the only parameter which will change as a result of climate change. Other variables (e.g. wave height, storm surge, tide) were assumed to be constant in time. This assumption seems reasonable given the focus of the study on the effects of SLR and the fact that other variables are assumed to be less influenced by climate changes effects. 6.2.2 Flood hazard The flood hazard assessment included a number of assumptions. First of all, the focus is on coastal flooding only, therefore discarding other forms of flooding (i.e. from river flooding, rainfall, etc). Extreme storm surges and offshore wave conditions were derived from storm events with different return periods, as estimated by global models. These global models do not explicitly resolve individual hurricane events. This assumption was assumed to be plausible, given the fact that the scope of the study was on assessing the relative effect that SLR may have on coastal flooding rather than on estimating the impact of individual extreme events. The estimation of flood hazards was done based on local DEM (when available) and, alternatively, by using globally available DEM on a 90 m resolution. The study pointed out the local differences in estimated flood hazard which may arise by the use of lower resolution global data. This difference becomes more evident especially at islands characterized by steep slopes (e.g. Grenada, Section 5.2.2). As a result of the coarse DEM and model resolution used, and the lack of local information, coastal protection measures could not be explicitly modelled in the flood hazard assessment. To take into account, in a very simplistic way, the effect of protection measures and prevent the overestimation of the total risks which would follow by excluding the presence of possible protection measures, the flood risk assessment was carried out by assuming a protection level of 1:10 years. This means that it was assumed that flood events with a return period lower than 1:10 years would not cause any damage. This is an assumption based on expert judgement and other similar regional frameworks (as the LISCOAST framework). In reality, protection levels will change dependent on the country and the land use. The effect of waves was included in the study by adding an estimated wave setup component to the total water level, computed as 20% of the offshore significant wave height. This simple approach does not require a full wave modelling to transform offshore conditions to nearshore conditions. However, this may result in an underestimation of the flood hazard as the dynamic component of flooding (runup, overtopping) resulting from individual waves is not explicitly included. This assumption is typical for regional flood hazard studies since no viable alternatives, computationally feasible and easy to implement, are yet available (see e.g. Vousdoukas et al., 2018). 51 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 6.2.3 Flood risk The estimation of risks induced by flooding includes several assumptions, as described in Section 4.2.3. These include: • The use of global population and building footprints • The estimated average value of assets per country derived based on global datasets (Huizinga et al., 2017) • The use of a regional depth-damage function estimated as a weighted averaged from different damage functions • A factor equal to 2 assumed to estimate damages to other items which are non-residential • A factor equal to 1.2 assumed to estimate indirect damages • The use of an averaged household size for the country • The use of a threshold in flood depth of 10 cm to identify people affected by flooding (i.e. people affected are the ones experiencing flooding, with a flood depth larger than 10 cm) As shown in recent studies, uncertainties resulting from the use of different depth-damage functions, protection levels and global DEM are the highest factor of uncertainties in flood risk studies in the current situation (Parodi et al., 2020; Vousdoukas et al., 2018c; Wagenaar et al., 2016). Differences between SLR scenarios will become of similar (or higher) relative importance towards the end of the century. While the effect of single variable on estimated damages can be up to 200% (Parodi et al., 2020; Vousdoukas et al., 2018c), the superimposition of the different effects can range between 200% and 500% (Wagenaar et al., 2016). Additionally, the estimation of the future development of flood risks, taking into account SLR in combination with socio-economic changes, was carried out assuming the validity of global SSP scenarios. These scenarios may not be fully representative for small SIDS countries, which may be more influenced by local national socio-economic developments. The estimation of the combined effect of SLR and SSP scenarios, as carried out in the present study may be an overestimation of the actual future flood risks. The assumption is that socio-economic development will continue in the same areas as in the current situation. In reality, people will tend to adapt (i.e. either by protecting, adapting or retreating) and therefore the risk development will most likely follow a less pronounced increase. 6.2.4 Sandy beach erosion and associated damages To estimate the erosion of sandy beaches, a number of assumptions were made. First of all, sandy beaches were identified by using satellite images, as published in Luijendijk et al. (2018). This means the detection is dependent on the resolution of available images and on algorithms for automatic detection of the sandy beaches, which was however done following state-of-the art methodologies. To translate SLR to potential coastal retreat, the global dataset of coastal slopes by Athanasiou et al. (2019) was used and applied in combination with the Bruun rule (Bruun, 1962). Although widely used, the Bruun rule is a simplification of the actual coastal retreat process, which assumes that the coastal profile can migrate landward by a rate estimated trough the SLR rate and the coastal slope. The assumption is that the cross-shore coastal profile can indeed move landward and therefore there is no structure behind the beach which could prevent such a landward migration and/or that beaches are sufficiently wide. This may not always be the case at many Caribbean beaches, which are relatively narrow and often characterized by infrastructures behind the beaches which could 52 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 prevent this landward shift. Therefore, the estimated erosion is only representative of the “potential” erosion, while the “actual” erosion may be of a lower value. Although the global dataset by Athanasiou et al. (2019) has been validated by means of several coastal profiles worldwide, it still remains a global dataset, which may not capture all the local variability. Athanasiou et al. (2020) assessed the uncertainty in the projections of sandy beach erosion due to sea level rise at the European scale. According to the authors, in the current situation the choice of the sandy beach locations, the nearshore slope and the sea level rise scenario lead to similar uncertainties on the prediction of the coastal retreat, while the uncertainty inherent to the choice of sea level rise scenario becomes the major source of uncertainty in 2100. It is also important to realize that the Bruun rule only estimates the potential retreat induced by SLR while erosion induced by short-term storm events are not accounted for. On a longer time-scale, and assuming there is no additional anthropogenic interference, this assumption is generally acceptable as beaches are in a dynamic equilibrium and therefore erosion during a storm event is compensated by accretion during periods with milder waves. Other type of erosion (e.g. induced by the presence of ports and/or cross-shore structures) are also not included in the study. Finally, the study only focusses on potential erosion of sandy beaches, while erosion of other type of coastlines are not explicitly accounted for. The estimation of damages related to beach erosion was based on the assumption that all potential beach erosion would be compensated by nourishments to counteract the erosion, and assuming a range of average nourishment cost with a lower (i.e. 10 USD/m3) and higher (i.e. 50 USD/m3) value, valid for the entire region. In reality, the value of beaches is different from country to country and within the country. Moreover, it is not expected that all beaches will be restored to the 2010 situation. Furthermore, we have not included as damage induced by erosion, additional indirect damages related e.g. by loss in economic activities dependent on beach use. The erosion costs have been spread equally over the century, assuming a linear rate of erosion until 2100. 6.2.5 Coastal adaptation The discussion about options for coastal adaptation are only focusing on one country (St. Lucia) and are purely exploratory, to show how the information derived from the study can be used to distinguish areas more susceptible to the effects of flooding and SLR and define possible options for adaptation. The study is therefore not meant as a feasibility or detailed design assessment of different adaptation measure, and for which a more detailed study accompanied by the use of higher accuracy local data and interaction with local stakeholders would be required. 6.3 Recommendations Following the discussion points as described in Section 6.2, the following recommendations are provided: • Collect, when possible, local high-resolution DEM data (e.g. by using lidar or drone techniques) as these will highly improve the quality of the flood hazard assessment and the evaluation of possible adaptation options. • Simulate additional (and more extreme) SLR scenarios to provide a broader range of possible impacts. • Include additional processes in the modelling framework. For example, other erosion processes (e.g. as a result of storms) and the impacts of hurricane events (i.e. including rainfall and wind). 53 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 • Include more accurate flood protection levels (e.g. accounting for differences between countries and difference between urbanized and non-urbanized areas). • The flood risk modelling includes several assumptions (e.g. value of assets, estimation of direct and indirect damages, etc). A validation of the flood risk modelling results by using local data from the different countries (e.g. after a flooding event), could provide useful insight on the estimated values. • The socio-economic development assessment could make use of more realistic local data accounting for urbanization trends in combination with future flood hazard. • To start a prioritization process of possible adaptation options for the different countries, it would be recommended to jointly discuss the results of the hazard modelling with different local stakeholders. The prioritization process should take into account existing plans and priorities at each country, possible existing limitations (i.e. physical, financial or institutional) which could preclude the implementation of some of the options, and future socio-economic developments foreseen. Such a prioritization process could set the basis for the setting up of a multi-sectoral investment plan for each country. • Further improvement of local models, in combination with the use of local data (in particular DEM), would be recommended for the testing of different solutions, therefore allowing the comparison between costs and benefits of different options. 54 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 7 Conclusions This study presents estimates of the effects of SLR4 for 18 countries in the Caribbean region, in terms of coastal flooding due to marine surge and wave setup and associated impacts and erosion of sandy beaches. Indicators were derived for each of the countries analyzed, to estimate the relative impact potential of SLR. Different scenarios were computed for different time horizons, storm events with different return periods, SLR scenarios (RCP5 4.5 and RCP 8.5, following estimates by Vousdoukas et al., 2018), and SSP6 pathways (SSP3 and SSP5, using the database of Gidden et al., 2019). The study estimates that the current (2010) EAAD 7 as a result of flooding, and excluding the effect of hurricane events, over the region is 0.37% of the total GDP in the Caribbean region. The effect of SLR only could lead to an increase in EAAD as % GDP at the regional level under RCP 4.5 (RCP 8.5) by almost 35% (45%) by 2050 and 85% (150%) by 2100, due to the effect of coastal flooding only. With respect to people affected, the current EAAP8 as % of the total population over all the countries is estimated to be almost 0.15%, a value which is expected to increase by almost 25% (40%) by 2050 and 75% (120%) by 2100 under RCP 4.5 (RCP 8.5). When considering both SLR and socioeconomic growth, the total risk is projected to grow significantly more than when considering SLR only. Expected average annual damage as % of GDP is expected to increase under the combination of RCP 4.5 and SSP3 (RCP 8.5 and SSP5) by almost 75% (120%) by 2050 and by 185% (300%) by 2100 relative to the 2010 values. EAAP is projected to increase of 25% (25%) by 2050 and 65% (75%) by 2100 under the combination of RCP 4.5 and SSP3 (RCP 8.5 and SSP5). The countries which will have a relatively larger impact (i.e. in terms of relative EAAD as % of the total GDP) are Guyana, Suriname and Saint Martin (FR) with values increasing from 3.1%, 3.4% and 1.4% respectively up to 6.9% (10.2%), 6.5% (8.8%) and 2.7% (3.6%) in 2100, when considering SLR scenario only under RCP 4.5 (RCP 8.5). The countries with the lowest impact are Turks and Caicos Islands, Trinidad and Tobago and Barbados with values increasing from 0.03%, 0.04% and 0.02% for 2010 up to 0.08% (0.10%), 0.05% (0.07%), 0.02% (0.02%) by 2100. The above mentioned values are under the assumption that protection levels against flooding are designed for flood events with a return period of 10 years (i.e. no flooding for events with a return period lower than 10 years) and no further adaptation is implemented before the end of the century. The values confirm that the resilience potential of some of these countries to SLR is low and further adaptation should be implemented to allow the economics of the countries to withstand the predicted impact. The result of the study can help prioritizing countries where adaptation is more urgent. The same three countries (i.e. Guyana, Suriname and Saint Martin) will also have the largest relative number of people affected by flooding, as most of the settlements are located in low-lying areas. On average, over all the 18 Caribbean countries of the study, a shoreline retreat of sandy beaches of 61 m (98 m) is projected under RCP 4.5 (RCP 8.5) by 2100. The country with the highest sandy beach loss is the Bahamas, where by 2100 almost 375 km2 of sandy land is projected to be lost due to SLR under RCP 8.5. This is the result of a combination of mild nearshore slopes, a quite long —————————————— 4 SLR = Sea Level Rise 5 RCP = Representative Concentration Pathway 6 SSP = Shared Socioeconomic Pathways 7 EAAD = Expected Average Annual Damages 8 EAAP = Expected Averaged Annual People 55 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 coastline (5,550 km) and a high percentage of sandy coastline (~62%). Other countries that are projected to lose a large amount of sandy beach area are Belize, the Turks and Caicos Islands, the Dominican Republic and Haiti with 35, 27, 23 and 20 km2 projected to be lost by 2100 under RCP 8.5. Under the same scenario, a total of 192 km2 (542 km2) is expected to be lost over all the 18 Caribbean countries of the present study by 2050 (2100). A reduction in carbon emission (i.e. following a RCP 4.5 scenario) can play an important role in reducing projected land loss by almost 20% by 2050 and almost 40% by 2100, relative to the high-emission RCP 8.5 scenario. As for many of these countries the touristic sector is crucial to their economies, these values indicate that the state of these sandy beaches should be monitored carefully. This should include the preparation of a sediment and shoreline management plan for each country, to adapt to the possible consequences of SLR and the expected shortage of sandy sediments on the beaches and foreshores. The results from the study were used to investigate possible options for adaptation at one specific country (i.e. St. Lucia) and should be further discussed and verified with local stakeholders as part of future studies. The results showed that several of the coastal settlements and infrastructures in the country are already prone to flooding in the current situation. The situation will aggravate in 2100 with an increase in the areas prone to flooding, which will also start affecting some critical infrastructures (e.g. local airport). A planned retreat option seems to be applicable for some of the assets located in areas more vulnerable to the effects of SLR. For some specific assets (e.g. airport) the only viable options will be to either to protect (e.g. using seawalls) or accommodate (e.g. by increasing the height of the runway of the airport). Given the importance of the sandy beaches for the country touristic sector, the use of nature-based approaches (e.g. using sand nourishments) should be explored to maintain the sandy beaches, assuming that sufficient sand (and of the right quality) is available. When sand is not sufficient, this could possibly be supplemented by the use of hard-structures (e.g. groynes, breakwaters). Existing coastal ecosystems (e.g. coral reefs, mangroves, etc.) should be carefully studied to capitalize on their beneficial effect as natural coastal protection and to minimize the impact on these ecosystems induced by human interference and/or climate change. The uncertainties in the acceleration of SLR which could arise from potential ice mass-loss from Antarctica was not explicitly taken into account in this study. This could have important implications not only on the measures to be applied but also on the lead time available to plan and implement the selected measures. 56 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 8 References Athanasiou, P., van Dongeren, A., Giardino, A., Vousdoukas, M., Gaytan-Aguilar, S. and Ranasinghe, R.: Global distribution of nearshore slopes with implications for coastal retreat, Earth Syst. Sci. Data, 11(4), 1515–1529, doi:10.5194/essd-11-1515-2019, 2019. Athanasiou, P., van Dongeren, A., Giardino, A., Vousdoukas, M.I., Ranasinghe, R., Kwadijk, J. Uncertainties in projections of sandy beach erosion due to sea level rise: an analysis at the European scale. Scientific Reports, 2020. AVISO: MDT CNES-CLS18, Available from https//www.aviso.altimetry.fr/en/data/products/auxiliary- products/mdt.html, 2018., 2018. Bamber, J. L. and Aspinall, W. P.: An expert judgement assessment of future sea level rise from the ice sheets, Nat. Clim. Chang., 3(4), 424–427, doi:10.1038/nclimate1778, 2013. De Bruijn, K. M.: Resilience and flood risk management: a systems approach applied to lowland rivers., 2005. De Bruijn, K. M., Wagenaar, D. J., Slager, K., De Bel, M. and Burzel, A.: Updated and improved method for flood damage assessment., 2014. Bruun, P.: Sea-Level Rise as a Cause of Shore Erosion, J. Waterw. Harb. Div., 88(1), 117 –132, 1962. Central Intelligence Agency - Library: The World Factbook, https://www.cia.gov/library/publications/the-world-factbook/geos/rn.html, 2020. Chatterton, J., Viavatenne, C., Morris, J., Penning-Rowsell, E. and Tapsell, S.: Delivering benefits through evidence: The costs of the summer 2007 floods in England. Environmental Agency., 2010. Church, J.A., P.U. Clark, A. Cazenave, J.M. Gregory, S. Jevrejeva, A. Levermann, M.A. Merrifield, G.A. Milne, R.S. Nerem, P.D. Nunn, A.J. Payne, W.T. Pfeffer, D. Stammer and A.S. Unnikrishnan, 2013: Sea Level Change. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. De Conto, R. and Pollard, D.: Contribution of Antarctica to past and future sea-level rise, Nature, 531, 591–597, 2016. Farr, T. G., Rosen, P. A., Caro, E., Crippen, R., Duren, R., Hensley, S., Kobrick, M., Paller, M., Rodriguez, E., Roth, L., Seal, D., Shaffer, S., Shimada, J., Umland, J., Werner, M., Oskin, M., Burbank, D. and Alsdorf, D.: The Shuttle Radar Topography Mission, Rev. Geophys., 45(2), RG2004, doi:10.1029/2005RG000183, 2007. Freire, S., MacManus, K., Pesaresi, M., Doxsey-Whitfield, E. and Mills, J.: Development of new open and free multi-temporal global population grids at 250 m resolution., in Proceedings of the 19th AGILE Conference on Geographic Information Science., 2016. 57 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Fujisada, H., Bailey, G. B., Kelly, G. G., Hara, S. and Abrams, M. J.: ASTER DEM performance, IEEE Trans. Geosci. Remote Sens., 43(12), 2707–2714, doi:10.1109/TGRS.2005.847924, 2005. Gauderis, J. and Kind, J. M.: MKBA WV21 Bijlage D: Bewerking schadegegevens. Deltares 1204144-006., 2012. Giardino, A., Nederhoff, K. and Vousdoukas, M.: Coastal hazard risk assessment for small islands: assessing the impact of climate change and disaster reduction measures on Ebeye (Marshall Islands), Reg. Environ. Chang., doi:10.1007/s10113-018-1353-3, 2018. Gidden, M. J., Riahi, K., Smith, S. J., Fujimori, S., Luderer, G., Kriegler, E., van Vuuren, D. P., van den Berg, M., Feng, L., Klein, D., Calvin, K., Doelman, J. C., Frank, S., Fricko, O., Harmsen, M., Hasegawa, T., Havlik, P., Hilaire, J., Hoesly, R., Horing, J., Popp, A., Stehfest, E. and Takahashi, K.: Global emissions pathways under different socioeconomic scenarios for use in CMIP6: a dataset of harmonized emissions trajectories through the end of the century, Geosci. Model Dev., 12(4), 1443–1475, doi:10.5194/gmd-12-1443-2019, 2019. Haasnoot, M., Brown, S., Scussolini, P., Jimenez, J. A., Vafeidis, A. T. and Nicholls, R. J.: Generic adaptation pathways for coastal archetypes under uncertain sea-level rise, Environ. Res. Commun., 1(7), 71006, doi:10.1088/2515-7620/ab1871, 2019. Haasnoot, M., Kwadijk, J., van Alphen, J., Le Bars, D., van den Hurk, B., Diermanse, F., van der Spek, A., Essink, G. O., Delsman, J. and Mens, M.: Adaptation to uncertain sea-level rise; how uncertainty in Antarctic mass-loss impacts the coastal adaptation strategy of the Netherlands, Environ. Res. Lett., 15(3), 34007, doi:10.1088/1748-9326/ab666c, 2020. Hallegatte, S., Green, C., Nicholls, R. J. and Corfee-Morlot, J.: Future flood losses in major coastal cities, Nat. Clim. Chang., 3(9), 802–806, doi:10.1038/nclimate1979, 2013. Hallegatte, S., Vogt-Schilb, A., Mook, B. and Rozenberg, J.: Unbreakable: Building the Resilience of the Poor in the Face of Natural Disasters., Washington, DC., 2017. Hinkel, J. and Klein, R. J. T.: Integrating knowledge to assess coastal vulnerability to sea-level rise: The development of the DIVA tool, Glob. Environ. Chang., 19(3), 384 –395, doi:10.1016/j.gloenvcha.2009.03.002, 2009. Hinkel, J., Nicholls, R. J., Vafeidis, A. T., Tol, R. S. J. and Avagianou, T.: Assessing risk of and adaptation to sea-level rise in the European Union: an application of DIVA, Mitig. Adapt. Strateg. Glob. Chang., 15(7), 703–719, doi:10.1007/s11027-010-9237-y, 2010. Hinkel, J., Nicholls, R. J., Tol, R. S. J., Wang, Z. B., Hamilton, J. M., Boot, G., Vafeidis, A. T., McFadden, L., Ganopolski , A. and Klein, R. J. T.: A global analysis of erosion of sandy beaches and sea-level rise: An application of DIVA, Glob. Planet. Change, 111, 150–158, http://doi.org/10.1016/j.gloplacha.2013.09.002, 2013. Hinkel, J., Lincke, D., Vafeidis, A. T., Perrette, M., Nicholls, R. J., Tol, R. S. J., Marzeion, B., Fettweis, X., Ionescu, C. and Levermann, A.: Coastal flood damage and adaptation costs under 21st century sea-level rise, Proc. Natl. Acad. Sci., 111(9), 3292–3297, doi:10.1073/pnas.1222469111, 2014. Hsiang, S.M. and Jina, A.S.: The Causal Effect of Environmental Catastrophe on Long-Run Economic Growth: Evidence from 6,700 Cyclones. NBER Working Paper Series. National Bureau of Economic Research. 58 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Huizinga, J., Moel, H. de and Szewczyk, W.: Global flood depth-damage functions: Methodology and the database with guidelines, JRC Work. Pap. JRC105688, Jt. Res. Cent. (sev. site), 2017. IMF, 2016. “Small States” Resilience to Natural Disasters and Climate Change-Role for the IMF. IMF Policy paper, Washington, D.C. IPCC, 1990. Strategies for adaptation to sea level rise. Report of the IPCC Coastal Zone Management Subgroup: Intergovernmental Panel on Climate Change. Geneva: Intergovernmental Panel on Climate Change. IPCC: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia,., 2013. Krieger, G., Moreira, A., Fiedler, H., Hajnsek, I., Werner, M., Younis, M. and Zink, M.: TanDEM-X: A Satellite Formation for High-Resolution SAR Interferometry, IEEE Trans. Geosci. Remote Sens., 45(11), 3317–3341, doi:10.1109/TGRS.2007.900693, 2007. Kron, W.: Flood Risk = Hazard • Values • Vulnerability, Water Int., 30(1), 58–68, doi:10.1080/02508060508691837, 2005. Kulp, S. A. and Strauss, B. H.: CoastalDEM: A global coastal digital elevation model improved from SRTM using a neural network, Remote Sens. Environ., 206, 231 –239, doi:10.1016/j.rse.2017.12.026, 2018. Leijnse, T. W. B.: Computationally Efficient Modelling of Compound Flooding due to Tropical Cyclones with the Explicit Inclusion of Wave-Driven Processes: Research into the required processes and the implementation within the SFINCS model, Delft Univ. Technol. Master thesis [online] Available from: https://repository.tudelft.nl/islandora/object/uuid%3A53037d4d-d7a9-4f71- 9b22-8c329f2b384c?collection=education, 2018. Leijnse, T. W. B., van Ormondt, M., Nederhoff, K. and van Dongeren, A.: Modeling compound flooding in coastal systems using a computationally efficient reduced-physics solver: including fluvial, pluvial, tidal, wind- and wave-driven processes, Coast. Eng. under Rev., 2020. Li, F.: Probabilistic Estimation of Dune Erosion and Coastal Zone Risk, Delft University of Technology., 2014. Luijendijk, A., Hagenaars, G., Ranasinghe, R., Baart, F., Donchyts, G. and Aarninkhof, S.: The State of the World’s Beaches, Sci. Rep., 8(1), 6641, doi:10.1038/s41598-018-24630-6, 2018. Marconcini, M., Metz-Marconcini, A., Üreyen, S., Palacios-Lopez, D., Hanke, W., Bachofer, F., Zeidler, J., Esch, T., Gorelick, N., Kakarla, A. and Strano, E.: Outlining where humans live -- The World Settlement Footprint 2015, [online] Available from: http://arxiv.org/abs/1910.12707, 2019. Mentaschi, L., Vousdoukas, M. I., Voukouvalas, E., Dosio, A. and Feyen, L.: Global changes of extreme coastal wave energy fluxes triggered by intensified teleconnection patterns, Geophys. Res. Lett., 44(5), 2416–2426, doi:10.1002/2016GL072488, 2017. Monioudi, I. Ν., Asariotis, R., Becker, A., Bhat, C., Dowding-gooden, D., Esteban, M., Feyen, L., Mentaschi, L., Nikolaou, A., Nurse, L., Phillips, W., Smith, D. Α. Υ., Satoh, M., Trotz, U. O. D., Velegrakis, A. F., Voukouvalas, E., Vousdoukas, M. I. and Witkop, R.: Climate change impacts on critical international transportation assets of Caribbean Small Island Developing States (SIDS): the case of Jamaica and Saint Lucia, Reg. Environ. Chang., 18, 2211–2225, 2018. 59 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Muis, S., Verlaan, M., Nicholls, R. J., Brown, S., Hinkel, J., Lincke, D., Vafeidis, A. T., Scussolini, P., Winsemius, H. C. and Ward, P. J.: A comparison of two global datasets of extreme sea levels and resulting flood exposure, Earth’s Futur., 5(4), 379–392, doi:10.1002/2016EF000430, 2017. Muis, S., Irazoqui Apecechea, M., Dullaart, J., de Lima Rego, J., Skovgaard Madsen, K., Su, J., Yan, K. and Verlaan, M.: A global high-resolution dataset of extreme sea level, including future projections, Front. Mar. Sci. under Rev., 2020. Nicholls, R. and Cazenave, A., 2010. Sea-Level Rise and its impact on Coastal Zones. Science, 328(18), 1517-1520. O’Neill, B. C., Kriegler, E., Ebi, K. L., Kemp-Benedict, E., Riahi, K., Rothman, D. S., van Ruijven, B. J., van Vuuren, D. P., Birkmann, J., Kok, K., Levy, M. and Solecki, W.: The roads ahead: Narratives for shared socioeconomic pathways describing world futures in the 21st century, Glob. Environ. Chang., 42, 169–180, doi:10.1016/j.gloenvcha.2015.01.004, 2017. Palacios-Lopez, D., Bachofer, F., Esch, T., Heldens, W., Hirner, A., Marconcini, M., Sorichetta, A., Zeidler, J., Kuenzer, C., Dech, S., Tatem, A. J. and Reinartz, P.: New Perspectives for Mapping Global Population Distribution Using World Settlement Footprint Products, Sustainability, 11(21), 6056, doi:10.3390/su11216056, 2019. Palanisamy, Hindumathi, Becker, M., Meyssignac, E., Henry, O., Cazenave, A., 2012. Regional sea level change and variability in the Caribbean Sea since 1950. Journal of Geodetic Science 2(2): 125-133. Parodi, M., Giardino, A., Dongeren, A. van, Pearson, S. and Bricker, J. Uncertainties in Coastal Flood Risk Assessments in Small Island Developing States, Nat. Hazards Earth Syst. Sci. Discuss., 1–29, doi:10.5194/nhess-2019-392, 2020. Pekel, J.-F., Cottam, A., Gorelick, N. and Belward, A. S.: High-resolution mapping of global surface water and its long-term changes, Nature, 540(7633), 418–422, doi:10.1038/nature20584, 2016. Pesaresi, M., Ehrlich, D., Ferri, S., Florczyk, A. ., Freire, S., Halkia, S., Julea, A. M., Kemper, T., Soille, P. and Syrris, V.: Operating procedure for the production of the Global Human Settlement Layer from Landsat data of the epochs 1975, 1990, 2000, and 2014., 2016. Ranasinghe, R., Callaghan, D. and Stive, M. J.: Estimating coastal recession due to sea level rise: beyond the Bruun rule, Clim. Chang., 110(3–4), 561–574, 2012. Riahi, K., van Vuuren, D. P., Kriegler, E., Edmonds, J., O’Neill, B. C., Fujimori, S., Bauer, N., Calvin, K., Dellink, R., Fricko, O., Lutz, W., Popp, A., Cuaresma, J. C., KC, S., Leimbach, M., Jiang, L., Kram, T., Rao, S., Emmerling, J., Ebi, K., Hasegawa, T., Havlik, P., Humpenöder, F., Da Silva, L. A., Smith, S., Stehfest, E., Bosetti, V., Eom, J., Gernaat, D., Masui, T., Rogelj, J., Strefler, J., Drouet, L., Krey, V., Luderer, G., Harmsen, M., Takahashi, K., Baumstark, L., Doelman, J. C., Kainuma, M., Klimont, Z., Marangoni, G., Lotze-Campen, H., Obersteiner, M., Tabeau, A. and Tavoni, M.: The Shared Socioeconomic Pathways and their energy, land use, and greenhouse gas emissions implications: An overview, Glob. Environ. Chang., 42, 153 –168, doi:10.1016/j.gloenvcha.2016.05.009, 2017. Rosenqvist, A., Shimada, M., Ito, N. and Watanabe, M.: ALOS PALSAR: A Pathfinder Mission for Global-Scale Monitoring of the Environment, IEEE Trans. Geosci. Remote Sens., 45(11), 3307 – 3316, doi:10.1109/TGRS.2007.901027, 2007. 60 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Simpson, M. C., Scott, D., Harrison, M., Sim, R., Silver, N., O’Keeffe, E., Harrison, S., Taylor, M., Lizcano, G., Rutty, M., Stager, H., Oldham, J., Wilson, M., New, M., Clarke, J., Day, O. J., Fields, N., Georges, J., Waithe, R. and McSharry, P.: Quantification and Magnitude of Losses and Damages Resulting from the Impacts of Climate Change: Modelling the Transformational Impacts and Costs of Sea Level Rise in the Caribbean, Barbados, West Indies., 2010. Slager, K., Burzel, A., Bos, E., de Bruijn, K., Wagenaar, D., Winsemius, H., Bouwer, L. and van der Doef, M.: User Manual Delft-FIAT version 1, publicwiki.deltares.nl/display/DFIAT/Delft-FIAT+Home, 2016. Thinh, N.A., Hens, L. Tourism and beach erosion: valuing the damage of beach erosion for tourism in the Hoi An World Heritage site, Vietnam. Journal of Environment Development and Sustainability. DOI: 10.1007/s10668-018-0126-y, 2018. Tiggeloven, T., Moel, H. de, Winsemius, H., Eilander, D., Erkens, G., Gebremedhin, E., Diaz Loaiza, A., Kuzma, S., Luo, T., Iceland, C., Bouwman, A., Huijstee, J. van, Ligtvoet, W. and Ward, P.: Global scale benefit-cost analysis of coastal flood adaptation to different flood risk drivers, Nat. Hazards Earth Syst. Sci. Discuss., (November), 1–28, doi:10.5194/nhess-2019-330, 2019. Trevor, A., Inci, O., Uma, R. and Krisna, S.: Unleashing Growth and Strengthening Resilience in the Caribbean., 2017. UNDRR, 2019. Global Assessment Report on Disaster Risk Reduction 2019. 978-92-1-004180-5, 425 pp. United Nations: United Nations Statistics Division - National Accounts, https://unstats.un.org/unsd/snaama/Index, 2020. Vernon, S., Tennant, R. and Garmony, N.: Landscape Architect’s Pocket Book, 2 edition., Routledge, New York., 2013. Vilier, J. J. T. A.: Assessment of the costs of business interruption due to large scale floods: A comparison of methods., Master Thesis Delft University of Technology., 2013. Vousdoukas, M., Voukouvalas, E., Mentaschi, L., Dottori, F., Giardino, A., Bouziotas, D., Bianchi, A., Salamon, P. and Feyen, L.: Developments in large-scale coastal flood hazard mapping, Nat. Hazards Earth Syst. Sci., 16(8), 1841–1853, doi:10.5194/nhess-16-1841-2016, 2016. Vousdoukas, M., Mentaschi, Lorenzo Voukouvalas, E., Verlaan, M., Jevrejeva, S., Jackson, L. and Feyen, L.: Global Extreme Sea Level projections, Eur. Comm. Jt. Res. Cent. [Dataset], doi:10.2905/jrc-liscoast-10012, 2018a. Vousdoukas, M., Mentaschi, L., Voukouvalas, E., Verlaan, M., Jevrejeva, S., Jackson, L. P. and Feyen, L.: Global probabilistic projections of extreme sea levels show intensification of coastal flood hazard, Nat. Commun., 9(1), 1–12, doi:10.1038/s41467-018-04692-w, 2018b. Vousdoukas, M. I., Bouziotas, D., Giardino, A., Bouwer, L. M., Mentaschi, L., Voukouvalas, E. and Feyen, L.: Understanding epistemic uncertainty in large-scale coastal flood risk assessment for present and future climates, Nat. Hazards Earth Syst. Sci., 18(8), 2127 –2142, doi:10.5194/nhess- 18-2127-2018, 2018c. Vousdoukas, M. I., Ranasinghe, R., Mentaschi, L., Plomaritis, T. A., Athanasiou, P., Luijendijk, A. and Feyen, L.: Sandy coastlines under threat of erosion, Nat. Clim. Chang., 10(3), 260 –263, doi:10.1038/s41558-020-0697-0, 2020. 61 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Wagenaar, D.J., de Bruijn, K.M., Bouwer, L.M., and de Moel, H. Uncertainty in flood damage estimates and its potential effect on investment decisions. Journal of Natural Hazards and Earth System Sciences. https://doi.org/10.5194/nhess-16-1-2016, 2016. Wagenaar, D. J., Dahm, R. J., Diermanse, F. L. M., Dias, W. P. S., Dissanayake, D. M. S. S., Vajja, H. P., Gehrels, J. C. and Bouwer, L. M.: Evaluating adaptation measures for reducing flood risk: A case study in the city of Colombo, Sri Lanka, Int. J. Disaster Risk Reduct., 37(December 2018), 101162, doi:10.1016/j.ijdrr.2019.101162, 2019. World Bank Group: Data - Indicators - GDP (Current US$), https://data.worldbank.org/indicator/NY.GDP.MKTP.CD, 2020a. World Bank Group: Data - Indicators - total population, https://data.worldbank.org/indicator/SP.POP.TOTL, 2020b. Yamazaki, D., Ikeshima, D., Tawatari, R., Yamaguchi, T., O’Loughlin, F., Neal, J. C., Sampson, C. C., Kanae, S. and Bates, P. D.: A high-accuracy map of global terrain elevations, Geophys. Res. Lett., 44(11), 5844–5853, doi:10.1002/2017GL072874, 2017. 62 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Appendix Table A.1: Relative changes in coastal flood impact indicators (in %) for each country under SSP 3 and SSP 5 (Gidden et al., 2019) for 2010, 2050 and 2100 time horizons and assuming present day flood hazard constant through time. Baseline Baseline EAAD as % of GDP EAAP as % of population Country name GDP pop. SSP 3 SSP 5 SSP 3 SSP 5 (M USD) (x1,000) 2010 2050 2100 2050 2100 2010 2050 2100 2050 2100 Antigua and Barbuda 1,611 96 0.31 0.31 0.31 0.30 0.30 0.24 0.24 0.24 0.23 0.23 Bahamas 12,425 386 0.40 0.37 0.31 0.41 0.40 0.42 0.42 0.42 0.42 0.42 Barbados 5,145 287 0.02 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0.01 Belize 1,871 383 0.58 0.59 0.67 0.74 0.79 0.29 0.29 0.29 0.28 0.28 Dominica 551 72 0.88 0.89 0.96 1.03 1.08 1.49 2.01 2.39 2.12 2.49 Dominican Republic 85,555 10,627 0.20 0.21 0.23 0.25 0.26 0.06 0.06 0.06 0.06 0.06 Grenada 1,186 111 0.24 0.24 0.24 0.24 0.24 0.16 0.16 0.16 0.15 0.15 Guyana 3,879 779 3.62 4.45 5.22 5.24 6.14 1.65 1.65 1.65 1.60 1.60 Haiti 9,659 11,123 0.50 0.71 1.12 0.82 1.03 0.06 0.06 0.06 0.06 0.06 Jamaica 15,714 2,935 0.65 0.80 1.18 0.93 1.16 0.25 0.28 0.27 0.31 0.42 St. Kitts and Nevis 1,011 52 0.16 0.16 0.17 0.16 0.16 0.31 0.34 0.33 0.38 0.44 St. Lucia 1,922 182 0.38 0.38 0.43 0.47 0.56 0.14 0.14 0.14 0.14 0.14 St. Vincent and the Grenadines 811 110 0.44 0.35 0.40 0.45 0.44 0.07 0.07 0.07 0.07 0.07 Saint Martin (FR) 562 37 1.38 1.38 1.39 1.44 1.44 1.40 1.44 1.43 1.46 1.54 Sint Maarten (NL) 1,059 41 0.38 0.38 0.38 0.38 0.38 0.46 0.46 0.46 0.45 0.45 Suriname 3,591 576 5.57 5.65 6.40 6.66 7.27 1.46 1.46 1.46 1.44 1.44 Trinidad and Tobago 23,808 1,390 0.05 0.05 0.04 0.07 0.06 0.03 0.03 0.03 0.03 0.03 Turks and Caicos 1,022 38 0.02 0.05 0.05 0.05 0.05 0.04 0.07 0.07 0.07 0.06 63 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Table A.2: Relative changes in coastal flood impact indicators (in %) for each country under combined SLR and SSP projections: RCP 4.5 and RCP 8.5 (following SLR projections by Vousdoukas et al., 2018) and SSP 3 and SSP 5 (Gidden et al., 2019) for 2010, 2050 and 2100 time horizons. EAAD as % of GDP EAAP as % of population Baseline GDP (M USD) Baseline pop. (x1,000) RCP 4.5 RCP 8.5 RCP 4.5 RCP 8.5 SSP 3 SSP 5 SSP 3 SSP 5 SSP 3 SSP 5 SSP 3 SSP 5 Country 2010 2010 name 2050 2100 2050 2100 2050 2100 2050 2100 2050 2100 2050 2100 2050 2100 2050 2100 Antigua and 1,611 96 0.31 0.37 0.47 0.37 0.47 0.38 0.26 0.38 0.26 0.24 0.28 0.35 0.27 0.34 0.29 0.19 0.28 0.19 Barbuda Bahamas 12,425 386 0.40 0.45 0.56 0.50 0.72 0.49 0.75 0.55 0.97 0.42 0.50 0.81 0.50 0.81 0.67 0.96 0.67 0.96 Barbados 5,145 287 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 Belize 1,871 383 0.58 0.64 0.77 0.79 0.91 0.65 0.84 0.80 1.00 0.29 0.30 0.31 0.29 0.30 0.30 0.34 0.29 0.33 Dominica 551 72 0.88 0.98 1.19 1.13 1.33 1.01 1.38 1.16 1.54 1.49 2.08 2.56 2.18 2.67 2.10 2.88 2.20 3.00 Dominican 85,555 10,627 0.20 0.27 0.36 0.31 0.41 0.28 0.46 0.33 0.52 0.06 0.07 0.08 0.07 0.08 0.07 0.10 0.07 0.10 Republic Grenada 1,186 111 0.24 0.33 0.49 0.33 0.49 0.36 0.76 0.35 0.76 0.16 0.21 0.31 0.21 0.30 0.23 0.45 0.22 0.44 Guyana 3,879 779 3.62 6.50 11.65 7.64 13.70 7.08 17.20 8.33 20.21 1.65 2.37 3.41 2.30 3.31 2.56 4.83 2.49 4.69 Haiti 9,659 11,123 0.50 0.82 1.50 0.94 1.38 0.84 1.74 0.97 1.59 0.06 0.06 0.08 0.06 0.08 0.07 0.09 0.07 0.09 Jamaica 15,714 2,935 0.65 1.00 1.78 1.17 1.83 1.05 2.15 1.23 2.24 0.25 0.35 0.43 0.39 0.62 0.37 0.52 0.41 0.73 St. Kitts 1,011 52 0.16 0.19 0.24 0.19 0.24 0.20 0.30 0.20 0.30 0.31 0.38 0.54 0.42 0.71 0.43 0.64 0.47 0.85 and Nevis St. Lucia 1,922 182 0.38 0.42 0.56 0.52 0.74 0.44 0.67 0.54 0.88 0.14 0.15 0.18 0.15 0.18 0.16 0.22 0.16 0.22 St. Vincent and the 811 110 0.44 0.38 0.47 0.48 0.51 0.39 0.51 0.49 0.55 0.07 0.08 0.08 0.07 0.08 0.08 0.08 0.07 0.08 Grenadine s Saint Martin 562 37 1.38 1.88 2.63 1.96 2.73 2.02 3.50 2.11 3.64 1.40 1.85 2.45 1.86 2.64 1.98 2.95 2.00 3.17 (FR) Sint Maarten 1,059 41 0.38 0.46 0.61 0.47 0.61 0.48 0.80 0.48 0.80 0.46 0.50 0.71 0.49 0.69 0.56 0.90 0.54 0.87 (NL) Suriname 3,591 576 5.57 7.64 12.42 9.01 14.11 8.18 16.79 9.65 19.07 1.46 1.84 2.49 1.82 2.46 1.95 3.14 1.92 3.10 Trinidad and 23,808 1,390 0.05 0.06 0.07 0.08 0.09 0.06 0.08 0.09 0.11 0.03 0.03 0.04 0.03 0.04 0.03 0.05 0.03 0.05 Tobago Turks and 1,022 38 0.02 0.03 0.04 0.03 0.04 0.03 0.05 0.03 0.05 0.04 0.04 0.05 0.04 0.05 0.05 0.06 0.04 0.06 Caicos 64 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Table A.3 : Estimates of yearly erosion damages under RCP 4.5 and RCP 8.5 (Vousdoukas et al., 2018b) for each country using a nourishment cost of 10 USD/m3. Yearly erosion cost (% GDP) Baseline Country name GDP RCP 4.5 RCP 8.5 (M USD) Antigua and 1,611 0.17 0.27 Barbuda Bahamas 12,425 0.81 1.34 Barbados 5,145 0.01 0.02 Belize 1,871 0.52 0.83 Dominica 551 0.00 0.00 Dominican 85,555 0.01 0.01 Republic Grenada 1,186 0.04 0.07 Guyana 3,879 0.11 0.17 Haiti 9,659 0.06 0.09 Jamaica 15,714 0.02 0.03 Saint Kitts and 1,011 0.06 0.10 Nevis Saint Lucia 1,922 0.03 0.05 Saint Vincent and the 811 0.06 0.10 Grenadines Saint Martin 562 0.16 0.25 (FR) Sint Maarten 1,059 0.07 0.12 (NL) Suriname 3,591 0.01 0.02 Trinidad and 23,808 0.01 0.02 Tobago Turks and 1,022 0.72 1.17 Caicos Islands Total 171,382 0.09 0.14 65 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 Table A.4 : Estimates of yearly erosion damages under RCP 4.5 and RCP 8.5 (Vousdoukas et al., 2018b) for each country using a nourishment cost of 50 USD/m3. Yearly erosion cost (% GDP) Baseline Country name GDP RCP 4.5 RCP 8.5 (M USD) Antigua and 1,611 0.85 1.36 Barbuda Bahamas 12,425 4.06 6.70 Barbados 5,145 0.06 0.09 Belize 1,871 2.62 4.14 Dominica 551 0.00 0.00 Dominican 85,555 0.04 0.06 Republic Grenada 1,186 0.21 0.34 Guyana 3,879 0.53 0.83 Haiti 9,659 0.30 0.47 Jamaica 15,714 0.09 0.15 Saint Kitts and 1,011 0.31 0.49 Nevis Saint Lucia 1,922 0.17 0.27 Saint Vincent and the 811 0.30 0.48 Grenadines Saint Martin 562 0.79 1.23 (FR) Sint Maarten 1,059 0.37 0.58 (NL) Suriname 3,591 0.07 0.11 Trinidad and 23,808 0.07 0.10 Tobago Turks and 1,022 3.61 5.83 Caicos Islands Total 171,382 0.43 0.70 66 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020 67 of 67 Assessing the impact of sea level rise and resilience potential in the Caribbean 20-03-0028, 22 September 2020