1 MODELING AND PLANNING FOR IMPACTS OF COASTAL FLOODING AND SEA LEVEL RISE ON CURRENT AND FUTURE DEVELOPMENT IN ST. JOHNS COUNTY, FLORIDA By ADAM HAMMOND CARR A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF URBAN AND REGIONAL PLANNING UNIVERSITY OF FLORIDA 2018
2 2018 A dam Hammond Carr
3 For my parents, David and Peggy
4 ACKNOWLEDGMENTS Many people have helped me over the course of my studies here at UF and, more broadly, over the course of my life. My parents have been extremely supportive and have offered important guidance to me throughout my life. From an early age they instilled in me a respect for the natural world and a desire to protect. This mindset has guided me through all of my ventures. I want to thank all of my colleagues at the GeoPlan Center, especially Crystal Goodison and L ex Thomas, who offered me a terrific environment to grow my GIS skills and knowledge. I hope to carry the character and quality of this work and learning environment with me in my future endeavors. I also want to thank all of my professors at UF, especia lly Dr. Paul Zwick and Dr. Katherine Frank, who helped me work through my research and studies. They have fostered in me a strong sense of how planners can have a positive impact on the world and the importance of thinking critically about planning proble ms. Finally, I want to recognize the friends I have made during my time at UF. Already I have seen them start their careers and work to improve the human environments that we live in. We have a wonderful opportunity to use our knowledge and education to work towards the greater good of the public and the planet. Despite variety of levels. I look forward to continuing to consider difficult problems, using data and sound a nalysis to make informed decisions.
5 TABLE OF CONTENTS p age ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURE S ................................ ................................ ................................ .......... 8 LIST OF AB B REVIATIONS ................................ ................................ ........................... 10 CHAPTER 1 I NTRODUCTION ................................ ................................ ................................ .... 13 S ea Level Rise and Coastal Flooding Threats in Florida ................................ ........ 13 F uture Growth in St. Johns County and Potential Impacts ................................ ..... 15 R esearch Outcomes ................................ ................................ ............................... 20 2 L ITERATURE REVIEW ................................ ................................ .......................... 22 S ea Level Rise Forecasting ................................ ................................ ................... 22 M odeling Coastal Flooding ................................ ................................ ..................... 25 E stimating Impacts of Coastal Flooding Events ................................ ..................... 31 3 M ETHODOLOGY ................................ ................................ ................................ ... 36 D ata Acquisition ................................ ................................ ................................ ..... 38 H azus MH Model Descriptions ................................ ................................ ............... 43 C urrent Development Economic Impact Analysis ................................ ................... 45 F uture Development Economic Impact Analysis ................................ .................... 47 4 C OASTAL FLOODING AND SEA LEVEL RISE RESUL TS ................................ ... 49 5 E CONOMIC IMPACT RESUL TS ................................ ................................ ............ 54 C urrent Development Impact Results ................................ ................................ ..... 54 F uture Development Impact Results ................................ ................................ ...... 55 6 D ISCUSSION AND CONCLU SION ................................ ................................ ........ 58 C oastal Flooding from Base 100 Year Storm and Impacts ................................ .... 58 C oastal Flooding from 100 Year Storms with Extra Sea Level Rise and Impacts .. 62 C oastal Flooding Impacts on Future Development ................................ ................. 66 A dditional Considerations ................................ ................................ ....................... 70
6 L imitations ................................ ................................ ................................ .............. 74 S uggestions for Future Research ................................ ................................ ........... 75 C oncluding Thoughts ................................ ................................ ............................. 77 APPENDIX A HAZUS MH COASTAL FLO ODING MODEL WORKFLOW ................................ .. 79 B ECONOMIC IMPACT ANAL YSIS WORKFLOW ................................ .................... 82 C COASTAL FLOODING MOD EL OUTPUT MAPS ................................ .................. 91 REFERE N CES ................................ ................................ ................................ ............ 107 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 111
7 LIST OF TABLES Table page 2 1 General Descriptions of Coastal Flooding Models 2 6 2 2 National Averages for Full Replacement Cost Models of Various Building Occupancies 4 4 1 Flood Statistics for Coastal Flooding Scenarios Incorporating Varying Amounts of Sea Level Rise in St. Johns County ... 9 5 1 Current Economic Impact Results ... 5 5 5 2 Future Economic Impact Results 5 7 5 3
8 LIST OF FIGURES Figure page 1 1 Florida 2070: Currently Developed Areas in St. Johns County 2010 Base Scenario 7 1 2 Florida 2070: Currently Developed Areas with Future Development in St. Johns County 8 1 3 Florida 2070: Currently Developed Areas with Future Development in St. Johns County 2070 Alternative Scenario 9 2 1 Mean Sea Level Trend Gauge Station 8720218 Maypo rt, Florid a 3 3 1 7 3 2 General Hazus 4 3 3 General Parcel Level 4 6 4 1 Hazus Coastal Flooding Output Base Flooding Scenario in the St. Augustine Area ... 5 1 4 2 Hazus Coastal Flooding Output 2070 Low Flooding Scenario in the St. Augustine Area ....... 5 2 4 3 Hazus Coastal Flooding Output 2070 High Flooding Scenario in the St. Augustine Ar ea ... 5 3 6 1 Hazus Coastal Flooding Output Base Flooding Scenario for St. J ohns County 60 6 2 Hazus Coastal Flooding Output Base Flooding Scenario in the St. Augustine Area ... 6 1 6 3 Hazus Coastal Flooding Output 2070 VLM Flooding Scenario for St. J 4 6 4 Hazus Coastal Flooding Output 2070 Extreme Flooding Scenario for St. Johns County .. 5 6 5 Hazus Coastal Flooding Output 2070 VLM Flooding Scenario, Impact on Future Development Near St. Augustine 8 6 6 Hazus Coastal Flooding Output 2070 Extreme Flooding Scenario, Impact on Future Development . 6 9
9 6 7 Hazus Coastal Flooding Output Base Flooding Scenario, Flooding Issue Along A1A Beach Boulevard Near Anastasia State Park .. 7 2 6 8 Hazus Coastal Flooding Output 2070 Low Flooding Scenario, Flooding Issue Along A1A Beach Boulevard Near Anastasia State Park . 7 3 C 1 Hazus Coastal Flooding Output Base Scenario Overview . 91 C 2 Hazus Coastal Flooding Output 2070 VLM Scenario Overview .. C 3 Hazus Coastal Flooding Output 2070 Low Scenario Overview .. C 4 Hazus Coastal Flooding Output 2070 IntLow Scenario Overview C 5 Hazus Coastal Flooding Output 2070 Intermediate Scenario Overview C 6 Hazus Coastal Flooding Output 2070 IntHighScenario Overview C 7 Hazus Coastal Flooding Output 2070 High Scenario Overview C 8 Hazus Coastal Flooding Output 2070 Extreme Scenario Overview C 9 Hazus Coastal Flooding Output Base Scenario Zoomed In C 10 Hazus Coastal Flooding Output 2070 VLM Scenario Zoomed In ... C 11 Hazus Coastal Flooding Output 2070 Low Scenario Zoomed In ........ ....101 C 12 Hazus Coastal Flooding Output 2070 IntLow Scenario Zoomed In C 13 Hazus Coastal Flooding Output 2070 Intermediate Scenario Zoomed In .. 103 C 14 Hazus Coastal Flooding Output 2070 IntHigh Scenario Zoomed In ...... .104 C 15 Hazus Coastal Flooding Output 2070 High Scenario Zoomed In ... .105 C 16 Hazus Coastal Flooding Output 2070 Extreme Scenario Zoomed In ..... 106
10 LIST OF AB B REVIATIONS ADCIRC Advanced Circulation BEBR Bureau of Economic and Business Research DEM Digital Elevation Model FEMA Federal Emergency Management Agency FGDL Florida Geographic Data Library FIS Flood Insurance Study GIS Geographic Information System GMSL Global Mean Sea Level Rise IPCC Intergovernmental Panel on Climate Change MEOW Maximum Envelope of Water MOM Maximum of Maximums NOAA Natio nal Oceanic and Atmospheric Administration RSL Regional Sea Level SLOSH Sea, Lake, and Overland Surges from Hurricanes SLR Sea Level Rise SWAN Simulating Waves Nearshore SWEL Stillwater Elevation USACE United States Army Corps of Engineers VLM Vertical Land Motion WHAFIS Wave Height Analysis for Flood Insurance Studies
11 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Urban and Regional Planning MODELING AND PLANNING FOR IMPACTS OF COASTAL FLOODING AND SEA LEVEL RISE ON CURRENT AND FUTURE DEVELOPMENT IN ST. JOHNS COUNTY, FLORIDA By Adam Hammond Carr May 2018 Chair: Paul Zwick Cochair: Katherine Frank Major: Urban and Regional Planning Coastal flooding and sea level rise pose major threats to co astal areas around Florida, the US, and around the world. Currently developed areas are at risk, as well as areas likely to be developed in the f uture. This research focu ses on St. Johns County, Florida in the northeastern part of the state. The goal is to estimate the physical and economic vulnerability currently developed areas in St. Johns County face from coastal flooding events. Along with current development, areas projected to be developed by 2070, according the Florida 2070 study, are also included in the analysis. Various sea level rise scenarios are included with the coastal flooding analysis to determine how variations in sea level will impact these areas in t he event of a 100 year hurricane event. The Hazus MH software is used to model the coastal flooding scenarios and generates flood depth grids as well as impacts to buildings. Economic impacts to buildings are also estimated using a parcel level analysis approach. The outcomes of the analysis show that between 40,000 and 5 8,000 acres of land may be inundated in the event of a 100 year storm, depending on the amount of sea level rise included in the
12 scenario. Currently developed areas are at risk of betwe en $850 million and $1.1 billion in damage to buildings (2018 dollars). Currently developed areas and additional areas projected to be developed by 2070 face the risk of between $4.2 and $16.6 billion of damage to buildings (2070 dollars). Using this inf ormation, communities in St. Johns County can better understand their vulnerabilities and work to minimize risks in the future. Limiting development in risky areas and instituting smart building regulations to mitigate vulnerability will help communities minimize economic impacts in the event of coastal flooding events in the future.
13 CHAPTER 1 INTRODUCTION Sea Level Rise and Coastal Flooding Threats in Florida C oastal f looding and sea level rise pose major threats to seaside communities around the world. Areas with developed coastlines have increased vulnerabilities to infrastructure, as well as human life, which is at risk of damage or loss due to these hazards. Florida face s particular vulnerability because large portions of the coastline have been developed coastal areas are at or very near sea level, and the peninsula is exposed to powerful tropical cyclones. Coastal areas are major contributors to Fl real estate and tourism markets, which form big pieces of the Fl orida Ocean Alliance, 2013; Bureau of Economic Analysis, 2017; Oxford Economics, 2018; Trigaux, 2016 ). In addition, many highly populated cities and areas of cultural significance can be found on or near the coasts with eight of Florid populous cities near the coast (Office of Economic and Demographic Research 2017 ). only likely to grow in the future. Over the next 50 years, it is estimated that about ten million people will move to Florida, putting the total population around 30 million (Florida Bureau of Economic and Business Research 2015 ; Carr& Zwick, 2016) A portion of these people will likely move to coastal areas, putting even more people and new infrastructure need ed to support them, at risk of storm surge events and sea level rise And it is predicted that changes to the global climate will lead to stronger and more frequent hurricanes, as well as sea level rise. This will put coastal communities at a greater ris k of damage and loss es. Impacts from coastal flooding, also known as storm surge and sea level rise have been researched and modeled in many communities around Florida (Frazier,
14 Wood, & Yarnal, 2010; Jones & Griffis, 2013: Linhoss et al., 2015; Peng, 2015 ) However, with storm surge and sea level rise research the specific location studied is extremely important. Small changes in topography and bathymetry can have large impacts on how an area will be impacted by storm surge and changing sea levels (Zimm erer et al., 2007; FEMA, 2018) Furthermore studies on the effects of storm surge and sea level rise have focused on threats to current infrastructure and human development. However, storm surge coupled with future sea level rise may impact areas that a re currently undeveloped but are likely to be built up projected population growth. With these considerations in mind, was selected as the study area. This area has been studied before in the context of coastal vulnerability but using different methods and with different research goals ( Frank, Volk, & Jourdan, 2015; Linhoss, Kiker, Shirley & Frank, 2015 ). St. Johns County is a significant study area because it is vulnerable to storm s urge and sea level rise and is home to special cultural assets In addition, St. Johns County is predicted to see significant development over the next 50 years as more people immigrate to Florida (Carr & Zwick, 2016) This research seeks to answer three primary questions: W hat is the vulnerability to a coastal flooding event in St Johns County and what are the estimated economic impacts on development within the study area? More specifically, what will be the economic damage to the building stock in the county? Answering this will offer new insight into how much damage could be expected in the event of a severe storm hitting the area.
15 Next, h ow will vulnerabilities change given different sea level rise scenarios? With updated data from the nearby tidal gauge and new sea level rise projections produced by NOAA, a more realistic view of future coastal conditions can be modeled on top of coastal flooding. Finally, h ow would future development be impacted by a coastal flooding event under these sea level rise scenarios ? This question will help estimate not only the physical risks that future development will face, but the economic risks too Using the current land use composition, land predicted to be built on in the future will be divided up proportiona lly into these land uses. With these land uses determined, the economic impacts from flooding storm surge and sea level rise can be estimated. Future Growth in St. Johns County and Potential Impacts The assumptions surrounding future development conditio ns in St Johns County used in this research is based off a prior research project called Florida 2070 In Florida 2070 Center, and the Florida Department of Agriculture and Con sumer Services used GIS analysis and scenario planning to consider how projected population growth in the year 2070 may be accommodated across the state. The study was tasked with allocating an additional 15 million people to different areas of the state. Two population allocation scenarios were created, one simulating a situation similar to the current pattern of land development, the Trend Scenario, and one simulating an alternative pattern w ith denser urban development and greater conservation of greenfield space, the Alternative Scenario. Figures 1 1 1 2 and 1 3 presented below, show the various developed area layers produced by Florida 2070 in St. Johns County.
16 As mentioned previously, St. Johns County was selected as the study area because of its current vulnerability to storm surge. However, it was also selected because St. Johns is projected to see some of the largest population growth, by percentage, of any county in the state. It is predicted that St. Johns will grow from a 2010 baseline of 190,039 to 591,272 in 2070, over a 3 fold increase (Carr & Zwick, 2016). Though some of the modeled development to accommodate population increases in St. Johns will take place away from the co ast, there are areas near the coast which will be vulnerable. The robust results of the Florida 2070 study combined with the coastal flooding and sea level rise models from Hazus enable novel research to take place. Analyzing the impact of storm surge co upled with future sea level rise on not only currently developed areas, but on areas likely to be developed in the future will give a greater understanding of the potential impacts of storm surge and sea level rise on the built environment in St. Johns Cou nty.
17 Figure 1 1 Florida 2070: Currently Developed Areas in St. Johns County 2010 Base Scenario
18 Figure 1 2 Florida 2070: Currently Developed Areas with Future Development in St. Johns County 2070 Trend Scenario
19 Figure 1 3 Florida 2070: Currently Developed Areas with Future Development in St. Johns County 2070 Alternative Scenario
20 Research Outcomes The outcome s of this research show t hat there is significant vulnerability to coastal flooding events in St. Johns County. A 100 year stor m with no additional sea level rise could flood an estimated 40,000 acres of land and cause between $850 million and $1.1 billion in damage to buildings With just over two feet of sea level rise, a 100 year storm could inundate nearly 50,000 acres of land And with almost five and a half feet of sea level rise, a 100 year storm could impact nearly 58,000 acres of land. These impacts from flooding with sea level rise translate to between about $1.0 billion and $4.5 billion of damage to buildings in St. Jo hns County, depending on the flooding scenario. Future development in St. Johns County is also expected to experience major impacts from coastal flooding events. An additional 180,000 acres of land is projected to be developed in St. Johns County. Of thi s total area, between 4,000 and 7,200 acres are predicted to be inundated by 100 year coastal flooding events, depending on the sea level rise scenario. Though this is a small amount compared to the total area of future developed, the economic impacts wou ld be significant. The estimated impact to buildings in future development is between $1.6 and $4.9 billion in 2070 dollars. Results will be discussed in greater detail in later chapters. In chapter two, the foundation for the research conducted in this paper will be laid out. The theoretical basis behind sea level rise forecasting, storm surge modeling, and estimating impacts to the built environment will be discussed. Chapter three will elaborate on the methodologies used in the study. This will cove r data acquisition, data management, the modeling workflow, and impact analysis workflow. Chapter four will present results of the coastal flooding and sea level rise analysis. T he modeled extent
21 of coastal flooding with and without sea level rise scenari os will be shown. Chapter five will present the results of the economic impact analysis. E stimated impacts to current and future development will be presented. Chapter six will provide a discussion and conclusion. The results and their implications for St. Johns County will be discussed. Conclusions regarding planning for future coastal flooding events and sea level rise, along with research limitations will close the paper.
22 CHAPTER 2 LITERATURE REVIEW The basis for this research is built on three broad areas of prior research: sea level rise forecasting, modeling of coastal flooding, and estimating impacts of f looding events Sea Level Rise Forecasting A variety of documents and tools are available which offer forecasts of how fu ture sea level rise may impact the globe as well as more specific regions. The Intergovernmental Panel on Climate Change (IPCC) is a well known international organization which has provided a variety of reports on climate change and its outcomes. The mos t recent major document produced by the IPCC is the Fifth Annual Report (AR5). In this document, among other climate change related issues, the IPCC describes various sea level rise concepts such as Global Mean Sea Level, Relative Sea Level Rise, Regional Sea Level, and Local Sea Level. These different concepts describe why the same amount of sea level rise is not seen everywhere and how in some locations negative sea level rise may be observed. In addition to these concepts, mu ltiple sea leve l rise scen arios are discussed. When considering global mean sea level rise, the IPCC predicts that between 2046 and 2065 sea levels rise will likely be 0.17 to 0.38 meters (0.56 to 1.25 feet) (IPCC, 2014). Between 2081 and 2100 the IPCC predicts a global mean sea level rise of between 0.26 and 0.82 meters (0.85 to 2.69 feet) will be observed (IPCC, 2014). With these broad ranges of global sea level rise in mind, other sources for regional or local sea level rise will be discussed. The National Oceanic and Atmosph eric A dministration (NOAA) produce data and reports describing sea level rise around the US. NOAA uses a network of tide gauge
23 stations to measure past sea level change trends and project future variation. There are four primary : Station 8720030, Fernandina Beach; Station 8720218, Mayport (Bar Pilots Dock); Station 8721120, Daytona Beach Shores; and Station 8723170, Miami Beach. The Mayport station is closest of the four Atlantic stati ons to St. Johns County, so is used as the source for sea level rise data. Sea level data for the Mayport station has been recorded monthly since ( National Oceanic and Atmosphe ric Administra tion, 2017). The trend data for this station can be seen in Figure 2 1 below Figure 2 1 Mean Sea Level Trend Gauge Station 8720218 Mayport, Florida (National Oceanic and Atmospheric Association, 2017) In Januar y 2017, NOAA released a report titled Global and Regional Sea Level Rise Scenarios for the United States This report presents updated projections of global mean sea level rise as well as regional sea level rise projections for areas in the US. Three re ports form the foundation for the research conducted in the Sweet et al. report: Global Sea Level Rise Scenarios for the United States National Climate Assessment Probabilistic 21st and 22nd
24 century sea level projections at a glob al network of tide gauge sites Regional Sea Level Scenarios for Coastal Risk Management: Managing the Uncertainty of Future Sea Level Change and Extreme Water Levels for Department of Defense Coastal Sites Worldwide (Hall et al. 2016). The work of Parris et al. set a precedent for sea level rise research in the US. It worked to coordinate sea level rise scenario planning and resulted in creating four Global Mean Sea L evel (GM S L) rise scenarios to help planners and policy maker s consider various risk levels in the future (2012). However, the work Parris et al. focused on global mean sea levels. Research by Kopp et al. (2014) and Hall et al. (2016) sought to generate more localized Regional Sea Level (RSL) rise estimates, accou nting for local variations such as land subsidence and variability in ocean ci rculation. The researchers used the Parris et al. (2012) results as a basis for their work. Kopp et al. generated probabilistic sea level projections at a global network of tide gauge sites (2014). Hall et al. uses the scenarios from Parris et (2016, pg. 2 9). Sweet et al. use these two reports as the foundation for developing RSL ris e scenarios for the U S Particularly important in the report is the incorporation the most up to date science and methodologies for adjusting GMSL rise scenarios to a specific region (Sweet et al., 201 7). The report produces six scenarios for 2100 based on GMSL scenarios. The GMSL rise scenarios are: Low (0.3 meters), Intermediate Low (0.5 meters), Intermediate (1 .0 meters), Intermediate High (1.5 meters), High (2.0 me ters), and Extreme (2.5 meters ) (Sweet et al., 2017). In addition, GMSL scenarios
25 were calculated for each de cade based on 19 year averages (Sweet et al., 2017). These GMSL scenarios were used to calculate RSL scenarios and can be viewed using a tide gau ge stati on as a reference point. A simple way to view the RSL rise projections at each station is by using the US Army Corps of Engineers (USACE) Sea Level Curve Calculator. Using the online tool provided by the USACE, a user can select from multiple sea level rise scenario sources. The NOAA 2017 report by Sweet et al. is included as an o ption. Simply by selected the Mayport gauge station and the NOAA 2017 scenario source the user can view the different GMSL projections at the Mayport location (USACE Sea Level Change Curve Calculator, 2017) Modeling Coastal Flooding Coastal flooding commonly called storm surge, poses major risks for communities in coastal areas Many techniques have been developed to model the extent and depth of flooding given a variety of parameters Typically, these models are produced by government bodies, like NOAA or FEMA, or by academic institutions. Examples include SLOSH, ADCIRC, SWAN, WHAFIS, and RUNUP The calculations and computer programs used to run these models are complex. The required inputs for these models varies based on the fluid mechanics and three dimensional models they use. Table 2 1 below provides an overview of the models reviewed and then they are described in greater detail.
26 Table 2 1 General Descriptions of Coastal Flooding Models Model Name Characteristics SLOSH Developed by National Weather Service. Uses grids with defined values. Can generate surge models using deterministic, probabilistic, or composite approach. ADCIRC Developed by research group. Uses various water and environmental characteristics to model s torm surge. SWAN Developed at Delft University of Technology. Models waves nearshore and accounts for various physical phenomena in wave generation. Should be paired with another model like SLOSH or ADCIRC. WHAFIS Developed by FEMA Uses transects to ca lculate wave characteristics in study area. User helps determine location of transects Frequently coupled with RUNUP. RUNUP Developed by FEMA. Uses environmental characteristics to determine height of water above still water elevation. Frequently coupled with WHAFIS. Hazus MH Developed by FEMA. Uses combination of updated/modified WHAFIS and RUNUP models. Offers many hazard analysis options. Relatively easy to use interface. Less technical knowledge needed. SLOSH Model SLOS H or S ea L ake and O verland S urge from H urricanes is a model developed by from historical, hypothetical, or predicted hurricanes by taking into account the atmospheric pressure, size, forward speed, and track data (N ational Hurricane Center 2018). There are three modeling approaches that SLOSH can use to calculate the extent and magnitude of a storm surge event: Deterministic Approach, Probabilistic Approach, and Composite Approach. The Deterministic Approach forecasts surge by solving physics equations using hurricane data as the inputs. This method is highly dependent on the quality of the input data and small changes in any hurricane characteristics can have a large impact on the model outputs, so this approach may not always produce accurate storm surge
27 models. The Probabilistic Approach uses statistics from prior forecast performances to create a group of SLOSH runs based on distributions of hurricane characteristics. The Compo site Approach generates thousands of SLOSH runs of hypothetical hurricanes with different variables. This approach generates the most robust outputs, providing worst case storm surge scenarios for hurricanes of varying intensities ( National Hurricane Cent er, 2018 ). SLOSH is a computationally efficient model and accounts for a variety of factors such as flow through barriers, overtopping of barrier systems, and coastal reflections of surges. However, SLOSH does not model impacts of waves nor can it account for rain flooding or river flow ( National Hurricane Center, 2018 ). Despite these drawbacks, SLOSH is an efficient and robust model that emergency management officials and planners can use to determine storm surge vulnerability. ADCIRC Model ADCIRC, AD van ced CIRC ulation, is a model developed by a group of researchers headed by Dr. Rick Luettich from the University of North Carolina at Chapel moving fluid on a rotating earth These equations have been formulated using the traditional hydrostatic pressure and Boussinesq approximations and have been discretized in space using the finite element (FE) method and in time using the finite (Luettich & Wester ink, 2017). ADCIRC can be run as a two dimensional depth integrated model or as a three dimensional model and can be run using a Cartesian or spherical coordinate system.
28 The ADCIRC model includes in its conditions factors like specified elevation, zero normal flow, external barrier overflow out of the domain, atmospheric pressure and outward radiation of waves. ADCIRC can also be forced with elements such as elevation boundary conditions, surface stress boundary conditions, and tidal potential (Luettich & Westerink, 2017). ADCIRC is a technical product that requires strong knowledge of fluid dynamics and physical modeling. SWAN Model SWAN, S imulating Wa ves N earshore, is a wave model developed at the Delft om, short crested wind generated Delft University of Technology 201 7 ). SWAN model runs can be made on a regular grid, a curvilinear grid, and triangular mesh in Cartesian or spherical coordinate system. SWAN c an be run in serial or in parallel ( Delft University of Technology, 2017 ). SWAN accounts for various physical phenomena, such as wave generation by wind, wav induced set up, propagation from laboratory up to global scales, diffraction, and three and four wave interactions. However, SWAN does not account for Bragg scattering or wave tunneling. SWAN produces numerical files containing tables, maps, and timeseries with outputs like significant wave height and period, average wave direction and directional spreading, dissipation, wave induced force, set up, and diffraction parameter ( Delft University of Technology, 2017 ). On its own, the SWAN model may not be extremely useful for modeling coastal flooding because of its focus on wave characteristics. Howev er, it can be paired with ADCIRC to provide very robust storm surge models. The combi ned SWAN and
29 ADCIRC models have been used in a variety of research efforts (Chen et al., 2013; Jones and Griffis, 2013 ; Sebastian et al., 2014; Xie, Zou & Cannon, 2016 ). WHAFIS Model WHAFIS, W ave H eight A nalysis for F lood I nsurance S tudies, is a model based program that uses representative Management A gency, 2018). Transects to be used in the model are selected with major topographic, vegetative, and cultural features in mind. Variations in these categories determine the frequency and placement of transects (Federal Emergency Management Agency, 2018). WHAFIS uses transects and other input data to calculate depth limited wave height on the ocean end of each transect. The user determines the location of the transects but should be careful when selecting their location to produce the most ideal distrib ution along the coastline being analyzed (Zimmerer et al., 2007) RUNUP Model based program for wave runup computation. This program uses stillwater elevation, shore profile and roughness and incident wave condition input information to compute a wave runup 2018). Wave runup is an important factor in considering the impact of storm surge because it determines how f ar above still water elevation a wave can be expected to reach. The RUNUP model is typically used in conjunction with other models, like WHAFIS.
30 Hazus MH The storm surge models listed above are fairly complex. They can require understanding of complex f luid dynamic models, with many required inputs to successfully run the program. This can make using these tools too cumbersome for planners complex physics. To offer another option, FEMA created the Hazus MH tool a program connected to the ArcGIS framework Hazus allows users to view impacts from a variety of natural hazards, including coastal flooding events. Hazus utilizes a nationally applicable standard methodology, offering fl exible options for generating outputs (FEMA, 2018). The coastal flooding modeling functionality in Hazus uses a general approach and methodology similar to those that FEMA uses to generate coastal Flood Insurance Rate Maps. These methods include drawing transects perpendicular to shorelines; calculating water surface elevations, flood depths, and flood hazard zones; and determining which models to run along each transect based on shoreline characteristics and wave conditions (Department of Homeland Securi ty, 2013 ) This ma y sound similar to the WHAFIS and RUNUP models. In fact, these models are used within Hazus to generate flood boundaries. However, it is noted in documentation that these models included in the Hazus tool contain simplifications compared to the full blow n WHAFIS and RUNUP models. These simplifications enable users to model storm surge events with less input and knowledge than required in the standard models. It is also noted in the documentation, though, that the models used in
31 the Hazus tool include im provements made to some aspects of the models by including more recent scientific developments (Department of Homeland Security 2013) Estimating Impacts of Coastal Flooding Events There is extensive literature focused on estimating the economic impacts of coastal flooding and sea level rise on the built environment. This research can be broken down into two varieties, one group considering only the direct impacts of a hazard on property and the other group considering additional indirect costs such as l ost income, relocation costs, and price changes Additionally, t hese studies typically consider on ly the impacts of storm surge or sea level rise (Yohe et al., 1996; Tol, 2002; Stanton, Davis, & Fencl, 2010; Linhoss et al., 2015 ) However, more and more researchers have recognized that future sea level rise will only make coastal flooding events more damaging by raising the base sea level ( Kleinosky, Yarnal, & Fisher, 2006; Frazier, Wood, & Yarnal, 2010 ; Peng, 2015; Withey, Lantz, & Ochuodho 2015 ) Geno vese and Green analyze how flood depth, using outputs from the SLOSH model, in South Florida may impact buildings, both property and contents in the absence and presence of coastal protection (2014). Using flood depth, flood extent, damage functions for different building classifications, and land use information for properties they are able to estimate the economic impacts when SLOSH outputs overlap with properties (Genovese & Green, 2014) Hazus MH Impact Analysis After determining flood depth and extent in Hazus, described in the section above, the software offers built in flood loss estimation analysis Hazus can implement two different methodologies specific to building damage, a full replacement cost or a
32 depre ciated cost model (Department of Homeland Security, 2013). The methodology for the full replacement cost is more significant as it relates to this research, so will be the focus of this review. Building replacement cost models in Hazus are built on indus try standard models available in the 2006 Means Square Foot Costs by R.S. Means. Replacement cost data is stored at the census block level for each type of building occupancy. The census block represents the highest level of economic impact analysis with in the Hazus framework. A default structure replacement cost, using cost for square foot as the measurement unit, is provided in the software for each occupancy class. These replacement costs can be viewed in Table 2 2 below. Square foot costs shown in the table were averaged for various building materials (Department of Homeland Security, 2013) Of the occupancy classes, the single family residential replacement cost model is the most intricate. It uses socio economic data from each census block to he lp provide a more accurate mix of replacement cost models. The algorithm used to determine total estimated valuation for single family residences includes the following factors: total area in square feet of single family residences in a census block, the Means construction class, the weighting factor for the construction class, the number of stories, the weighting factor for number of stories, the cost per square foot of single family residential given the construction class and number of stories, the pres ence of a basement, the weighting factor for basement presence, the additional cost per square foot for a finished basement given the construction class and number of stories, the weighting factor for garage type, additional replacement cost for garage typ e, and count
33 of single family residential structures within a census block (Department of Homeland Security, 2013). T he remaining algorithms used to estimate replacement cost for the different occupancy classes are far less complex. First, a specific alg orithm for replacement cost of manufactured housing, more specifically mobile homes, is used. The factors used to determine total valuation for manufactured housing includes: total floor area in square feet of manufactured housing and the cost per square foot. The other algorithm used includes all other residential building types and all non residential building types. This algorithm incorporates potential for number of stories. The factors used to determine the other building valuations include: total floor area in square feet for a specific building type in a census block and the cost per square foot of the building occupancy type (Department of Homeland Security, 2013).
34 Table 2 2 National Averages for Full Replacement Cost Models of Various Building Occupancies ( R.S. Means, 2006)
35 Hazus Impact Analysis Comparison The Hazus methodology is robust and includes region specific information based on the location of the study area. However, with the highest resolution of data being at the census block scale, some even higher level local variations can be smoothed over in the analysis. Karamouz et al. utilize an alternative method to the Hazus impact analysis methodology (2016) In the study they compare Hazus outputs with their alternative methods. They use four general inputs to model impacts: the floodplain, the DEM, land use data, and depth damage functions They use zonal statistics to assign the maximum water level to each polygon of land use, classified into three major categories: residential, industrial, and commercial. Afte r gathering this necessary informatio n by land use category, depth damage curves were applied based on land use to determine the economic impact of flooding (Karamouz et al., 2016). Karamouz et al. conducted a case study on the island of Manhattan in New York City, NY (2016). They ran the Ha zus impact analysis methodology then their own methodology to compare estimations of impacts to buildings. Overall, their models estimated lower economic loss estimates for structural impacts compared to the Hazus predictions when using Flood Insurance St udy information. In their conclusions, they discuss how the incorporation of higher resolution tax parcel data into the land use and damage modeling can lead to more accurate estimations of building impacts from flooding (Karamouz et al., 2016)
36 CHAPTE R 3 METHODOLOGY The goal of this research is to produce a clear and systematic analysis of storm surge impacts on property to consider how current and future development will need to act to minimize and, if possible, avoid losses from natural hazards. By laying out a repeatable methodology, others may use these methods to estimate impacts in other communities. A flow chart of the research process steps can be seen in Figure 3 1 below. A n online US Army Corps of Engineers resource was used to estimate future sea level rise and a FEMA Flood Insurance Study (FIS) was used to determine expected sea levels during a storm event in the study area. The Hazus MH tool was selected to model coastal flooding. Hazus was selected because of its relative ease of u se, its intuitive connection within ArcGIS architecture, the relative simplicity of required inputs, and the useful. ArcGIS software was used further to conduct impact analysis on developed areas and understand how future development may be impacted by st orm surge and sea level rise.
37 Figure 3 1 Generalized Research Process Steps Study Area Boundary Elevation Data Property & Parcel Data Florida 2070 Data Base 100 year (1% Chance) Storm 100 year (1% Chance) Storm with Additional Sea Level Rise Hazus Economic Impact Analysis Parcel Based Economic Impact Analysis Future Development Economic Impact Analysis
38 Data Acquisition There are four broad groups of data that are necessary to complete this research. They are: inputs into the Hazus MH m odel, Hazus MH outputs, parcel data from St. Johns County, and Florida 2070 development projection data. Hazus MH Inputs There are three major inputs into the Hazus MH model. Hazus requires the user to provide the study region which outlines the area of interest, a Digital Elevation M odel (DEM) which details elevations across the study region, and specific parameters for a coastal flooding scenario. With these inputs, the user can successfully run a coastal flooding analysis. Some of the data is provided out of the box in the Hazus s oftware or can be acquired through the prescribed workflow. However, the user can provide their own data for some inputs and must provide their own inputs for the scenario parameters. The study region i s selected within the Hazus model T he user can cho ose based on a spatial hierarchy from census block level up to an entire state. St. Johns county was selected as the study area. A custom DEM was used. The DEM was a mosaic of two DEMs, one created by the UF GeoPlan Center called FLIDAR and the other fr om the United States Geological Service National Elevation Dataset. The two DEMs were processed to ensure they covered St. Johns county, then pieced together using the Raster Calculator tool in ArcToolbox. The specific calculation can be found in the wor kflow described in Appendix A. Finally, scenario parameters were derived from FEMA FIS reports. These reports are created for each county and updated periodically. The FIS report describes 100 year flood still water elevation levels (SWEL) for different
39 coastal transects throughout the county. Using these values, the user can generate a single 100 year SWEL to be used for the entire coastline in the model or the study region coast line can be split up into multiple sections with different SWEL values. W ith these values, the user can then add on any additional potential sea level rise as desired The result is a number of different SWELs, each representing a certain coastal flooding scenario. The user must also determine the spatial datum, in this case NAVD88, which defines certain elevation reference points in the model. With these inputs in hand the Hazus coastal flooding model can be run. The parameters used in the various coastal floodi ng scenarios are determined from Flood Insurances Studies (FIS) produced by FEMA. These studies are conducted for each county and are updated periodically. Currently there is an updated FIS for St. Johns County that is under review, meaning that FEMA pro duced a final report, but may need to make some minor edits before releasing the finalized version. The repo r t was released on May 16 th 2016. Despite th e preliminary nature of the report, it was used in this study to help provide still water elevations (SWELs) for the study area because it supplied the most up to date information and is likely a good representation of conditions in St. Johns County. The 100 year SWELs wer e derived from the FIS report and used later the coastal flooding analysis. It is essential to recognize that the study area selected only models coastal flooding on the eastern side of St. Johns county. Though in reality flooding impacts may be observed on the western side of the county from the St. Johns River, these are not included in the model outputs. If the inclusion of this potential flooding source is desired, the study region would need to be expanded to so that hydro connectivity
40 between the Atlantic Ocean and the St Johns River is explicitly present. As the study region i s, the river is isolated from the ocean so only ocean side coastal flooding is modeled. Hazus MH Outputs There are two important outputs that the Hazus MH model generates which are used in further analysis in the research. These are flood depth grids and impact analysis outputs Flood depth grids come in the form of raster grids and are produced for the entire study area. Each cell in the grid contains a specific flood water depth and the size of the cell is determined by the resolution of the DEM input by the user. The flood depth grids are then used within the Hazus software as well as in analyses outside of Hazus. Maps showing the extent and depth of the flood depth grids for each flooding scenario can be seen in Appendix C The impact analysis outputs describe economic impacts to infrastructure such as buildings, essential facilities, and the transportation network. These outputs can be visualized in polygon layers within the ArcGIS framework or can be output into PDF tables. These values are used later to compare economic impacts of various flooding scenarios. Further description of these economic impact outputs can be seen in the St. Johns County Parcel Data St. Johns County parc el data is downloaded from the St. Johns County Property Appraiser. The property appraiser maintains the most up to date parcel and property datasets for the county. This data is used analysis, which will be describ ed later. Having the most up to date property information
41 ensures that the impact analysis is as accurate as possible. These datasets were downloaded from the St. Johns County Property Appraiser website ( ht tps://www.sjcpa.us/formsdata/ ). The files were downloaded on December 20, 2017 and came in the form of zipped file folders. The files downloaded were: GIS Data Bundle, CAMA Data Bundle, and CAMA Data Supplemental Bundle. The data in the GIS Data Bundle c ame in the form of a file geodatabase, while the two CAMA Bundles came in Microsoft Access databases and excel spreadsheets. The key feature class from the GIS Data Bundle, within the file geodatabase, was the Parcel feature class. The key tables within t he Access databases were the ParcelView table, from the CAMA Data Bundle and the BldView and the StructElemViewUnit tables, from the CAMA Data Supplemental Bundle. The Access tables were exported to excel spreadsheets within Access and then converted from excel files to file geodatabase tables using the Excel to Table tool in ArcMap. This method was used because it preserved a key piece of information, the Parcel ID Number, most effectively. This ensure d that linking attribute information between the par cel feature class and other tables through joins would be successful. Florida 2070 Development Projection Data In the Florida 2070 study, population growth projections were taken from the Florida Bureau of Economic and Business Research (BEBR) and a statew ide population of 33.7 million residents in 2070 was used. The baseline population condition used in the study is from the 2010 census, where a total population of
42 18,801,310 was counted. An increase of nearly 15 million people over 60 years serves as a defining value for the study (Carr & Zwick, 2016). Development densities for the Trend and Alternative scenarios were determined on a county by county basis and these densities were used to allocate population based on various land use categories. In the Trend Scenario, no new population was added to allocated to existing developed areas and was allocated to new development using suitability surfaces and the gross development density for each county. In the Alternative Scenario, some of the new population was added to existing urban areas with the remaining population added using similar suitability layers but with an increased development density. These differences in population allocation show how more undeveloped lands can be kept in their current stat e, protecting important agricultural resources and natural habitats. It also can help prevent development from straying into areas that may be vulnerable to future impacts of storm surge and sea level rise (Carr & Zwick, 2016). The Florida 2070 project produce d a variety of outputs These include reports outlining data inputs, methodology, and results. These documents provide stakeholders and other researchers insight into the project, showing how the methods may be replicated to generate the outputs d escribed or may be tweaked to fit other locations. However, the key datasets for this thesis research are spatial datasets describing the extent of development in Florida. T wo polygon layers are particularly important: one detailing the developed land ar ea in 2010 and one predicting the developed land area in 2070 if current trend land development practices continue These datasets will be used
43 to consider how predicted future development in St. Johns County may be impacted by different coastal flooding scenarios. Though Florida 2070 also produced the Alternative Scenario, this was left out of the analysis. This decision was made because assumptions regarding future development densities were more difficult to accurately replicate in the economic impact analysis. Only including the Trend development scenario in the future economic impact analysis minimizes assumptions that must be made when processing the data. The breakdown of land use types for the future developed ar eas in the Trend scenario can be derived from parcel information from St. Johns County Property Appraiser, mentioned above The 2010 Base and 2070 Trend spatial datasets were downloaded from the Florida Geographic Data Library FGDL ( https://fgdl.org/metadataexplorer/explorer.jsp ). Maps showing these datasets can be seen in Chapter 1, displayed in Figures 1 1 & 1 2 Hazus MH Model Descriptions The Hazus MH model is developed by FEMA Accordi nationally applicable standardized methodology that contains models for estimating potential losses from earthquakes, floods, and hurricanes. Hazus uses Geographic Information Systems (GIS) technology to estimate physical, economic, and social (FEMA, 2018) Hazus is a very useful tool because of the standardized and straightforward methodology for modeling impacts. This research utilizes flooding analysis within Hazus and more specifically the coastal flooding methodology The coastal flooding model enables the user to simulate anywhere from a 10% annual chance to a 0.2% annual chance or a 10 year to a 500
44 year, storm event. For this study, a 1% annual chance, or 100 year storm event, wa s used. The input data needed to run the coastal flooding model is described in the Hazus MH Inputs section above, however it is important to note the data that the user must input to successfully run the Hazus model is relatively light compared to othe r coastal flooding models The models that Hazus uses in coastal flooding analyses are simplifications of the erosion, Wave Height Analysis for Flood Insurance Studies (WHAFIS), and RUNUP models ( Department of Homeland Security, 2013 ). To successfully ru n a coastal flood model, there are a series of specific steps that must be followed. The workflow is somewhat lengthy, but is not complex. The specific steps can be found in Appendix A Figure 3 2 seen below, generally outlines the model workflow. Following this workflow, the user is provided with flood depth rasters. T he various flooding scenarios can then be used to estimate economic impacts. Figure 3 2 General Hazus MH Coastal Flood Model Workflow
45 Current Development Economic Impact Analysis Estimating the economic impact of coastal flooding events, with and without increases in sea level, is a key goal of this research. Listing the flood depths for and showing the extent of flooding in the study area gives readers a visual and physical understanding of imp acts. However, it can be difficult to gauge how those values translate to damage to infrastructure that people value, like homes and businesses. A simple way to gauge economic impacts is to consider the impact of flooding on buildings. Buildings have an inherent value that is not necessarily related to the location of the building, but is tied to the type of building, age, size, and materials used. This makes considering impacts on buildings an efficient proxy for measuring economic impact. In the anal ysis, only impacts to buildings are considered. Other infrastructure such as transportation and utilities network is not included. The Hazus software has this capability built in to the software. These have been described in the literature review. An issue noted with Hazus is that economic impacts to buildings are aggregated at the census block level. In doing this, some of the granularity of th e flooding analysis is lost. This is because aggregating parcels to a census bock generalizes impacts an d building valuations are spread evenly across census blocks. As a result, it is possible that Hazus overestimates economic impacts from flooding events because buildings are frequently built on higher elevation portions of parcels. This means that build ings may not actually be impacted by a flooding event even though Hazus counts it. In addition, the specific valuations of infrastructure are generalized, so an expensive building that may not be impacted will have its value averaged out across the census block.
46 In an effort to limit this overestimation of impacts an alternative economic impact analysis was performed outside of the Hazus tool. This alternative analysis was conducted at the parcel level, increasing the resolution to capture variations in flood area and depth over small areas. A detailed description of the workflow can be found in A ppendix B The general outline of the workflow can be seen in F igure 3 3 below. After completing the parcel level economic impact analysis, these values can b e compared to the economic impact outputs generated by Hazus. Figure 3 3 General Parcel Level Economic Impact Analysis Workflow It is important to note that the economic impact analysis conduct ed in Hazus and the parcel level economic impact analysis conducted outside of Hazus do not use the exact same input datasets or analysis procedures. Th e differences in methodology will lead to different impact results. With this in mind, comparing the outcomes can help gauge the strengt hs and weaknesses of each methodology. Comparing the outputs can still provide useful insights into what portions of the study area face particular risk and where steps may need to be taken to reduce hazard vulnerability. Detailed steps for 2. Clean Parcel Data
47 completing th e current development economic impact analysis can be found in A ppendix B Future Development Economic Impact Analysis Measuring impact of coastal flooding and sea level rise on current development is a crucial first step in ability and taking steps to improve resiliency moving forward. However, the ability to estimate the amount of impact on developed areas in the future is more difficult because it is hard to say where and how development will take place. The Florida 2070 report offers us this glimpse into the future, using sound assumptions and analysis methods to determine what areas will likely be built up in the future. Using the area outlined which is projected to be developed in 2070, we can estimate economic impact from coastal flooding events with various additional amounts of sea level rise added onto a 1% annual chance coastal flooding event. The future development economic impact analysis is similar to the current development economic impact analysis, with the addition of two other steps. After combining parcel data with depth damage curves, the area to be considered for future development must be determined This is done by using t he 2070 Trend devel oped lands layer dataset from the Florida 2070 study. Th is dataset w as acquired from FGDL. The development boundaries in th is layer include s the developed area from the base scenario. These areas were removed from the data layers so that only lands proj ected to be developed were considered. In addition, the parcels used in the current economic impact analysis were used to erase areas developed after 2010 from the 2070 Trend layer.
48 After isolating the future development area, the land use breakdown of th e future developed areas had to be established. For the Trend scenario, the same composition of land uses seen in the parcel data was applied to the future development lands. Th e land use breakdown w as determined from specific occupancy or generalized l and use, descriptions tied to specific depth damage curves. Based on the specific occupancy information, average valuations of the land could also be determined. Hazus draws from a database where this information is stored, and it can be used, along with inflation rates, to determine the value of future developed lands. It should be recognized that the 2070 Trend development scenario permits land to be developed in areas that may be vulnerable to coastal flooding and sea level rise. There are not necessa rily any limitations to development, such as not permitting building in floodplains. The qualifications for land that is considered for future development can be found in the Florida 2070 project documentation. By adding in these steps future developed areas could then be overlaid with the storm surge scenarios and the economic impact analysis process can be carried through. Detailed steps for the future development economic impact analysis can be found in A ppendix B
49 CHAPTER 4 COASTAL FLOODING AND SEA LEVEL RISE RESULTS Hazus Coastal Flooding and Sea Level Rise Outputs The first major set of outputs generated from th e analysis are flood depth grids. These grids, generated by the Hazus coastal flooding analysis, represent flood depth and extent for the study are a To make the flood depth grids more visually intuitive, the raster layers were clipped to land areas using a land cover dataset from FGDL Table 4 1 below summarizes some basic flood depth grid statistics for the various scenarios tha t were modeled. St. Johns County boundary was used as the area of consideration for calculating flood statistics Table 4 1 Flood Statistics for Coastal Flooding Scenarios Incorporating Varying Amounts of Sea Level Rise in St. Johns County Flooding Scen ario Feet of SLR Minimum Flood Depth Maximum Flood Depth Average Flood Depth Flood Area Base 0 0 ft 18.07 ft 4.42 ft 40826 ac 2070VLM 0. 13 0 ft 18.19 ft 4.54 ft 41669 ac 2070Low 0.98 0 ft 19.05 ft 5.11 ft 44576 ac 2070Int ermediate Low 1.25 0 ft 19.3 2 ft 5.24 ft 46776 ac 2070Int ermediate 2.23 0 ft 20.29 ft 6.02 ft 49617 ac 2070In termediateHigh 3.31 0 ft 21.38 ft 6.77 ft 52320 ac 2070High 4.49 0 ft 23.79 ft 7.68 ft 55265 ac 2070Extreme 5.41 0 ft 24.79 ft 8.32 ft 57760 ac Increases in sea level rise have a noticeable impact on variations in maximum flood depth, average flood depth, and area flooded. With additional amounts of sea level rise added on, it can be seen how the maximum flood depths increase. The maximum flood depth estimated in the study area ranges from just over 18 feet at the base sea level scenario to nearly 25 feet at the most extreme scenario. It is important to recognize that this increase is not linear, meaning that the maximum flood depth does not inc rease by the same amount of sea level rise. For example, in the VLM scenario 0.13 feet of sea level is added but the maximum flood depth increases by only 0.12 feet.
50 On the other hand, in the extreme scenario 5.41 feet of sea level is added and the maxim um flood depth increases by 6.72 feet. Average flood depth is also important to consider. With nearly 5.5 feet of sea level rise over the base, average flood depth for the study area nearly doubles. This increase in flooding is noticeable especially whe n looking at the outputs visually. In Figures 4 1 4 2 and 4 3 below, a portion of the study area, near the city of St. Augustine, is zoomed in to. Other maps of the flooding outputs can be seen in Appendix C.
51 Figure 4 1 Hazus Coastal Flooding Output Base Flooding Scenario in the St. Augustine Area
52 Figure 4 2 Hazus Coastal Flooding Output 2070 Low Flooding Scenario in the St. Augustine Area
53 Figure 4 3 Hazus Coastal Flooding Output 2070 High Flooding Scenario in the St. Augustine Area
54 CHAPTER 5 ECONOMIC IMPACT RESULTS Economic impacts were estimated for current and future conditions in the study area. Results are presented below. Current Development Impact Results Current Economic impacts for the study were using the Hazus program and using a parcel level analysis. First, the Hazus results will be presented, then the parcel level impacts will be presented. Hazus Economic Impact Outputs Hazus offers a variety of metrics describing economic impac ts. These ra nge from analyzing impacts to critical facilities, to transportation networks, to vehicles. In this research, economic impacts to buildings, excluding impacts to contents, were the focus. As described in the methodology, a variety of land use types were used to help determine how a range of flood depths would impact different building types. The Hazus software calculates the impacts within the software, with the user simply selecting a few options for what to include in the economic impact analysis. Res ults are provided in Table 5 1 below. Parcel level Economic Impact Outputs The parcel level economic impact analysis was conducted to compare how using finer detail in input data would affect economic impact outcomes. The parcel level analysis used data from the St. Johns County Property Appraiser and flooding outputs from Hazus to calculate impact estimates. Parcels were assigned a depth damage curve from the Hazus database based on the land use and building type. Using this
55 information in conjunction with flooding outputs resulted in economic impact estimates at the parcel level. Results are provided in Table 5 1 below. The Hazus impact results are greater than the parcel level impact results for the Base, 2070 VLM, 2070 Low, and 2070 Intermediate Low scenarios. However, the parcel level impact results are greater than the Hazus impact results for 2070 Intermediate, 2070 Intermediate High, 2070 High, and 2070 Extreme. Table 5 1 Current Economic Impact Results Scenario Hazus based Economic Impact (2018$) Hazus: Percent Increase in Damage over Base Parcel level Economic Impact (2018$) Parcel level: Percent Increase in Damage over Base Base 1,142,143,000 N/A 856,270,088 N/A 2070VLM 1,224,770,000 7.23% 934,325,935 9.12% 2070Low 1,635,292,000 43.18% 1,423,526,573 66.25% 2070IntermediateLow 1,822,872,000 59.60% 1,671,588,136 95.22% 2070Intermediate 2,474,940,000 116.69% 2,502,162,020 192.22% 2070IntermediateHigh 3,050,304,000 167.07% 3,265,762,801 281.39% 2070High 3,839,244,000 236.14% 4,108,012,602 379.76% 2070Extreme 4,436,897,000 288.47% 4,562,493,639 432.83% Future Development Impact Results Future economic impacts for the study were estimated using Hazus results and parcel level analysis coupled with Florida 2070 data. First, the Hazus results will be presented, then the parcel level impacts. Hazus Economic Impact Outputs Hazus only really offers economic impact analyses options using data from its database. Initially, the valu es presented only show the estimated value of the impacts in 2018 dollars. However, with the seven scenarios that use sea level rise estimations for the year 2070, some projections for the value of impacts at that date can be applied. Assuming an annual inflation rate of 2.0%, the value of the Hazus current economic
56 impact results, presented in the previously, can be calculated for the year 2070. The results are shown in Table 5 2 below. Parcel level Economic Impact Outputs Similar to the Hazus outputs, the seven parcel level impact outputs that included sea level rise estimations in 2070 initially only present impact estimations in 2018 dollars. Using the same annual inflation rate, 2.0%, the value of these impacts can be projected for 2070. The results are shown in Table 5 2 below. More significantly, future economic impacts have been calculated for areas projected to be developed by the Florida 2070 project. In these areas slated for new development, a ratio of land use/building types from currently developed parcels was used to forecast the makeup of the projected developed lands in 2070. Using th is information, valuation of these land uses per square foot, the area of the future developed areas inundated by coastal flooding and the average flood depth for each scenario the estimated economic impact to future development in 2070 from coastal flood ing events with additional sea level rise could be calculated. These values are presented in Table 5 3 below in both 2018 and 2070 dollars. Future development impact results show that from $1.6 billion to nearly $4 billion (2070$) of additional property may be impacted by coastal flooding in 2070 if current development continues in a similar pattern This increases vulnerability by about 30 % to 60%, depending on the flooding scenario.
57 Table 5 2 Future Economic Impact Results Scenario Hazus based Economic Impact (2018$) Hazus based Economic Impact (2070$) Parcel level Economic Impact (2018$) Parcel level Economic Impact (2070$) Base 1,142,143,000 N/A 856,270,088 N/A 2070VLM 1,224,770,000 3,429,757,725 934,325,935 2,616,419,076 2070Low 1,635,292,000 4,579,353,976 1,423,526,573 3,986,341,321 2070IntermediateLow 1,822,872,000 5,104,639,502 1,671,588,136 4,680,995,061 2070Intermediate 2,474,940,000 6,930,643,780 2,502,162,020 7,006,874,366 2070IntermediateHigh 3,050,304,000 8,541,851,699 3,265,762,801 9,145,207,014 2070High 3,839,244,000 10,751,142,472 4,108,012,602 11,503,782,715 2070Extreme 4,436,897,000 12,424,766,902 4,562,493,639 12,776,478,687 Table 5 3 Future Economic Impact Results Continued Scenario Future Development Economic Impact (2018$) Future Development: Percent Increase in Damage over VLM Future Development Economic Impact (2070$) Base N/A N/A N/A 2070VLM 589,527,807 N/A 1,650,871,224 2070Low 853,782,412 44.82% 2,390,870,793 2070IntermediateLow 884,521,028 50.04% 2,476,949,002 2070Intermediate 972,227,246 64.92% 2,722,555,180 2070IntermediateHigh 1,079,192,437 83.06% 3,022,092,799 2070High 1,244,143,437 111.04% 3,484,009,703 2070Extreme 1,386,474,730 135.18% 3,882,584,009
58 CHAPTER 6 DISCUSSION AND CONCLUSION Discussion of the outcomes from the coastal flooding analysis and economic impact analysis is found below. Conclusions from the outcomes are drawn and limitations are elaborated on, as well as opportunities for further research. Coastal Flooding from Base 100 Year Storm and Impacts Considering the vulnerability to a coastal flooding event in St. Johns County and estimated economic impacts, t he coastal flooding analysis shows how a 100 year storm could impact St. Johns County. Figure 6 1 and 6 2 presents the flooding that coastal areas in the county may experience in the event of a 100 year storm. In addition to some flooding on the beaches, e xtensive flooding can be seen along th e Matanzas, Guana, and Tolomato Rivers as well as along the Intracoastal Waterway. In many cases, it is the low lying areas close to the river mouths and along the river runs that are most vulnerable to flooding. This is somewhat counter intuitive becaus e coastal flooding events caused by hurricanes are generally expected to have the greatest impact on properties on or near the beach. However, the extensive dune system and slightly higher elevations found near the coastline reduces the vulnerability of t hese areas to flooding. The lower elevation of land along the rivers exposes them to greater flooding risk. Storm surge pushing water through inlets and up the rivers can impact many properties in these lower elevation areas. With this flooding vulnerability in mind, a 100 year storm with no sea level rise is estimated to flood nearly 40,000 acres of land in St. Johns County. This translates to about $1.1 billion of damage to buildings in the Hazus impact model and about $850 million of damage t o buildings in the parcel level impact model. The reason for this
59 difference is tied to the method for calculating impacts in each model. Hazus, because it is built for use anywhere in the US, can only predict economic impacts from flooding at the Census Block level. As a result, more specific information about building valuation and the impacts that certain parcels experience from flooding are lost in data aggregation. However, at the parcel level, these specifics are not lost They are factored into the overall model, taking into account valuations for individual properties and their location across a Census Block. By limiting aggregation and considering parcel level impacts, the parcel level model minimizes generalization and gives a more accurate e stimation of flooding impacts on buildings. It is also important to note that the Hazus database of building information is not as current as the parcel data used. This also helps explain discrepancies between the two impact estimations Additionally, t his impact is only to buildings in the study area and does include any other infrastructure.
60 Figure 6 1 Hazus Coastal Flooding Output Base Flooding Scenario for St. Johns County
61 Figure 6 2 Hazus Coastal Flooding Output Base Flooding Scenario i n the St. Augustine Area
62 Coastal Flooding from 100 Year Storms with Extra Sea Level Rise and Impacts Seven sea level rise scenarios were modeled to show a range of sea level rise increases and coastal flooding events may impact St. Johns County in the fu ture. It is clear that coastal flooding events with additional sea level rise will have a major impact. At the lowest sea level rise scenario, 2070 VLM, 0.13 feet of sea level rise translates to about 41,000 acres of land being inundated by a 100 year st orm event, 1000 acres more than the Base event. In the 2070 Extreme scenario, nearly five and a half feet of sea level rise coupled with a 100 year storm event is estimated to flood nearly 58,000 acres of land about 18,000 more than the base scenario M ore specifics for the various 100 year storm scenarios can be seen i n Table 4 1 in Chapter 4. Figures 6 3 and 6 4 show the extent of flooding in St. Johns County for the 2070 VLM and 2070 Extreme scenarios, respectively. All m aps of the various coastal f looding scenarios can be seen in Appendix C The economic impacts are also projected to increase with additional sea level rise. For the Hazus model, impacts to buildings range from $3.4 to $12.4 billion in 2070 dollars. F or the parcel level model, impacts to buildings range from $2.6 to $12.7 billion in 2070 dollars. Curiously, there is a point where impacts modeled in the parcel level analysis overtake the impacts modeled in the Hazus analysis. From the Base scenario to the 2070 Intermediate Low scenario, Hazus predicts greater impacts to buildings than the parcel level analysis. However, from the 2070 Intermediate scenario to the 2070 Extreme scenario, the parcel level analysis predicts greater impacts to buildings than Hazus. Though half of th e scenarios follow expectations, the scenarios with larger amounts of sea level rise do not. This may be for a variety of reasons, however there
63 are two that are most likely: first, the parcel level analysis captures flooding of more valuable buildings th at are only inundated by the flooding scenarios that include larger amounts of sea level rise, while Hazus generalizes impacts to such properties across Census Blocks. Second, the more current data used in the parcel level analysis has greater values for areas inundated in the higher sea level rise flooding scenarios than the Hazus database. Either of these reasons will cause the parcel level analysis to result in higher impact estimations compared to the Hazus analysis Regardless of the discrepancies b etween the models, variations in future sea level will clearly increase the economic impacts on buildings from a coastal flooding event in St. Johns County.
64 Figure 6 3 Hazus Coastal Flooding Output 2070 VLM Flooding Scenario for St. Johns Coun ty
65 Figure 6 4 Hazus Coastal Flooding Output 2070 Extreme Flooding Scenario for St. Johns County
66 Coastal Flooding Impacts on Future Development Future development in St. Johns County, projected by the Florida 2070 study, was overlaid with the various coastal flooding scenarios including sea level rise. Of the almost 180,000 acres of land projected to be developed in the 2070 Trend scenario, between about 4,000 and 7,200 acres was estimated to be inundated by the v arious coastal flooding scenarios. With this information it was possible to calculate the estimated economic impact to buildings in the future developed areas. Again, only impacts to buildings were included in the analysis, no other infrastructure was c onsidered. It was estimated coastal flooding would cause between $1.6 and $3.9 billion in damage to buildings Though only between about 2% to 5% of the total area projected to be developed would be inundated, it would add significantly to the impacts of a coastal flooding event in St. Johns Cou nty. Figures 6 5 and 6 6 below show examples of the flooding impacts to future developed areas an d Table 5 2 and 5 3 in Chapter 5 show dollar values of the impacts. A key consideration of this part of the analysis is how the 2070 Trend future development area was broken up into different land use/building type categories. The technique of using the land use breakdown for future development derived from the parcels used in the parcel level anal ysis is the most logi cal method. The current development pattern can be projected into the future to represent the trend of development. Beyond this method, it is difficult to estimate building values with any strong theoretical basis that far into the future. There are lon g range development plans, such as county comprehensive plans which lay out future land use, but these
67 typically extend no more than 30 years into the future. Using current development patterns, for which there is building valuation information, permits t he estimation of future development values. Also keep in mind that the location of future development may be in somewhat risky locations. In the scenarios with more dramatic sea level rise, local governments and regulatory agencies may be more hesitant t o permit building in low lying areas. As a result, the future economic impacts may be overestimated. However, it is difficult to judge and properly model the combination of factors affecting future development trends at this level. Simply continuing tre nds into the future generate useful outputs that can be used as a baseline
68 Figure 6 5. Hazus Coastal Flooding Output 2070 VLM Flooding Scenario, Impact on Future Development Near St. Augustine
69 Figure 6 6. Hazus Coastal Flooding Output 2070 Extr eme Flooding Scenario, Impact on Future Development Near St. Augustine
70 Additional Considerations It is interesting to note how some areas in the model are spared from flooding because of the presence of roads. In reality, this will likely not occur becau se of presence of culverts and other stormwater management infrastructure. Water will more easily pass under and around roads, flooding areas of similar elevation that are simply separated by a road. However, the Hazus coastal flooding model lacks some o f the finer scale hydro connectivity information so roads can act as a flood barrier. However, even small variations in input stillwater elevations lead to the overtopping of these barrie rs Figures 6 7 and 6 8 below show an example of this along A1A Be ach Boulevard near Anastasia State Park. After having spent a significant amount of time using the Hazus tool to generate coastal flooding and economic impact outputs, some seemingly straightforward modifications should be made to enhance t he capability o f the software. The parcel level analysis I conducted was necessary because this sort of analysis is difficult, if not impossible, to complete within the Hazus application. This is because Hazus conducts analyses a large scale, from the Census Block and h igher. It would be far too difficult and memory intensive for FEMA to include parcel data for every state in the country. However, the application developers could offer users the ability to add around five fields to an inventory data table. These field s c ould contain property attribute information such as specific occupancy ID, parcel/building value, number of floors, and first floor elevations. In addition, latitude and longitude information could be included for each record, representing the center of a parcel or the center of a building footprint.
71 With this information entered, a user would be a ble to produce economic impact outputs specific to their location within the Hazus tool. Creating a model in Model Builder or python script to use these fields as inputs and generate impact outputs would be straightforward. Overlaying this information wi thin the Hazus tool would streamline the process and reduce the potential for errors when trying to get more specific impact estimates.
72 Figure 6 7 Hazus Coastal Flooding Output Base Flooding Scenario, Flooding Issue Along A1A Beach Boulevard Near An astasia State Park
73 Figure 6 8. Hazus Coastal Flooding Output 2070 Low Flooding Scenario Flooding Issue Along A1A Beach Boulevard Near Anastasia State Park
74 Limitations As with all models, there is uncertainty. Literature can comparisons with other flooding studies has shown that Hazus outputs, similar to outputs of MOMs in the SLOSH model, may be representative of the worst case scenario for coastal flood ing events. This may be attributed to simplifications of various fluid dynamics and physics mod els included in the Hazus software. However, using these coastal flooding outputs give stakeholders in coastal areas useful information when considering where to locate critical infrastructure. It also can help inform planners and emergency management st aff under what storm conditions areas should begin to be evacuated. And stakeholders may begin to understand what economic losses from buildings can be expected given certain storm and sea level characteristics. Additionally, comparing Hazus impact and p arcel level impact model outputs is property datasets and spatial resolutions for estimating economic impacts from coastal flooding. However, understanding the differences in the outputs can help researchers more effectively model impacts in the future. By considering these data limitations, improvements can be made to the models to enhance accuracy of outputs. An important assumption made in the parcel based economic impact analysis is that all buildings in the study area have their first f loor elevation at ground level. This assumption was made because first floor elevations for all buildings in St. Johns County w ere not found The parcel data acquired from the county property appraiser d id not have this leve l o f detail This assumption may lead to overestimation of economic impacts because some buildings in the study area likely have first floors that are
75 elevated. Buildings with first floor elevations above ground level will face lower risk to flooding damage and see lower economic losses in the event of a flood. However, until more information is available for first floor elevations of buildings across St. Johns County, assuming ground level first floor elevations provides a standardi zed method in the analysis process. Property appraisers, especially in counties facing major flooding risks, could begin compiling first floor elevation information for properties. Maintaining this information would supply risk managers and planners with a valuable resource to better understand vulnerabilities to flooding Using only the value of buildings on a parcel as a proxy for economic impacts from a coastal flooding event can lead to additional underestimations of estimated impacts. Using only building values does not include any valuation of the land on which the building sits. The value of the buildings plus the value of the land determines the taxable value of the parcel, which determines a significant portion the property tax ba se for local governments If these combined values are permanently impacted by a coastal flooding event and/or sea level rise, local governments will face loss es to their tax base on top of having to manag e other impacts from these hazards. Other analyse s can be conducted where the taxable value of a parcel is used, rather than just the building value, to give local officials an idea of how they may be impacted by coastal flooding and sea level rise Looking into the potential permanent loss of income fr om property taxes will show additional vulnerability of a community to hazards. Suggestions for Future Research Future research efforts can be made to increase the resolution of the impact analysis. Specifically, the methods used in the parcel level anal ysis can be narrowed
76 further, so that flooding on building footprints is analyzed. Focusing the flooding impacts even further should help create an even more accurate estimate of the impacts from flooding on buildings. However, this effort would be very intensive. The process of creating building footprints for an entire county would require major investments of time and money to produce a high quality result. This research only focused on coastal flooding and sea level rise on the eastern side of St. Jo hns County, which is bordered by the Atlantic Ocean. However, further flooding analysis should be conducted on the western side of the county which is bordered by the St. Johns River. The St. Johns River in this area may not be exposed to as significant storm surge effects, but sea level rise will certainly have an impact on the western side of the county. To get a complete picture of how St. Johns County may be impacted by coastal flooding and sea level rise, the western side of the county should be inc luded in future research. In addition to using Hazus as the sole model to produce coastal flooding outputs, other models, like those mentioned in Chapter 2, could be used to create alternative flooding scenarios. Other models that require greater technic al knowledge may be able to produce more representative flooding outputs and, therefore, could give a more realistic idea of how a 100 year coastal flooding event may impact St. Johns County. Finally, estimating the impacts of coastal flooding events on i nfrastructure beyond buildings, like the transportation network, can give stakeholders a more complete view of how coastal flooding events may impact their community. Buildings area significant piece of the fabric of communities and represent major assets for
77 individuals and organizations, but they are not the only infrastructure impacted by a coastal flooding event. Concluding Thoughts It is clear that St. Johns County faces major flooding risk from strong hurricanes. Furthermore, if development is allow ed to continue in the future as it has leading up to now, many more built up areas will be vulnerable by 2070. It is in the best interest of stakeholders around coastal areas to consider how they may be impacted by storm surge right now and how the combin ation of storm surge and sea level rise may make outcomes worse in the future. An important step in working to minimize future vulnerabilities is to l imit development in low lying areas near inlets and along the river and intracoastal N ew ordinances or c odes preventing development in floodplains should be enacted to limit this vulnerability Further, implementing regulations that require buildings which were catastrophically damaged by a flooding event to be reconstructed with enhanced structural qualiti es or higher first floor elevations will minimize the risk of repeated damage to a single property. If development in risky areas cannot be curtailed, then imposing certain criteria for new development, such as establishing minimum first floor elevations above estimated 100 year flood levels with a n additional sea level rise value, would be a way to help mitigate vulnerability If steps are not taken by local governments to curtail unwise development in the future, they may be exposing themselves to major legal and financial risk. An example can be seen in the Jordan et al. v St. Johns County 2011 case If a local government permits development and serves that development with utility or transportation
78 infrastructure, that entity is responsible with maintaining the original level of service provided to the development In the face of more frequent severe storms and sea level rise, this could be an ultimately ba nkrupting proposition in the long run as maintenance costs progressively rise. Other lega l issues may arise if landowners decide to abandon their property in the face of storms and sea level rise which renders their property ultimately worthless. It is unclear who would be responsible for managing physical and environmental risks posed by der elict buildings slowly becoming inundated by rising seas and ripped apart by storms. Coastal areas around Florida, the US, and the world are vulnerable to storm surge and sea level rise. In the future these vulnerabilities are likely to increase as areas continue to develop and as risks of more powerful storms and rising seas increase. It is important for communities to understand these risks and how they expose infrastructure in their area to damage. Modeling the physical and economic impacts of current coastal flooding events and future events coupled with sea level rise can help communities better understand current and future vulnerabilities. Moving forward leaders can use this information to take steps to minimize potential impact s and increase the resiliency of their communities.
79 APPENDIX A HAZUS MH COASTAL FLOODING MODEL WORKFLOW Process Steps for Coastal Flooding Analysis Hazus MH 1. Start Hazus. From Hazus main menu, choose CREATE NEW STUDY AREA a. Enter name for study area b. Select hazard type: Earthquake, Flood, Wind. In this case, select FLOOD c. Choose analysis scale: state, county, watershed, census tract, census block, etc. In this case, select CENSUS TRACT d. Go through the various dialogue boxes until you have selected the desired census tracts for your study area. e. After selection is done, click OK and the study area will be created. You will then be returned to the Hazus main menu. 2. From Hazus main menu, choose OPEN STUDY AREA a. In new window, select the study are you want to conduct analysis on, likely the one you just created above. 3. In new ArcMap style window that opens, make note of the 4 Hazus specific drop down menus across the top ribbon: Inventory, Hazard, Analysis, Results 4. Click the Hazard drop down and click FLOOD HAZARD TYPE a. In the dialogue box that opens up, select the COASTAL ONLY option and click the OK button 5. Then click the Hazard drop down again and click USER DATA a. In the dialogue box that opens up, make sure that the DEM tab at the top is selected. Then click the Bro dialogue box.
80 b. Then you must select the DEM you want to use as the base data. After selecting the DEM data you want to use, click the OK button. Another dialogue box will open up to say that raster processing will occu r. Click OK. 6. After completing the USER DATA steps, click the Hazard drop down again then click SCENARIO a. In the dialogue box that opens up, enter a name for the new flooding scenario you want to run. b. After creating the new scenario, select the coastal shor eline you want to analyze using the button. Then click the save button and then OK. c. In the new dialogue box that opens up, manage the shorelines as needed. You can edit the shorelines or keep them as default. Then click the NEXT button. d. In the new dialogue box that opens up, enter in the still water elevations that you have derived from the FIS for the study area. Be sure to select the proper vertical datum and choose if you want to include wave setup and other information. 7. After entering in the necessary scenario information, click the Hazard drop down again and then click COASTAL a. In the dialogue box that opens, ensure that the information is correct for the coastal flood analysis and click OK. Another dialogue box will open that raster processi ng will occur. Click OK. The coastal flooding analysis will run until the flood depth grid has been produced.
81 8. After the process is completed, the depth grid will be produced
82 APPENDIX B ECONOMIC IMPACT ANALYSIS WORKFLOW Current Development Parcel Level Economic Impact Analysis Steps After downloading parcel data and property data, in the form of CAMA files, from the St. Johns County Property Appraiser, the data had to be processed and reorganized to facilitate data analysis. The following steps describe in detail the steps taken to reorganize the data. 1. Parcel data from the St. Johns Property Appraiser came in an ArcGIS file geodatabase. Within the file geodatabase, a single parcel feature class was present con taining all of the parcel boundaries and limited descriptive data for each parcel. The projected coordinate system of the data originally was NAD_1983_StatePlane_Florida_East_FIPS_0901_Feet However, to match the coordinate system of the various flood de pth grids, the parcel data was re projected to NAD_1983_UTM_Zone_17N. The new parcel layer was named 2. Next, the CAMA files were processed. The CAMA files came in the form of Microsoft Access Databases, with multiple tables found within each database. These databases contained property information such as parcel values, building values, and building structural information. a. Three tables were taken from the two access databases and converted into excel files: ParcelView, BldView, and StructElemView. b. These three excel files were then converted into file geodatabase tables in
83 3. After converting the tables to file geodatabase tables, the data was cleaned to remove unnecessary records and attribute information. a. The ParcelView table contained a field offering information about the total had no buildings, or no building value, were not included in the analysis. Records with tot_bld_val equal to zero were removed from the dataset by selecting records which had building values greater than zero and exporting these records into a new table call ed b. The StructElemView table contained information describing building specifications based on parcel number, with multiple records for a single parcel. One of the records contained information on the number of stories for each building. These records were selected and exported to a new c. No extra processing was required for the BldView table. 4. After cleaning the data tables, they could be joined. First, the 5. Next, the structelemunit_storiesonly table was joined with the parcel_bld_join table using the strap field as the join field. The joined tables were exported to a
84 6. The parcel_bld_struct_join table was joined to the parcels_projected feature class. The Str ap field was used as the join field. The joined feature class was 7. only parcels with a tot_bld_val greater than zero was selected. This selection After cleaning and reorganizing the parcel and property data, the flood depth grids had to be processed to ensure that only areas that needed to be analyze d were in the flood depth grid boundaries. This was accomplished by using land use and land cover data from St. Johns River Water Management District, supplied by the Florida Geographic Data Library (FGDL). 8. aded from FGDL 9. The lu_sjrwmd_2009_feb12 was then project from the FGDL Albers projection to UTM_17N using the project tool to match the flood depth grid layers. The projected layer was called lu_sjrwmd_2009_feb12_utm. 10. The lu_sjrwmd_2009_feb12_utm layer th en needed to be narrowed to the study area, so a select by location operation was done using the St. Johns county boundary. If any part of the lu_sjrwmd_2009_feb12_utm layer intersected the county boundary, it was selected. The selected features from the lu_sjrwmd_2009_feb12_utm layer were exported to a new layer called lu_sjrwmd_2009_feb12_utm_stjohncnty. 11. The lu_ sjrwmd_2009_feb12_utm_stjohncnty layer was then selected by attribute using a specific string. The selection was made to select open water area s and
85 then remove them so that a land only layer could be used to mask flood depth grids. The selection string used can be seen below: a. "LCCOD_DESC" = '9999: MISSING LUCODE OR OUTSIDE SJRWMD' OR "LCCOD_DESC" = '8370: SURFACE WATER COLLECTION BASINS' OR "LC COD_DESC" = '5100: STREAMS AND WATERWAYS' OR "LCCOD_DESC" = '5200: LAKES' OR "LCCOD_DESC" = '5300: RESERVOIRS PITS, RETENTION PONDS, DAMS' OR "LCCOD_DESC" = '5400: BAYS AND ESTUARIES' OR "LUCOD_DESC" = '8370: SURFACE WATER COLLECTION BASINS' OR "LCCOD_DE SC" = '5250: OPEN WATER WITHIN A FRESHWATER MARSH / MARSHY LAKES' OR "LCCOD_DESC" = '5430: ENCLOSED SALTWATER PONDS WITHIN A SALT MARSH' b. After making the selection, the selection was switched and the lu_sjrwmd_2009_feb12_ utm_stjohncnty layer was exported. The new layer was called lu_sjrwmd_2009_feb12_utm_watermask_final 12. The lu_sjrwmd_2009_feb12_utm_stjohncnty_watermask_final layer was then used to mask out undesired water areas from the flood depth grids. After cleaning and reorganizing the parcel and property data, the property data could be processed further to add information, such Specific Occupancy ID and depth damage curves, to be used in economic impact analyses. 13. dded. Then, based on the land use, building information, and number of stories, the Specific Occupancy was calculated. The attribute information in this field is coordinated with attributes from the depth
86 damage curves to ensure that information from the depth damage curves can be added to the parcel information. 14. A join was executed between the combined depth damage curves, derived from the parcel layer was matched with a simila r field in the depth damage curve layer and the attribute information for the layers was matched. The layer was then exported to make the join permanent. Preparing flood information to determine flooded area of parcels 15. Flood depth rasters for each coastal flooding scenario that had been selected to include only land areas were opened. These rasters were then converted to integer rasters using the Int tool. 16. These new integer rasters were multiplied by zero using the Times tool. 17. The zero rasters were then converted to polygons. 18. The polygon flood boundaries then needed to be clipped to the parcels, so that only the flooding in each parcel is considered. The Clip tool was used to isolate the flooding polygons to only the parcel boundaries. 19. The clipped fl ood polygons then need to have their information added to the parcel boundaries. This was done using the Intersect tool, but first a field called determine the square foot area of the parcel. The Intersect tool was then run once for each flood scenario polygon. Each intersected layer had a field added to
87 flood scenario title. Calculate geometry was run for each field so the square feet of flooding by scenario in each polygon was calculated. 20. After intersecting the parcels and the flood polygons, the resulting polygon layers fiel d (or similar field depending on the flooding scenario). In the spatial join tool, the match option used was CONTAINS. These spatial joins were added on top of prior spatial joins to produce a single parcel layer containing the flooded area information for each parcel for each flooding scenario. 21. Finally, 8 fields were added. Each to serve as a location to calculate the economic impact on a parcel for each flooding scenario. Impact calculations 22. With the final parcel layer in hand, zonal statistics could be r un for each flood depth grid using the final parcel layer as the zone. The zonal statistics could then be added to the parcel layer and the impact value calculated based on the minimum flood depth in each parcel. Minimum flood depth was used to provide a m ore conservative estimate of the flood impacts on a parcel.
88 Future Development Economic Impact Calculation Steps Prep the Future Development Layer 1. First, load the following Florida 2070 datasets. a. Florida2070_Dev_Base2010 b. Florida2070_Dev_Trend2070 c. Final parcel feature class used to calculate economic impact from flooding d. St Johns County, the study area boundary 2. First, project the two Florida2070 layers into the coordinate system matching the parcels feature class. 3. Next, clip the Trend2070 layer to the St Johns County boundary. 4. Next, erase the erased Trend2070 layer with the final parcels feature class. 5. Next, select only the features in the Trend2070 layer which have a DESCRIPT equal to TREND2070. Export the selection to a new layer. 6. Next, dissolve the features in the most recent output layer so there is only one record. Include as much information as you would like from the prior layer. 7. the area in square feet for the Trend2070 feature. 8. The final output, call it Trend2070_Final, is the future development layer that can be used to calculate future economic impact. Calculate Flooding Depths for Impact Analysis 9. Use the polygon flood rasters created during the parcel level flooding analysis and clip them to the Trend2070_Final layer.
89 10. Then, intersect each of those with the Trend2070_Final layer, creating 7 intersected feature classes. Then add a field to each of the intersected layers various flood scenario titles as necessary. 11. Next, spatial join these intersected layers to the Trend2070_Final feature class, flood scenario). In the spatial join tool, the MATCH OPTION used was INTERSECT. These spatial joins were added on top of prior spatial joins to produce a single parcel layer containing all of the spatial information regarding flooding extent in the future development ar ea. The output is called Trend2070_Final_spatialjoinFinal. Add 7 fields to serve as a location for summing total impact value for each flooding scenario. 12. Using the Trend2070_Final_spatialjoinFinal layer, run zonal statistics on the flood depth grids for the various 2070 flooding scenarios. a. 2070vlm b. 2070low c. 2070intlow d. 2070int e. 2070inthigh f. 2070high g. 2070extreme 13. Then join all of these zonal stats tables to the spatialjoinFinal layer, so that each flooding scenario has flood depth information in the table.
90 Econo mic Impact Analysis Calculations 14. Using the Trend2070_Final_spatialjonFInal_join_final layer, you can calculate economic impacts. 15. For each scenario, get the total Trend2070 future development area that is flooded and the mean flood depth from the Trend2070_ Final_spatialjonFInal_join_final layer and information from the future_devel_impact_info table. 16. Will need to add fields to the future_devel_impact_info table to help capture information about flooded area for each scenario and average flood depth for each scenario
91 APPENDIX C COASTAL FLOODING MODEL OUTPUT MAPS Figure C 1. Hazus Coastal Flooding Output Base Scenario Overview
92 Figure C 2 Hazus Coastal Flooding Output 2070 VLM Scenario Overview
93 Figure C 3 Hazus Coastal Flooding Output 2070 Low Scenario Overview
94 Figure C 4 Hazus Coastal Flooding Output 2070 IntLow Scenario Overview
95 Figure C 5 Hazus Coastal Flooding Output 2070 Intermediate Scenario Overview
96 Figure C 6 Hazus Coastal Fl ooding Output 2070 IntHigh Scenario Overview
97 Figure C 7 Hazus Coastal Flooding Output 2070 High Scenario Overview
98 Figure C 8. Hazus Coastal Flooding Output 2070 Extreme Scenario Overview
99 Figure C 9. Hazus Coastal Flooding Output Base Scenario Zoomed In
100 Figure C 10. Hazus Coastal Flooding Output 2070 VLM Scenario Zoomed In
101 Figure C 11. Hazus Coastal Flooding Output 2070 Low Scenario Zoomed In
102 Figure C 12. Hazus Coastal Flooding Output 2070 IntLow Scenario Zoomed I n
103 Figure C 13. Hazus Coastal Flooding Output 2070 Intermediate Scenario Zoomed In
104 Figure C 14. Hazus Coastal Flooding Output 2070 IntHigh Scenario Zoomed In
105 Figure C 15. Hazus Coastal Flooding Output 2070 High Scenario Zoomed In
106 Figure C 16. Hazus Coastal Flooding Output 2070 Extreme Scenario Zoomed In
107 REFERE N CES Bureau of Economic Analysis. (2017). Florida Washington, DC: US Department of Commerce. Retrieved from https://www.bea.gov/regional/bearfacts/pdf.cfm?fips=12000&areatype=STATE&g eotype=3 Carr, M., & Zwick, P. (2016). Florida 2070 Technical Report Gainesville, FL: University of Florida GeoPlan Center. Retrieved from http://1000friendsofflorida.org/florida2070/wp content/upload s/2016/09/florida2070technicalreportfinal.pdf Chen, C., Beardsley, R., Luettich, R., Westerink, J., Wang, H., & Perrie, W. et al. (2013). Extratropical storm inundation testbed: Intermodel comparisons in Scituate, Massachusetts. Journal Of Geophysical Res earch: Oceans 118 (10), 5054 5073. http://dx.doi.org/10.1002/jgrc.20397 Delft University of Technology (2017). SWAN Simulating Waves Nearshore Retrieved 5 January 2018, from http://swanmodel.sourceforge.net/ Department of Homeland Security. (2013). Flood Model Hazus MH Technical Manual Washington, D.C.: FEMA. Retrieved from https://www.fema.gov/media library data/20130726 1820 25045 8292/hzmh2_1_fl_tm.pdf Federal Emergency Management Agency. (2018). Hazus Overview fema.gov Retrieved 5 January 2018, from https://www.fema.gov/hazus mh overview Federal Emergency Management Agency. (2018). RUNUP, Version 2.0 fema.gov Retrieved 5 January 2018, from htt ps://www.fema.gov/runup version 20 Federal Emergency Management Agency. (2018). Wave Height Analysis for Flood Insurance Studies, Version 4.0 FEMA Retrieved 5 January 2018, from https://www.fema.gov/wave height analysis flood insurance studies version 40 Florida Bureau of Economic and Business Research. (2015). Florida Estimates of Population 2015 Gainesville, FL: College of Liberal Arts and Sciences, Universit y of Florida. Retrieved from https://www.bebr.ufl.edu/sites/default/files/Research%20Reports/estimates_2015 .pdf Florida Ocean Alliance. (2013). Florida's Oc eans and Coasts: An Economic and Cluster Analysis Fort Lauderdale, FL. Retrieved from http://www.floridaoceanall iance.org/wp content/uploads/2015/08/FLORIDAS_OCEANS_AND_COASTS_AN_ECONOMI C_AND_CLUSTER_ANALYSIS.pdf
108 Frank, K., Volk, M., & Jourdan, D. (2015). Planning for Sea Level Rise in the Matanzas Basin: Opportunities for Adaptation Gainesville, FL: University of Florida. Retrieved from https://planningmatanzas.files.wordpress.com/2012/06/planning for sea level rise in the matanzas basin1.pdf Frazier, T., Wood, N., & Yarnal, B. (2010). Stakeholder perspectives on land use strategies for adapting to climate change enhanced coastal hazards: Sarasota, Florida. Applied Geography 30 (4), 506 517. http://dx.doi.org/10.1016/j.apgeog.2010.05.007 Genovese, E., & Green, C. (2014). Assessment of storm surge dama ge to coastal settlements in Southeast Florida. Journal Of Risk Research 18 (4), 407 427. http://dx.doi.org/10.1080/13669877.2014.896400 Hall, J.A., S. Gill, J. Obeysekera, W. Sweet, K. Knuuti, and J. Marburger. 2016. Regional Sea Level Scenarios for Coastal Risk Management: Managing the Uncertainty of Future Sea Level Change and Extreme Water Levels for Department of Defense Coastal Sites Worldwide Al exandria, VA: U.S. Department of Defense, Strategic Environmental Research and Development Program. 224 pp. IPCC, 2014. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 pp Jones, C. & Griffis, L. (2013). Advances in Hurricane Engineering Learning from Our Past 54. Advanced Estimation of Coastal Storm Surge: Application of SWAN+ADCIRC in Georgia/Northeast Florida Storm Surge Study American Society of Civil Engineers (ASCE). Retrieved from https://app.knovel.com/hotlink/pdf/id:kt00CCL575/advances in hurricane/advanced estimation coastal Karamouz, M., Fer eshtehpour, M., Ahmadvand, F., & Zahmatkesh, Z. (2016). Coastal Journal Of Irrigation And Drainage Engineering 142 (6), 04016016. http://dx.doi.org/10.1061/(asce)ir.1943 4774.0001017 Kleinosky, L., Yarnal, B., & Fisher, A. (2006). Vulnerability of Hampton Roads, Virginia to Storm Surge Flooding and Sea Level Rise. Na tural Hazards 40 (1), 43 70. http://dx.doi.org/10.1007/s11069 006 0004 z Kopp, R., Horton, R., Little, C., Mitrovica, J., Oppenheimer, M., & Rasmussen, D. et al. (2014). Probabilistic 21st and 22 nd century sea level projections at a global network of tide gauge sites. Earth's Future 2 (8), 383 406. http://dx.doi.org/10.1002/2014ef000239
109 Linhoss, A., Kiker, G., Shirley, M., & Frank, K. (2015). Sea Level Rise, Inundation, and Marsh Migration: Simulating Impacts on Developed Lands and Environmental Systems. Journal Of Coastal Research 299 36 46. http://dx.doi.org/10.2112/jcoastres d 13 00215.1 Luettich, R., & Westerink, J. (2017). ADCIRC adcirc.org Retrieved 5 January 2018, from http://adcirc.org/ National Oceanic and Atmospheric Association, (2017). Sea Level Trends tidesandcurrents.noaa.gov Retrieved 11 November 2017, from https://tidesandcurrents.noaa.gov/sltrends/sltrends_station.shtml?stnid=8720218 National Hurricane Center. (2018). SLOSH Mod el Sea, Lake, and Overland Surges from Hurricanes (SLOSH) Retrieved 5 January 2018, from http://www.nhc.noaa.gov/surge/slosh.php Office of Economic and Demographic Research. (2017). Population and Demographic Data Florida Products edr.state.fl.us Retrieved 12 January 2018, from http://edr.state.fl.us/Content/population demographics/data/index fl oridaproducts.cfm Oxford Economics. (2018). The Economic Impact of Out of State Visitors in Florida: 2016 Calendar Year Analysis Tallahassee, FL: Visit Florida. Retrieved from https://www.visitflorida.org/media/30679/florida visitor economic impact study.pdf Parris, A., Bromirski, P., Burkett, V., Cayan, D., Culver, M., & Hall, J. et al. (2012). Global Sea Level Rise Scenarios for the United States National Climate Assessment Silver Spring, MD: National Oceanic and Atmospheric Administration Office of Oceanic and Atmospheric Research. Peng, B. (2015). The Cost Benefit Analysis of Building Levees to Mitigate the Joint Effects of Storm Surge and Sea Level Rise. [electronic resource] A Case Study in City of Miami, Florida [Gainesville, Fla.] : University of Florida, 2015. R.S. Means. (2006). Means Square Foot Costs (27th ed.). Duxbury, Mass. Sebastian, A., Proft, J., Dietrich, J., Du, W., Bedient, P., & Dawson, C. (2014). Characterizing hurricane storm surge behavior in Galveston Bay using the SWAN+ADCIRC model. Coastal Engineering 88 171 181. http://dx.doi.org/10.1016/j.coastaleng.2014.03.002 Stanton, L., M. Davis, and A. Fencl. 2010. Costing Climate Impacts and Adaptation. A Canadian Study on Coastal Zones The National Round Table on the Environment and the Economy Somerville, MA: Stockholm Environmental Institute.
110 Sweet, W., Kopp, R., Weaver, C., Obeysekera, J., Horton, R., Thieler, E., & Zervas, C. (2017). Global and Regional Sea Level Rise Scenarios for the United States Silver Spring, MD: National Ocean Service Center for Operational O ceanographic Products and Services National Oceanic and Atmospheric Administration. Retrieved from https://tidesandcurrents.no aa.gov/publications/techrpt83_Global_and_Regional_ SLR_Scenarios_for_the_US_final.pdf Tol, R. (2002). Estimates of the Damage Costs of Climate Change. Part 1: Benchmark Estimates. Environmental And Resource Economics 21 (1), 47 73. http://dx.doi.org/10.1023/a:1014500930521 Trigaux, R. (2016). Florida GDP Outpacing U.S. Economic Output Thanks to Strong Real Estate, Construction Sectors. Tampa Bay Times Retrieved from http://www.tampabay.com/news/business/florida gdp outpacing us economic output thanks to strong real estate/2287087 USACE Sea Level C hange Curve Calculator (2017). Climate Preparedness and Resilience Retrieved 11 November 2017, from http://www.corpsclimate.us/ccaceslcurves.cfm Withey, P., Lantz, V., & Ochuodho, T. (2015). E conomic costs and impacts of climate induced sea level rise and storm surge in Canadian coastal provinces: a CGE approach. Applied Economics 48 (1), 59 71. http://dx.doi.org/10.1080/00036846.2 015.1073843 Xie, D., Zou, Q., & Cannon, J. (2016). Application of SWAN+ADCIRC to tide surge and wave simulation in Gulf of Maine during Patriot's Day storm. Water Science And Engineering 9 (1), 33 41. http://dx.doi.org/10.1016/j.wse.2016.02.003 Yohe, G., Neumann, J., Marshall, P., & Ameden, H. (1996). The economic cost of greenhouse induced sea level rise for developed property in the United Sta tes. Climatic Change 32 (4), 387 410. http://dx.doi.org/10.1007/bf00140353 Zimmerer, G., Anderson, M., Bellomo, D., Blanton, B., Buckerfield, B., & Collier, K. et al. (2007). Atlantic Ocean and Gulf of Mexico Coastal Guidelines Update Denton, TX: FEMA. Retrieved from https://www .fema.gov/media library data/1388780453134 c5e577ea3d1da878b40e20b776804736/Atlantic_Ocean_and_Gulf_of_Mexico_C oastal_Guidelines_Update_(Feb_2007).pdf
111 BIOGRAPHICAL SKETCH Adam Carr began his post secondary education at the University of North Caroli na at Chapel Hill in the Fall of 2009. While there he focused his studies i n e nvironmental s cience, minoring in g eography and e ntrepreneurship. He was fortunate enough to participate in a variety of fantastic programs, including study abroad trips to the US Virgin Islands and Bangkok, Thailand. The experiences he had especially in the Thailand program, got Adam interested in ur ban planning. After graduating from UNC in 2013, Adam spent two years working for the National Park Service and the US Forest S ervice. These positions allowed him to use some of his skills learned during his undergraduate studies and see different parts of the country before returning to school. Adam began his graduate studies at UF in the Fall of 2015 and quickly became interes ted in focusing his studies on the application of GIS and spatial analysis in the context of planning. A Florida native, the issue of sea level rise and impacts from hurricanes to the built environment was always of interest to him. Focusing his research on analyzing vulnerability of coastal communities he frequented growing up and helping these communities understand what steps can be taken moving forward has been a highlight of his graduate studies at UF. After completing his studies at UF, Adam hopes to continue using GIS as a tool to help provide valuable data and analysis to stakeholders involved in any sort of planning process. This will help generate informed planning decisions that minimize negative externalities and encourage sustainable, equita ble, and resilient outcomes.