A National Study on Protecting Infrastructure and Public Buildings against Sea Level Rise and Storm Surge

The national study analyzes sea level rise (SLR) impacts based on 36 different SLR and storm surge scenarios across 5.7 million geographic locations and 3 time periods. Taking an approach based on engineering design guidelines and current cost estimates, the study details projected cost impacts for states, counties, and cities. These impacts are presented from multiple perspectives including total cost, cost per-capita, and cost per-square mile. The purpose of the study is to identify specific locations where infrastructure is vulnerable to rising sea levels. The study finds that Sea Level Rise (SLR) and minimal storm surge is a $400 billion threat to the United States by 2040 that includes a need for at least 50,000 miles of protective barriers. The research is limited in its scope to protecting coastal infrastructure with sea walls. Additional methods exist and may be appropriate in individual situations. The study is original in that it is a national effort to identify infrastructure that is vulnerable as well as the cost associated with protecting this infrastructure.


21
Climate change presents a wide range of challenges for infrastructure owners, plan-22 ners, and users. Of these potential impacts, the one that is predicted to have one of the 23 largest ramifications in the United States is Sea Level Rise (SLR). If a rising temperature 24 continues to manifest on a global scale, then sea level rise will occur due to a combination of 25 thermal expansion of sea water and the melting of land-based ice into the ocean [1]. The con-26 sequences of this sea level rise on coastal road networks, buildings and infrastructure due 27 to economic, social and environmental costs are predicted to be substantial [2][3][4]. 28 According to NOAA, the United States' tidal shoreline is 95,471 miles long [5], of 29 which 60,342 miles is along the lower 48 states. Estimates for the number of people that 30 could be impacted by SLR vary, based on the expected depth associated with the SLR, but 31 range from a potential 4.2 million people at immediate risk of inundation given 0.9 m SLR 32 to 13.1 million people at immediate risk given a 1.8 m risk in the year 2100 [6]. Median 33 estimates for the level of SLR range from 30 centimeters to 111 centimeters depending on 34 the progress in mitigating greenhouse gas emissions [7,8]. 35 Because coastal areas have historically been attractive areas to establish communi-36 ties, the predicted SLR leaves many systems along the coast vulnerable to damage. A 37 total of 60,000 miles of US roads and bridges in the existing coastal floodplain are already 38 at risk to extreme storms and hurricanes [2]. When put in perspective of the overall US 39 economy, coastal zone counties account for 48% of GDP or $8 trillion dollars [9]. of what areas are at the greatest risk continues to dominate the discussion, the challenge 46 of how to respond to this vulnerability requires greater attention. Specifically, the ques-47 tion of whether SLR vulnerability requires new policies to relocate communities, or new 48 investments to construct protection barriers, or whether communities should take a wait- 49 and-see approach is one that needs further attention. 50 Currently, the approaches to SLR response can be divided into three broad catego-51 ries: protection, accommodation, and retreat. The protection category includes creating 52 dikes, surge barriers, closure dams, constructing dunes, nourishment and sediment man-53 agement of wetlands, creating coast defenses, sea walls and land claims, creating saltwater 54 intrusion barriers and implementing drainage systems/polders. The accommodation cat-55 egory includes implementing building codes to minimize the flooding of critical building 56 spaces, ensuring land use planning that accommodates for wetland loss, changing water 57 extraction practices, using freshwater injections to stop saltwater intrusions and increas-58 ing the delineation of natural hazard areas. Finally, practices that pertain to retreating 59 focus heavily on policies that minimize new building in areas where SLR threatens infra-60 structure, as well as considering the movement of existing structures in threatened areas 61 [1]. 62 While protection, accommodation, and retreat present a large array of approaches to 63 protecting against SLR, historically, the implementation of sea walls, also known as rock 64 revetments or armoring, has been the most common approach to reducing the impact of 65 ocean activity on coastal communities [14][15]. Similarly, the building of inland sea walls, 66 also known as bulkheads, along the banks of inland waterways have been a common ap-67 proach to protecting property against rising waterway levels. While these are the predom-68 inant approaches to protecting coastal properties, seawalls do not work in every circum-69 stance. Specifically, in cases where porous materials such as limestone form the bed of the 70 waterway, water can infiltrate through the rock and under the seawall. In these cases, 71 alternatives including the addition of pumping may be necessary. 72 Based on the historic focus on sea walls as a protection strategy, the current study 73 provides a national estimate of the construction costs associated with armoring areas of 74 the coast that are projected to be flooded and which contain built assets. These assets in-75 clude both public and private assets such as roads, rails, and public buildings. Private 76 residences are not specifically modelled in this effort but are included indirectly by pro-77 tecting the locations that include roads and other public infrastructure elements that sup-78 port these properties. 79 The intent of the current study is to provide the best estimate of expenses that have 80 the highest likelihood of being incurred over the next 5-10 years. The study utilizes in-81 undation projections from the lower bounds of those published to ensure that the overall 82 results provide an indication of hard costs that are likely to be incurred by local, regional, 83 and national entities. 84 The cost estimates presented here are considered conservative in that they are esti-85 mated construction costs that may increase due to specific conditions in local areas. The 86 costs also do not include long-term maintenance costs or the potential for cost increases 87 due to inflationary pressures. Thus, the actual costs incurred by municipalities is likely to 88 be higher than the costs presented in this study. The analysis of sea level rise impact is a growing field of study. A variety of ap-92 proaches have been used to model the potential economic, social and environmental costs 93 created by sea level rise. These procedures vary in methodology used, geography as-94 sessed, and scale implemented. 95 The first class of models consists of systems that are used to predict and model the 96 amount of land along the coastline that will be inundated which may include; inundation 97 models and Sea Level Affecting Marshes Model (SLAMM) [16]. Inundation models 98 utilize climate models to predict what areas will be flooded, using Geographic Infor-99 mation Systems (GIS) and Digital Elevation Models (DEM). SLAMM specializes in incor-100 porating habitat changes and processes that are likely to be impacted by climate change. 101 Kirshen et al. take a more simplified approach to modeling the impacts of sea level rise by 102 "developing damage-flooding depth probability exceedance curves for various scenarios 103 over a given planning period and determining the areas under the curve" [17]. 104 Neumann at el. incorporate "a tropical cyclone simulation model, a storm surge 105 model and a model for economic impact and adaptation" to estimate the impacts of sea 106 level rise for the US coastline through 2100. The model integrates site-specific elevation, 107 land subsidence and property value data to estimate the costs incurred due to shoreline 108 armoring, beach nourishment and property abandonment. Neumann et al. identify that 109 coastal areas are densely populated, further intensifying the impact of sea level rise on 110 human populations [18][19][20]. 111 In addition to measuring the financial impact of land inundation, ecological land-112 scape spatial simulation models account for the detrimental environmental consequences 113 of SLR. The Ecological landscape spatial simulation models are a category of model that 114 broadly analyze environmental factors such as subsidence, sea-level rise, changes in river 115 discharge, and climate variability and their impacts on coastal habitats. These models are 116 able to incorporate a larger range of variables including "coastal and estuarine hydrody-117 namics, water-borne particle transport, vegetation growth and infrastructure risk". How-118 ever, due to this level of detail, these models can be expensive and time-consuming to run. 119 The Dynamic Interactive Vulnerability Assessment (DIVA) model integrates biophysical 120 and socioeconomic consequences of sea-level rise and is able to assess the costs and ben-121 efits of adaptation to the predicted impacts. DIVA is designed to incorporate a large vari-122 ety of factors and components at a global scale and size. Similarly, SimCLIM models the 123 biophysical and socioeconomic impacts of climate change and variability. The tool esti-124 mates how future climate and sea-level changes impact sectors and associates a sensitivity 125 analysis with the modeling. The tool can be applied to local and global scales, depending 126 on the availability of data. 127 The final category of model used to predict and understand the impacts of SLR ac-128 count for the social impact on coastal communities and economies. Hsiang et al. utilize 129 SEAGLAS (Spatial Empirical Adaptive Global-to-Local Assessment System) to estimate 130 the cost of climate change to the sectors of agriculture, crime, coastal storms, energy, hu-131 man mortality and labor using a "risk-based approach" which is "grounded in empirical 132 longitudinal analyses of nonlinear, sector-specific impacts". The results suggest that cli-133 mate change costs approximately 1.2% of the gross domestic product per +1°C [21]. 134

135
Building on existing methodologies as well as new approaches developed for the 136 current study, the current study employs a multi-step process incorporating climate pro-137 jections, geoprocessing of detailed coastline flooding maps, the computational assessment 138 of where coastline needed protection, and the calculation of the costs associated with this 139 protection. The process developed for this estimation is based on previous climate impact 140 work developed by the authors for infrastructure impacts locally, regionally, and globally 141 [22-24]. The first step in the impact process was the identification of areas where inundation 145 and flooding were projected along the coastline and inland waterways. The study utilized 146 SLR and 1-year storm surge inundation projections from Climate Central through high-147 resolution data sets for the contingent United States coast based on published sea level 148 rise projections as well as research conducted by Climate Central [25][26][27]. The data sets 149 included analysis of the coastline at a 5-meter x 5-meter grid to ensure accurate capture of 150 tidal inlets. Each geographic location in the data set provided information on whether the 151 location was projected to be inundated and if so, to what depth the inundation was ex-152 pected. 153 As stated previously, the focus of the current study is to provide the best estimate of 154 expenses that have the highest likelihood of being incurred in the next 5-10 years. From 155 this perspective, the study includes the two lower projection RCP pathways, RCP 2.6 and 156 RCP 4.5, to provide an estimate of expenses that have the highest likelihood of being in-157 curred. The inundation data sets provided for the study were derived from a Monte Carlo 158 simulation set of 10,000 projections generated by Climate Central. Specifically, the 5 th , 50 th , 159 and 95 th percentile inundation projections for each of the two pathways from the overall 160 dataset were selected for the current study. Three time periods were selected from the 161 results for the impact analysis; 2040, 2060, and 2100. Additionally, the inundation data 162 was included with and without 1-year storm surge projections to capture both the base 163 SLR impact and the potential for regular flood impacts. These combinations resulted in a 164 total of 36 different scenarios for use in the study. 165 Similar to the projection data, Climate Central provided GIS files of infrastructure 166 locations based on previous work and public database information [28] ( Figure 1). This 167 infrastructure included a broad range of public infrastructure including schools, hospitals, 168 medical facilities, government buildings, airports and all public horizontal infrastructure 169 (roads, railways and runways). Although the study does not consider private residences 170 directly, the location of most residential areas can be determined through the location of 171 public roads that are used to access residential areas. Therefore, by considering all areas 172 that contain a road (both paved and unpaved), the majority of residential areas were also 173 considered. Areas that do not have any public infrastructure, such as national parks or 174 protected wildlife areas, were not included in the study. 175 A transformation of the data was performed to reduce the datasets to an indication 176 of whether infrastructure was in a specific area and whether that area was projected to be 177 impacted by SLR or storm surge. Although the original climate data was provided at an 178 ultra-high resolution, for processing speed, usability and accuracy, the data was con-179 densed to a uniform grid size. Sensitivity analysis tests were performed to determine an 180 appropriate grid size that would allow for the most accuracy in results while still main-181 taining computability speed. The sensitivity analysis focused on determining the largest 182 grid size that would both retain the underlying inundation detail as well as accurate loca-183 tion information for the infrastructure being analyzed. Through a series of test runs of 184 increasing grid sizes, the sensitivity analysis found that a grid system of 150 m 2 would 185 achieve the research objectives ( Figure 2).  192 Once the flooding files were processed, the second step of the process required de-193 termining what areas of coastline needed protection to remove the threat of flooding. This 194 determination requires a series of logic tests to understand if a flooded grid is directly 195 impacted by flooding from adjacent waterways, or if it is indirectly affected by other grids 196 that are adjacent to waterways. 197 The first step in this process was to determine if any given gridded square is located 198 within an area that is expected to flood, according to a specific climate scenario. This ques-199 tion is nuanced in that there must be a determination as to how much of a grid cell needs 200 to be flooded for it to be considered a flooded grid. The need for this determination orig-201 inates from the issue of how to limit the protection of coastal grids that appear in the study 202 with minimal flooding along the edge of the coastline. For example, a grid covering an 203 inlet which is indicated to have inundation over an area covering just a few yards onshore, 204 and does not include flooding of any infrastructure, can be eliminated in terms of needing 205 protection. The study adopted a threshold of 15% minimum inundation area to eliminate 206 overprotection scenarios. The value of 15% was chosen based on engineering judgement 207 upon inspection of protection patterns using 5%, 10%, 15% and 20%. 208 The second issue focused on whether a grid was flooded due to direct flooding or 209 indirect flooding. The model works from the assumption that wherever flooding occurs, 210 the shoreline directly impacting that flooded area needs to be protected. The case of direct 211 flooding occurs when a grid is adjacent to a waterway and the scenario indicates that grid 212 is flooded due to an overtopping of the adjacent waterway. In this case, the adjacent 213 shoreline needs to be protected to prevent the grids from incurring flooding. The indirect 214 case occurs when an inland grid is flooded due to being connected directly or indirectly 215 to a water-facing grid. In this case, the model must trace the path of the flood back to its 216 origin which is the grid adjacent to the coastline. The model then protects the coastline 217 adjacent to the grid to eliminate the threat to the overall flood area. 218 The identification of the flood areas provides the entry point for the final step in the 219 process of calculating the length of coastline to be protected. The current study utilizes the 220 NOAA Medium Resolution Shoreline Data in order to determine what is considered 221 shoreline. This data set does not include HI, AK, PR, Guam, or other US territories. How-222 ever, this data was selected to ensure that the results were consistent with the original 223 data provided by Climate Central which was based on the NOAA Medium Resolution 224 Shoreline Data. As illustrated in the map of Charleston, South Carolina in Figure 4, the 225 NOAA map provides detailed imaging of the inlets and tidal areas within that geographic 226 location. This is in contrast to lower resolution maps such as the coastline map from the 227 Census Bureau which simplify the coastline as indicated by the red line that eliminates 228 the inland portions of the coastline. The result of using this higher resolution map is that 229 the actual length of coastline increases, in some cases significantly, as the true length of 230 coastline can be calculated. In the current study this translates to a study length of approx-231 imately 135,000 miles in comparison to the NOAA measurement of approximately 95,000 232 miles. 233 After completing the protection length calculation, the model analyzed the coastline 234 for every grid that was determined to have a flooding impact on identified infrastructure. 235 For each of the identified grids, the length of coastline in that grid was calculated to a 236 linear foot (Figure 3).

239
Costing assessments for the study were created using a combination of national cost 240 databases and local estimates from seawall design and construction companies to estab-241 lish localized per-foot costs. The estimates were adjusted for location factor by state, based 242 on the location of the sea wall. 243 The cost estimates are divided into two categories, coastal seawalls and inland sea-244 walls. In terms of the former, coastal seawalls are comprised of armored revetments that 245 are either adjacent to shore structures or serve as standalone offshore structures. The cur-246 rent study utilizes a typical design approach of using field stone to create an armored 247 revetment on the shoreline. This design is utilized in the model wherever the coast expo-248 sure is direct to open water. 249 Inland seawalls, often referred to as bulkheads, focus on the protection of shoreline 250 from an increased water level as well as from indirect wave action. Bulkheads are gener-251 ally constructed of steel sheet piling, wood, or concrete where more permanent protection 252 is required. The primary cost factor in these solutions is the installation which may vary 253 depending on where the bulkhead is located.

256
The protection of the coastline across 23 states is a significant task that will require 257 the cooperation of national, regional, and local entities. As illustrated in Figure 8, a mid-258 level climate scenario in 2040 projects that every coastal-facing state is threatened by sea 259 level rise and storm surge at a national cost conservatively placed at $416 billion. This 260 exposure elevates the SLR issue from a local problem that places the burden on local offi-261 cials to a national issue that requires collaboration at all levels. In this section, the results 262 of the SLR study are presented at the national, state, and county levels to emphasize the 263 multi-jurisdictional impact of SLR. 264 As detailed in the methodology section, the current study analyzed 36 different cli-265 mate scenarios to determine the potential impacts of SLR and storm surge. The 50 th per-266 centile of the RCP4.5 scenario at 2040 is highlighted in the following sections. The results 267 reflect the SLR and 1-year storm surge projections.

270
The construction cost to protect coastal infrastructure from SLR and storm surge is 271 conservatively placed at $416 billion in 2019 dollars when considering a projection using 272 the 50 th percentile estimate of RCP 4.5. This estimate grows to $518 billion in 2019 dollars 273 when the same scenario is extended to 2100 (Figure 9). The number is slightly higher at 274 $530 billion when the 95 th percentile of RCP 4.5 is considered in 2100. This number does 275 not include maintenance costs, future replacement costs, or potential inflationary pres-276 sures due to a limitation of material or personnel resources. 277 Regionally, the western states (California, Oregon, and Washington) see a combined 278 impact of $53.5 billion by 2040 with California having the largest impact with a potential 279 impact of $22.0 billion. While this region is not discussed as often as Atlantic coast and 280 Gulf coast states, the impact should not be minimized. California and Washington are 281 ranked 6 th and 7 th respectively in terms of potential impacts from SLR and storm surge by 282 2040. Porter, 2018). However, this number grows significantly larger when the cost of protecting 289 assets beyond real estate to include infrastructure as a general category is considered. 290 The southern half of the Atlantic states (Virginia, North Carolina, South Carolina, 291 and Georgia) account for $101.2 billion of costs by 2040. Notable among this group are 292 North Carolina and Virginia which rank 2 nd and 3 rd in projected costs respectively. This 293 region is already incurring damage-related costs in the billions of dollars due to increased 294 nuisance flooding as well as from recent hurricanes. However, the projected impact from 295 chronic SLR issues must not be overlooked. The long-term cost of chronic impacts will 296 likely surpass those of one-time events. 297 Finally, the northern half of the Atlantic states (Maine, New Hampshire, New York, 298 Massachusetts, Rhode Island, Connecticut, Delaware, Pennsylvania, New Jersey, Wash-299 ington DC, and Maryland) account for the remaining $118.7 billion. This region is notable 300 for having states with considerable coastal infrastructure that requires protection such as 301 in Maryland, Delaware, and New Jersey. In these locations, as in many of the high-cost 302 states, infrastructure has been constructed over time in low-lying areas that extends to the 303 coast with little or no buffer areas to protect property and infrastructure.

306
As illustrated in Table 1, Florida, Louisiana, and Virginia are the top 3 states in terms 307 of projected protection costs. Each of these states have similar concerns in terms of SLR. 308 Specifically, the extensive low-lying areas that exist together with the extensive infrastruc-309 ture that has been built in the low-lying areas. While these vulnerabilities have been ex-310 posed in large events such as hurricanes, the potential risk from the topography is in-311 creased when the impact of SLR and storm surge is taken into account along the entire 312 length of low-lying areas.    Table 2 uses the same scenarios as that in Table 1, but places the information in terms 364 of length of miles of seawall required to protect the infrastructure within the given state. 365 As listed, it is projected that a minimum of 50,000 miles should be considered for protec-366 tion by 2040.

369
Cities are often the focus of SLR concerns and reports. Large cities including New 370 York City, San Francisco, and Miami are often associated with the potential costs and risks 371 of SLR. However, the risks of SLR go beyond these limited examples to cities of all sizes 372 located in coastal zones. The current study addresses this oversight by specifically organ-373 izing the output in the context of the coastal cities of all sizes. The results are presented 374 in multiple formats including total costs, per-capita and per-square mile contexts. These 375 multiple perspectives were created by translating the underlying grids from the study to 376  A combination of inland waterway and coastal exposure creates one scenario where 386 many of the top-listed cities such as Jacksonville, FL, are at risk. This scenario is similarly 387 found in the other top cities including New York City, Virginia Beach, and Charleston. 388 A second scenario exists in cities such as Marathon, FL, Fire Island, NY, and Galveston TX 389 where the cities occupy an island location. In these scenarios, the protection require-390 ments can extend throughout the island to protect infrastructure from SLR and storm 391 surge.  The Top 20 cities not only represent a diversity of geography, but also represent a 416 diversity in per-capita and per-square area. Table 4 presents the overall list of Top 20 cities 417 from a per-capita perspective. Many of these locations are either small towns located on 418 islands, or beach towns located on coastal or inland waterways.  The projected costs per population highlight the challenge of small cities in terms of 426 the cost of protection versus the size of the city. The same challenge exists in terms of the 427 size of the town in comparison to the threats facing the location. Of particular concern, are 428 the small islands and coastal areas that either line the coast or lie adjacent to the mainland 429 in areas along the Atlantic coast in particular. These areas often have small populations, 430 but are popular tourist destinations. Protecting these areas, in particular the islands, can 431 translate to significant costs when put in the context of the size of the area. 432 Table 5 illustrates a number of these smaller geographic areas with a minimum 1 433 square mile of land area and their costs on a per-square mile basis. As listed, these small 434 areas can exceed $100 million per square mile of area to protect the location. 435 As a comparison for the small area challenge, cities with a large cost, but also a large 436 area where this cost is distributed include; New York City at $6.57 million per square mile, 437 Miami at $4.82 million per square mile, and San Francisco at $5.47 million per square mile. 438 These results highlight a challenge for policy makers as to where to prioritize limited pro-439 tection funds. Should priority be given to popular tourist areas with small populations, 440 but a large risk factor, or should priority be given to urban areas that provide a larger 441 return-on-investment ratio in terms of protection costs to area and population.

444
Seal Level Rise and storm surge present a new risk and projected reality for coastal 445 communities. This study highlighted a middle-of-the-road projection in 2040 and 2100 to 446 emphasize a likely scenario of costs that cities and counties will face over the next 5-10 447 years. Figure 4 illustrates the 2040 and 2100 costs for each state under the conservative 448 RCP 4.5 scenario. As illustrated, the majority of states incur the primary protection costs 449 by 2040 and only a few see significant increases in 2100. 450 The following sections focus on three issues that emerge from the current study; the 451 response timeline, the urban versus rural challenge, and the protection feasibility chal-452 lenge.

456
A key message from the data developed in this study is that the timeline for respond-457 ing to the threat of sea level rise and storm surge begins now. The data for all scenarios 458 included within this study indicate that SLR and storm surge will have a national impact 459 by 2040 in the majority of locations. In addition to the projected cost factor, there is con-460 siderable effort required to plan and design for protective barriers. Issues such as envi-461 ronmental impacts, site-specific engineering solutions, and availability of expertise are is-462 sues that will extend the time required to implement protection solutions. 463 Given that additional cost and time will be required for almost all of the protection 464 projects, consideration must be given to initiating discussions on this issue if they are not 465 already started. The data in this study indicates that within 20 years, approximately 80% 466 of the protection needed by 2100 to protect infrastructure from the SLR risk will already 467 be required. In terms of the number of miles of protection required by 2040, there is a 468 projected need for over 50,000 miles of protection. This number only increases by 15% to 469 just over 61,000 by 2100. This increase reduces to 10% or less in states including Dela-470 ware, Louisiana, New Jersey, North Carolina, South Carolina, and Georgia. 471 The message from the data and the accompanying protection analysis is that the time-472 line for decision-making is sooner rather than later. The majority of impacts to infrastruc-473 ture will occur by 2040.

476
The second challenge arising from the current study is the issue of where to prioritize 477 protection from SLR and storm surge. As documented previously, depending on the per-478 spective chosen, the cities at greatest risk will change in terms of cost. Table 6 provides a 479 comparison of three cities with differing prioritization depending on the perspective. 480 Holden Beach, NC is significantly higher in cost when viewed from a perspective of per-481 square mile costs. This measure indicates that the limited area of Holden Beach will re-482 quire a significant investment to protect the infrastructure located in the town. In contrast, 483 Jacksonville has a much smaller investment when viewed on a per-square mile basis. This 484 would indicate that the relative cost of protecting the infrastructure in Jacksonville is less 485 than that facing the community of Holden Beach. When viewed from a perspective of 486 population, Dames Quarter, MD is a much more significant investment than the other two 487 communities. Jacksonville appears to be a much smaller investment when placed in the 488 context of a per-capita investment. Finally, from a total cost perspective, the perspective 489 changes to show that Jacksonville will require a much greater investment than the other 490 two communities. From this perspective, Holden Beach is the lowest cost investment at 491 only $432 million by 2040.  The final issue that is highlighted here from the current study is the issue of the fea-498 sibility of implementing the required protection by 2040. With 50,000 miles of sea wall to 499 construct by 2040, the issue arises as to the feasibility of constructing this volume of pro-500 tection in time. The issue of feasibility incorporates multiple issues including; availability 501 of design and construction personnel, availability of materials, and the potential from 502 price increases due to micro-inflationary pressures. 503 In terms of personnel availability, the question focuses on whether there will be suf-504 ficient numbers of design and construction personnel available to design and construct 505 over 2,000 miles of sea walls per year for the next 20 years. While there may be sufficient 506 numbers of personnel in locations such as Florida and Louisiana where coastal engineer-507 ing is a constant requirement, there may be issues in areas such as Washington where over 508 1,600 miles of sea walls is projected to be required. 509 Similar to the availability of personnel is the availability of construction materials. 510 While coastal revetments primarily require rock and concrete which is more readily avail-511 able, inland bulkhead seawalls require materials such as steel sheet piling. A large push 512 to construct these bulkhead seawalls will put pressure on material suppliers in terms of 513 how the prioritization will be made between seawall construction and the continuing re-514 quirements of materials for other projects. Delays in providing materials could stall the 515 required protection projects for extended periods of time. 516

517
The current study is intended to open a new conversation on the impact of sea level 518 rise and storm surge. It is not focused on the science behind the SLR risks, nor does it 519 intend to add to the scientific question of whether there will be SLR or how much SLR 520 there might be. Rather, the current study addresses the critical question of what is the 521 potential impact of the projected SLR and storm surge. In the same manner as public offi-522 cials plan for earthquakes, hurricanes, floods, and tornadoes, SLR is a natural hazard that 523 deserves a complete range of discussions including how to minimize the damage if and 524 when such an event occurs. At a conservative projection of a $400 billion impact by 2040, 525 SLR can no longer be ignored or treated as a purely theoretical argument by public offi-526 cials responsible for the health and safety of the general public. 527 In addition to encouraging public officials to include SLR in planning discussions, 528 this study should encourage communities of all sizes to consider the monetary commit-529 ment required for protection against SLR. Whether the community is limited in physical 530 area and population such as Dames Quarter, MD, or is one of the larger cities such as San 531 Francisco, the impact of SLR will have significant financial impact. 532 The decision to address SLR is only the first step in addressing this complex issue. A 533 single property owner, or even a single community, is not enough to address the overall 534 threat from SLR. While a single owner may choose to retreat from the coastal area, or a 535 community may elect to aggressively address SLR, this is an issue that requires coopera-536 tion and collaboration at the state, regional, and national levels. The successful imple-537 mentation of a protection system requires neighboring communities and states to work 538 together to ensure that engineered or natural systems work seamlessly along the coastline. 539 The issue of collaboration returns to the challenges addressed in the previous section. 540 Prioritization must be considered when implementing any protection plan. The question 541 of how to develop this prioritization is one with no easy solution. However, public offi-542 cials have a choice; focus on the differences between the communities (size, population, 543 total risk), or focus on possible solutions that can be mutually beneficial. The choice that 544 is made will set the stage for the future of many communities.