Utiliizing Hazus-Mh and Sea Level Affecting Marshes Model (SLAMM) to Identify Floodplain Conservation Areas for Coastal Hazards

UTILIZING HAZUS-MH AND SEA LEVEL AFFECTING MARSHES MODEL (SLAMM) TO IDENTIFY FLOODPLAIN CONSERVATION AREAS FOR COASTAL HAZARDS A PAPER SUBMITTED FOR COMPLETION OF SENIOR RESEARCH FOR THE COLLEGE OF ARTS AND SCIENCES STETSON UNIVERSITY BY Emily Alyssa Niederman IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE(S) OF BACHELOR OF SCIENCE ENVIRONMENTAL SCIENCE ADVISOR Dr. Jason M. Evans, Ph.D. MAY 2017 i Table of Contents List of Illustrations.……………………………………………………………………………….……….ii List of Tables……………………………………………………………………………………..……….iii Acknowledgments……………………………………………………………………………….…….….iv Abstract…………………………………………………………………………………………………....v Introduction………………………………………………………………………………………………..1 Literature Review………………………………………………………………………………………….2 Study Area…………………………………………………………………………………………………5 Methodology……………………………………………………………………………………………….5 Hazus-MH………………………………………………………………………………7 SLAMM………………………………………………………………………………...9 The Overlay…………………………………………………………………………….10 Results…………………………………………………………………………………………………….12 Discussion and Conclusions………………………………………………………………………………12 Figures and Tables………………………………………………………………………………………...15 Appendix…………………………………………………………………………………………………..23 Works Cited……………………………………………………………………………………………….29 ii List of Illustrations Figure 1: Hazus-MH Flood Model of the 100 Year Flood Return Period……………………………....18 Figure 2: Vulnerability Assessment for a 100 Year Flood Return Period………………………………19 Figure 3: Prioritized U.S. Census Blocks for Coastal Floodplain Mitigation…………………………...20 Figure 4: Sample Area of Prioritized U.S. Census Blocks for Coastal Floodplain Mitigation………….22 iii List of Tables Table 1: SLAMM Parameters………………………………………………………………………….15 Table 2: SLAMM Landcover Classification Code…………………………………………………….16 Table 3: SLAMM Crosswalk for Tabulation…………………………………………………………..17 Table 4: Tabulated Results from the Overlay………………………………………………………….21 iv Acknowledgments I acknowledge Florida Sea Grant, the Dean’s Fund at Stetson University, and Stetson University’s Institute for Water and Environmental Resilience for sponsoring this project. This enabled the research to be shared at the Association of American Geographer’s Annual Meeting in April 2017. My deepest gratitude goes to the support from my family and friends and especially for the mentorship that I received from Dr. Evans and Dr. Abbott. v Abstract Local governments have limited options when facing coastal floodplain hazards, such as hurricanes, storm surge, flooding, and sea level rise. Some of the most common strategies include active planning (adaptation), passive observation (do nothing), or wholesale retreat (abandonment). One suite of adaptation actions involves conservation and restoration of natural ecosystems to buffer the potential damages caused by coastal hazards. However, a challenge for coastal floodplain managers is the identification of locations where ecosystem restoration is most likely to be sustainable and resilient under the range of possibilities implied by climate change. This study proposes and implements a geographic information systems (GIS) methodology for selecting adaptation action areas in Florida’s Upper Indian River Lagoon watershed. Two GIS models, the Federal Emergency Management Agency’s Hazus Multi-Hazard (Hazus-MH) and the Sea Level Affecting Marshes Model (SLAMM), are used to create a weighted overlay analysis that identifies and ranks the suitability of sites for sustainably implementing an integrated ecosystem restoration and floodplain mitigation strategy. For Volusia County, Florida, U.S.A., there were a total of 13,539 substantially damaged buildings within the vulnerable area, of which 774 buildings were located in suitable floodplain mitigation strategy sites with over 80 percent modeled landcover change. The results of this analysis can inform long-term disaster mitigation planning and climate change adaptation strategies at a local, regional, and federal level. 1 Utilizing GIS-based models, Hazus-MH and Sea Level Affecting Marshes Model (SLAMM), to identify floodplain conservation areas for coastal hazards Introduction Coastal communities are challenged to develop practical strategies of adaptation to increase the community’s resilience when facing hazards, such as sea level rise (SLR), flooding, coastal erosion, tropical storms, and hurricanes. The level of expertise, funding, and time for a thoughtful research study can be quite extensive. For example, the City of Satellite Beach, Florida, has an ongoing history with SLR vulnerability assessments, ranging from a geographic information systems and science (GIS) based bath-tub model by Parkinson and McCue (2011) to an ongoing critical vulnerability assessment (Evans et al. 2015). The City of Satellite Beach was the second city in the state of Florida to adopt the Adaptation Action Area language to prepare for sea level rise. However, the City of Satellite Beach has not developed an adaptation strategy beyond identifying threatened areas and naming them Adaptation Action Areas. Currently, there is an ongoing project led by Evans et al. (2015) that is specifically focusing on identifying the vulnerable areas using a GIS-based model from the Federal Emergency Management Agency (n. d.) called Hazus Multi-hazard (Hazus-MH). That information will then be used to initiate the adaptation conversation through community engagement. If an adaptation strategy was proposed with data to support it, then it could be added to the conversation that is currently ongoing at Satellite Beach. A forward-thinking adaptation strategy would be to add wetland remediation areas into the future land use codes for mitigating the influx of salt and fresh water and thus prevent flood damage to human property. This idea of natural area restoration for disaster hazard mitigation has been examined elsewhere. For example, the cost-benefit analysis by Kousky and Walls (2014) utilized the Hazus-MH flood model in St. Louis County, Missouri looking at incorporating a floodplain conservation strategy for mitigating future flood-related damages (Federal Emergency Management Agency n. d.). Contrasting with Kousky and Walls (2014), Satellite Beach faces coastal hazards instead of riverine flooding hazards and the ongoing project in the City of Satellite Beach is only identifying the vulnerable areas rather than a cost-2 benefit analysis of a mitigation strategy to those vulnerable areas (Evans et al. 2015). Additionally, neither the ongoing study for the City of Satellite Beach nor the study by Kousky and Walls (2014) looked at the floodplain mitigation conservation strategy from the side of the imposed and/or restored ecosystems themselves. Therefore, the aim of this project is to study where natural areas and wetlands could be reclaimed for hazard mitigation by comparing the modeled results of projected future locations of wetlands with the projected frequently flooded areas. These suggested areas could then be used to further the ongoing conversations in places to implementing a floodplain conservation mitigation strategy against coastal hazards. In this way, the objective was to establish a methodology by which coastal communities could not only assess, but then prioritize where to implement adaptations against coastal hazards for the protection of the vulnerable areas and thus the resilience of the communities. For this study, the region of Volusia County was selected as it is a part of the Upper Indian River Lagoon system and can be easily used as an example system. Literature Review There are many different types of models when it comes to vulnerability assessment for sea level rise. For a global picture, Bosello and De Cian (2014) compared what they termed as bottom-up versus top-down models. The bottom-up approach involves assigning an arbitrary value to the vulnerable areas, while the top-down approach involves summing up the entire economies and then having the modeled impacted areas then detract from the economic system. However, Bosello and De Cian (2014) stressed that the data should be reviewed in local conditions and that coastal mitigation strategies seem to be cost-effective overall in their conclusion. Out of all the GIS models when pertaining to sea level rise, the generic bathtub model is one of the most well-known and it can be used for several applications. The basic idea is that land at the lowest elevation will flood first. For instance, Parkinson and McCue (2011) modeled in three dimensions the City of Satellite Beach based off a Digital Elevation Model (DEM) in GIS and then evaluated the 3 magnitude of sea level rise necessary to flood the study area. Though Parkinson and McCue (2011) did create a three-dimensional model for their variant of a bathtub model, it is not required. Johnston et al. (2014), in their infrastructure assessment for Portland, Maine, based the elevation ranges off the Highest Astronomical Tide (HAT), with the HAT being ground zero, with another 0.6 meters (2 ft.) for coastal hazard coverage, and the worst historical storm’s data of that area for an additional 0.6 meters, totaling a 1.2 meters (4 ft.) above HAT. These ranges were gathered from a tide gauge at Portland, Maine, and the general elevation data was based upon Light Detection and Ranging (LiDAR) aerial surveying techniques and DEMs. Similarly, Murali and Kumar (2014) ran a bath-tub model in India, but instead of analyzing it for critical infrastructure, they assessed landcover and land use change. Although these three papers are all examples of bathtub models, the first two use the bathtub model as a vulnerability assessment whereas the third analyzes land cover change. Although these studies independently cover vulnerability assessments separately from land cover change, the combination of a coastal hazard analysis with a landcover change for a locality is unique in the literature. In terms of coastal hazard modeling, guidance from the Federal Emergency Management Agency (FEMA) is to prepare for the disaster before it comes and to be prepared for the disaster requires vulnerability assessments, which is why FEMA developed the Multi-hazard Mapping Initiative (MMI) tool and Hazus-MH (Mahendra et al. 2011; Federal Emergency Management Agency n. d.). Mahendra et al. (2011) took this advice and assessed the vulnerabilities of a region, which was one of the worst hit in the 2004 Indian Ocean tsunami—Cuddalore, Villupuram, India—with an inclusion of sea level changes, shoreline changes, extreme storm surges and return periods, and elevation. This vulnerability assessment led to a refined disaster evacuation plan and has led to higher community preparedness. Similarly, a study on a possible mitigation strategy, such as this one, for coastal disaster could be used to prevent damages in the scenario of a disaster and thus raise disaster preparedness levels. But beyond coastal hazard vulnerability assessments, the question should be raised of whether any sort of mitigation plan could be beneficial economically to a community. Kousky and Walls (2014) used Hazus-MH to designate the floodplain for the Meramec River in St. Louis County, Missouri 4 (Federal Emergency Management Agency 2016). This was the prep work to run a cost-benefit analysis to see whether the benefits of floodplain conservation do outweigh the costs of implementing the adaptation strategy, and they found that it does indeed have a modeled favorable return on investment. Although the study area for Kousky and Walls (2014) faces riverine flooding rather than coastal hazards, the application of Hazus-MH to generate the floodplain area is a viable method to use. The results suggest that a floodplain mitigation strategy does indeed benefit a community when facing disasters and Kousky and Walls (2014) propose that similar results could be found elsewhere. What Kousky and Walls (2014) did not do was to suggest which areas of the floodplain are modeled to be prioritized when facing climate change. In anticipation of landcover changes associated with global climate change, the U.S. Fish and Wildlife Service teamed up with Warren Pinnacle to run a GIS-based model called Sea Level Affecting Marshes Model (SLAMM). SLAMM uses various sea level rise curves to show where the ecosystems will migrate because of the changes in the saltwater flooding regime (Warren Pinnacle 2016). Of the various studies completed by Warren Pinnacle for this research project, three studies are in relative proximity to the City of Satellite Beach: Archie Carr National Wildlife Refuge (2010a), Merritt Island National Wildlife Refuge (2010b), and Pelican Island National Wildlife Refuge (2011). These studies were models of how the landscape was anticipated to change due to sea level rise within the wildlife refuges, but similar studies can be computed for different regions using the SLAMM software. Because SLAMM can be used to show how the natural areas could tend to migrate “naturally,” placing conservation areas where there is already a “natural” tendency from a developed landscape to a “natural” landcover could be used to designate the floodplain conservation mitigation areas. The idea of using mitigation areas to protect against coastal hazards can be found in the literature (Warren Pinnacle 2016; Federal Emergency Management Agency n. d.). In the Monterey Bay Area, California, Langridge et al. (2014) ran a coastal vulnerability model called InVEST which quantifies the vulnerability of a coastline depending on what kinds of “natural infrastructure” is serving to mitigate the potential damages. For their study area, Langridge et al. (2014) ran the model with multiple variants as to 5 how much “natural infrastructure” was present and then the model assessed the coastal hazard vulnerability and the results showed that “natural infrastructure” reduced the study region’s coastal hazard vulnerability. Langridge et al. (2014) shared the results with the regional planners to aid in climate change adaptation planning, similarly to how the project in the City of Satellite Beach by Evans et al. (2015) is in collaboration with the regional planners, and how the results from this proposed project will be shared. The results of the Langridge et al. (2014) study along with the Kousky and Walls (2014) suggest a valid legitimization of the adaptation strategy as a possibility to be studied from both the coastal hazard vulnerabilities and the floodplain conservation sides of the literature as to being a cost-saving strategy. Study Area The extent of the study region will entail a portion of the Upper Indian River Lagoon, focusing on Volusia County, although aspects of the research extend beyond the political boundaries, due to a consideration for the boundaries of the watershed. For instance, the actual extent of the digital elevation models and landcover datasets have larger boundaries than Volusia County itself, so the region SLAMM extends beyond the county borders (Warren Pinnacle 2016). Additionally, the digital elevation model (DEM) necessary is derived from Hazus-MH, the model defines the exact study region extent (Federal Emergency Management Agency n. d.). Because the DEM is required to run both SLAMM and Hazus-MH, the first model that is run for the study is Hazus-MH. Methodology The purpose of this methodology is to identify first where a coastal community is vulnerable to coastal hazards and then secondly to identify where a flood plain mitigation strategy would best be implemented in the vulnerable areas. This would be accomplished by overlaying the results from running Hazus-MH with the results from SLAMM—creating a selected region in which the “natural” landcovers are modeled to overtake a populated region that is vulnerable to coastal hazards (Federal Emergency 6 Management Agency n. d.; Warren Pinnacle 2016). The overlying areas would then be proposed as a possible wetland mitigation area following floodplain conservation methodologies as the results from SLAMM suggest that the desired types of mitigation ecosystems are already projected to inhabit that area. When running Hazus-MH itself, the flood model will be run to identify the areas that are considered vulnerable to coastal hazards in the study region. These vulnerable areas are used to identify the location for the coastal hazard mitigation areas by following the floodplain conservation methodology. There are numerous data outputs for Hazus-MH, but the item of interest are the regions that Hazus-MH determines to be vulnerable. These are polygon shapefiles which were used in the final overlay later. The next model is the Sea Level Affecting Marshes Model (SLAMM) (Warren Pinnacle 2016). SLAMM projects future locations of wetlands based on relative elevation changes due to SLR. SLAMM has built-in parameters that include the soil accretion rates with time, which differentiates it from a simple bathtub model. As to the parameters, a comparison was made with the three SLAMM runs of proximity to the study region by Warren Pinnacle: Archie Carr NWR (2010a), Merritt Island NWR (2011), and Pelican Island (2010b). For this study, SLAMM was run for one SLR curve: 1.5 meters (>5 feet) by 2100. Finally, the modeled run was done without protecting the developed land, which means that what Langridge et al. (2014) called “natural infrastructure” was modeled to overtake the developed land. It is these areas where the developed land is overcome that was used for selecting out the mitigation areas from the previously determined vulnerable areas. Finally, the resultant shapefiles from SLAMM was overlayed with the results from Hazus-MH to select the areas which are modeled to have high potential for floodplain conservation out from the identified vulnerable areas. This is accomplished is by tallying the area per vulnerable U.S. Census block that is predicted by SLAMM to go from a “dry” ecosystem to a “wet” ecosystem. The more area that is predicted to turn into a “wet” system is considered to be an area where the floodplain mitigation strategy would be an optimized implementation. In this way, the vulnerable areas were identified and then can be prioritized for the coastal hazard mitigation strategy by local planners. 7 With this overview in mind, the systematic process of running the models are detailed step-by-step. Hazus-MH For this study, the required materials to run Hazus-MH entails having ArcGIS 10.2.2 and Hazus-MH 3.1. In the processing of running the model, the Digital Elevation Model (DEM) is acquired through the Hazus-MH link to the National Elevation Dataset that is put out by the United States Geological Survey (2013). In this study, the DEMs that were gathered were for the following coordinates: 28N 081W, 28N 082W, 29N 081W, 29N 082W, 30N 081 W, and 30 N 082W. 1. Create a region and name it accordingly. 2. Choose flood hazard for the census tract level of analysis. Then select the state of Florida, Volusia County, all the census tracts, all the block groups, and then finally all the blocks. After this is completed, then click the finish button to create the region follow through with the prompt to switch to open a region. 3. Select open a region and then choose the name of the region that was just created. Next click through to accept the choosing opening the flood model the dialog box. 4. In the flood model, click the hazard tab and then choose the flood hazard type. This will open a dialog box where there will be an option for coastal only. This is the model that will be run, so choose the option and then continue through the dialog box until it closes. 5. Then open the hazard tab again and click on user data. It is in this dialog box that the extent of the region in terms of the Digital Elevation Model (DEM) is determined and is retrieved, by clicking on the determine the extent button. 6. After the extent is determined, there is an option to download the required DEMs. Once the DEMs are downloaded, they need to be stitched together. To do that requires opening ArcMap and under the toolbox find the Data management toolbox to the Raster category and down to the Raster dataset to open the Mosaic to New Raster tool. 8 7. In the Mosaic to New Raster tool some data is required informationally as to the type of rasters are going into the tool, such as the pixel type (32-bit float), cell size (0.0027777778), number of bands (1), and finally where the data will be stored and under what name. Once these DEMs have been stitched together, it is recommended to start a new region and redo steps 1 through 7 instead of picking up where the model left off in step 8. This is due to helping the model run consistently all the way through to avoid any hidden errors. 8. At this point, in the new study region, when it comes to the User Data dialog box under the Hazard Tab, instead of determining the extent of the study region to obtain the DEM, browse to the mosaic DEM that was geoprocessed in ArcMap during step 10. Then initiate Raster Preprocessing and let it run until it finishes. 9. Click on the Hazard tab again and then on the Scenario and slide over to click on New to creating a new scenario and name it accordingly. With the dialog boxes that appear, click the select all when it asks to add to selection and then save the selection. Use the default Startline and for the Volusia County run, the still water elevation was set to be 11 feet (NAVD88). This was chosen after consulting FEMA’s Flood Insurance Rate Map. After this click on finish and save the map file. 10. Once the new scenario has been created, click on Hazard scroll down to Coastal and then choose to delineate the floodplain. Afterwards save the study region, close Hazus completely, and reopen the study region. 11. After the study region is open again, click on the Analysis tab and then Run. In the dialog box choose to select all and then unselect the what-if and the indirect economic cost options to improve the run-time efficiency. After setting up the model options as desired, run the analysis. At one point the model will prompt what date would the flood take place and for this study the date September 15th was chosen because its placement during hurricane season. 12. At this point the Hazus-MH model has created results. The resulting census tracts will then be saved as a set of shapefiles for later comparison. 9 SLAMM The required materials utilized for the part of this study include SLAMM 6.2 and ArcMap 10.2.2 or 10.4. Additionally, the only data piece that was not previously obtained that is required to run SLAMM is a landcover dataset (Hazus-MH previously produced the DEM). For this study, the landcover dataset was obtained from the Florida Fish and Wildlife Conservation Commission and Florida Natural Areas Inventory (2016). Lastly, the parameters need to be developed for SLAMM. These include the NWI Photo Date, DEM date, direction offshore, historic trend, erosion rates, accretion rates, frequent over wash, whether to use elevation pre-processor values, salt elevation, GT Great Diurnal Tide Range, and the MTL-NAVD88 conversion factor. These were determined by comparing the established Warren Pinnacle reports for the Archie Carr NWR (2010a), Merritt Island NWR (2011), and Pelican Island (2010b) and can be found in Table 1. Additionally, while within SLAMM itself, click on the save simulation button as often as possible. 1. Open ArcMap and add in the landcover dataset. This landcover dataset must be crosswalked, or recategorized, into the same categories that SLAMM which can be found in Table 2. The values that correspond to the various landcover classifications will be defined to be equal and will be transferred over by means of the Reclassification tool. If there is a projection mismatch with the DEM, this would also be the time to adjust it so that all the data matches. For the classifications used in this study, please view Appendix A. 2. In ArcMap, bring in the DEM that was previously mosaiced together and using the Slope tool, create a slope layer. 3. Next entails the preprocessing of making sure that the Slope, DEM, and the crosswalked landcover datasets have the same exact cell size and alignment. Once this is done, the three aligned layers are converted to ASCII format because it is compatible with SLAMM. 4. With all the data in hand, startup SLAMM and then click on new simulation. Name the simulation accordingly and click on file setup to put in the Slope, DEM, and landcover ASCII 10 files by means of the browse button. To make sure that SLAMM accepts the files, click on Re-check Files’ Validity. Click the Save Simulation button and then OK. 5. In a similar fashion to the file setup, the site parameters need to be imputed by clicking on the Site Parameters and then typing in the parameters into their appropriate boxes. When finished, click Save Simulation and OK. 6. Once everything is set up, then it is on to the execute button and inserting the actual model scenarios. This is where the 1.5 meters by 2100 along with the year output 22100 are inputted. Also, it is important to make sure that the Don’t Protect is checked to make sure that the environment can take over developed land. When finished setting up the scenario, click the save simulation button one more time and then execute. The output files will be saved for further research purposes. The Overlay To perform the overlay, from which the final product is obtained, some of the results from SLAMM and Hazus-MH are required. From Hazus-MH, the U.S. Census blocks that are characterized by substantial damage need to be exported and the SLAMM outputs for the initial and 2100 projection of the landcover for the study region are required. The data should all be put into the same projection and opened in ArcMap. 1. Both the initial condition landcover and the 2100 projected landcover need to be reclassified into four categories which is detailed in Table 3: (1) developed dry, (2) undeveloped dry, (3) water saturated environments, and (4) open water. These categories for simplicity of communication can also be referenced as “dry” and “wet” environments. 2. Now that the initial condition and the 2100 projected condition have been classified into the four categories, they each need to be reclassified in a binary fashion such that the desired landcover type is equal to 1 and the other three categories are equal to 0. For instance, to isolate the developed land, the value used for developed land in the reclassification is equal to 1 and the values for the other three categories would be 0. To keep each of the classifications 11 separate, a naming scheme such as yinitial_1 was employed to designate that it was the isolated developed dry land (which has the value of 1) and it was for initial condition year. Similarly, y100_4 would designate the open water (which identification characteristic was assigned to be 4) for the year 2100. 3. Use the Times tool to isolate the location where the following landcover changes occurred: a. Initial landcover of developed dry land with 2100 landcover of open water b. Initial landcover of developed dry land with 2100 landcover of water saturated environments c. Initial landcover of undeveloped dry land with 2100 landcover of open water d. Initial landcover of undeveloped dry land with 2100 landcover of water saturated environments For each of the outputs, a naming scheme was utilized in the fashion such as Times_1_4 to designate the selection of area designated by the cross between developed dry land under the initial condition (1) with the 2100 landcover of open water (4). 4. To assign the areal designation of the output from the Times tool to each of the vulnerable U.S. Census Blocks, the Zonal Statistics as Table tool was utilized with the input feature zone data being the Building Damage U.S. Census Blocks and the input value raster being the results from the Times tool. This was repeated for all four of the Times tool raster outputs and resulted with four tables. 5. Join the Zonal output tables to the U.S. Census Blocks of the Building Damage and by adding a field (double format) and field calculator, the values from the tables can be saved to the U.S. Census Blocks. Furthermore, a field was added and the field calculator was used to tally the amount of “dry” environments that were changed into “wet” environments. Other results were gathered and tallied using the statistics button for a field in the attribute table, or the means of a calculator. 12 Results The premise of the methodology is that the areas where Hazus-MH deemed to be vulnerable and the identified census block is also designated by SLAMM to be converting to water based ecosystems, those are places to consider implementation of floodplain mitigation. There is a progression of information that follows from the methodology, the first being a map of the flood inundation itself, which can be found in Figure 1. From the modeled flood depth, the vulnerable U.S. Census Blocks which were selected and are depicted in Figure 2 according to the number of substantially damaged buildings. After the data outputs from SLAMM were tabulated per U.S. Census Block, the result can be found in Figure 3, which is the map of the percent landcover change from “dry” to “wet” for each vulnerable census block. For this study on the Upper Indian River Lagoon and Atlantic Coast of Volusia County, Florida, U.S.A., there were a total of 13,539 buildings of the vulnerable region’s 33,024 buildings (<41%) that were classified as having substantial damage (more than 50 percent damaged) in a 100 year flooding scenario. Of these, 11,234 buildings were more than 90 percent damaged. Of the substantially damaged buildings, 774 buildings are in U.S. Census Blocks were there is over 80 percent landcover change. Additionally, the tabulation of the U.S. Census Blocks per the landcover change is in Table 4. The sample section, as can be seen in Figure 4, demonstrates the impact of the overlay between the landcover change percentages with the number of substantially damaged buildings. Discussion and Conclusions There are several strengths to the methodology as it stands. Hazus-MH and SLAMM are not simple bath-tub models as they consider more factors than just elevation (Federal Emergency Management Agency n.d.; Warren Pinnacle 2016). Furthermore, utilizing SLAMM to model the landcover change provides the cross of a long-term coastal hazard, namely SLR, to overlay with the current assessment of the 100 year flood plain as determined by Hazus-MH. As the methodology was executed in this study, however, there is room for improvement. A major improvement is that the parameters to run SLAMM should be backed with stronger literature background and field studies 13 (Warren Pinnacle 2016). For instance, to come up with the parameters for the soil accretion rates, data was compared between the Georgia shoreline and the Florida Gulf Coast shoreline and not the Atlantic Coast of Florida (Warren Pinnacle 2010a; Warren Pinnacle 2011; Warren Pinnacle 2010b). This study did find a paper by Parkinson et al. (2006) in which the accretion rates of what Parkinson et al. referenced originally as wetlands but classified by this study as a regular flooded marsh as 3.2 mm/year does compare closely to the 3.9 mm/year value as assigned by Warren Pinnacle (2011). Nevertheless, a field study would be highly encouraged as to verify the actual parameters when the proposed methodology is put into practice. It is the proposition of this methodology that the places which are modeled for “wet” ecosystems are taking over “dry” ecosystems could be more successful, and therefore are prioritized for the mitigation strategy to be implemented. Instead of applying the strategy in all the areas where vulnerability assessments deem the population to be vulnerable, the combination of the landcover change analysis with the results of a vulnerability assessment provides selective assessment. Additionally, the auxiliary information that comes with Hazus-MH helps in the decision as to which prioritized area to implement the adaptation strategy, such as the types and amounts of vulnerable structures and industrial versus residential vulnerabilities. Similarly, SLAMM can produce the specific types of landcovers at designated timesteps to break down the timing of the landcover change. Furthermore, the landcovers as are modeled by SLAMM are not the simplistic developed dry land, undeveloped dry land, moist ground ecosystems, and open water categorizations but rather over 20 different categories. The specific landcover categories could be taken by a team of landscape ecologists, planners, and other informed knowledge base holders as to what kinds of plants should be plant in certain locations. As to the implementation of the strategy itself, the acting governmental authority would still need to decide where to start. Several prioritizations can come from this, and the decision-making process is best accompanied by the sample area that can be found in Figure 4. Questions such as whether implementation should start where the most structures or fewest vulnerable structures need to be considered carefully by the acting agency and it related closely to a cost-benefit analysis. For instance, if 14 the ecosystem restoration as a mitigation strategy was implemented where there were the most structures, then it would be anticipated that there would be the most buildings displaced, the most potential damage prevented, and a potential for ecosystems inserted in more densely populated areas. Conversely, if the strategy was implemented where there were fewer buildings, then the building counts of prevented damages would be lower along with the total prevented damages. A major benefit to implementing the strategy where there are fewer buildings is the potential for greater connectivity to currently established ecosystems. It is the suggestion of this study that the vulnerable houses are not simply removed to put in the ecosystem infrastructure, but rather as the vulnerable areas become available for the local jurisdiction to obtain the land, then the government can implement the adaptive strategies knowledgably. Also, it is not the suggestion of the study that the floodplain mitigation strategy is the only method suitable for adaptation, but rather it is the aim of this study to find a methodology to allow for the floodplain mitigation strategy to become a feasible option of consideration. This is largely due to the selective method of area prioritization includes the premise that the strategy is not applied everywhere, but rather where it is modeled to be most suited. 15 Tables and Figures Parameter Global Description Atlantic NWI Photo Date (YYYY) 2016 DEM Date (YYYY) 2015 Direction Offshore [n,s,e,w] East Historic Trend (mm/yr) 2.37 MTL-NAVD88 (m) -0.26 GT Great Diurnal Tide Range (m) 0.44 Salt Elev. (m above MTL) 0.37 Marsh Erosion (horz. m / yr) 1.8 Swamp Erosion (horz. m / yr) 1 T.Flat Erosion (horz. m / yr) 0.5 Reg.-Flod Marsh Accr (mm/yr) 3.9 Irreg.-Flood Marsh Accr (mm/yr) 4.7 Tidal-Fresh Marsh Accr (mm/yr) 5.9 Inland-Fresh Marsh Accr (mm/yr) 5.9 Mangrove Accr (mm/yr) 7 Tidal Swamp Accr (mm/yr) 1.1 Swamp Accretion (mm/yr) 0.3 Beach Sed. Rate (mm/yr) 0.5 Freq. Overwash (years) 25 Use Elev Pre-processor [True,False] FALSE Table 1: These are the parameters that were used for SLAMM in this study that were based on studies by Warren Pinnacle (2010a; 2010b; 2011). 16 SLAMM Identification Code SLAMM Landcover Category 1 Developed Dry Land 2 Undeveloped Dry Land 3 Swamp (Nontidal) 4 Cypress Swamp 5 Inland-Fresh Marsh 6 Tidal-Fresh Marsh 7 Trans. Salt Marsh 8 Regularly-Flooded Marsh 9 Mangrove 10 Estuarine Beach 11 Tidal Flat 12 Ocean Beach 13 Ocean Flat 14 Rocky Intertidal 15 Inland Open Water 16 Riverine Tidal 17 Estuarine Open Water 18 Tidal Creek 19 Open Ocean 20 Irreg.-Flooded Marsh 21 Not Used 22 Inland Shore 23 Tidal Swamp Table 2: This is the reclassification table for the landcover categories per the SLAMM classification that is necessary for the crosswalk (Warren Pinnacle 2016). 17 SLAMM Identification Code SLAMM Landcover Category Identification Value Assessment Category Name 1 Developed Dry Land 1 Developed Dry Land 2 Undeveloped Dry Land 2 Undeveloped Dry Land 3 Swamp (Nontidal) 3 Water Saturated Environments 4 Cypress Swamp 3 Water Saturated Environments 5 Inland-Fresh Marsh 3 Water Saturated Environments 6 Tidal-Fresh Marsh 3 Water Saturated Environments 7 Trans. Salt Marsh 3 Water Saturated Environments 8 Regularly-Flooded Marsh 3 Water Saturated Environments 9 Mangrove 3 Water Saturated Environments 10 Estuarine Beach 3 Water Saturated Environments 11 Tidal Flat 3 Water Saturated Environments 12 Ocean Beach 3 Water Saturated Environments 13 Ocean Flat 3 Water Saturated Environments 14 Rocky Intertidal 3 Water Saturated Environments 15 Inland Open Water 4 Open Water 16 Riverine Tidal 4 Open Water 17 Estuarine Open Water 4 Open Water 18 Tidal Creek 4 Open Water 19 Open Ocean 4 Open Water 20 Irreg.-Flooded Marsh 3 Water Saturated Environments 21 Not Used 0 Not Used 22 Inland Shore 3 Water Saturated Environments 23 Tidal Swamp 3 Water Saturated Environments Table 3: This is the crosswalk for the reclassification of the SLAMM results to the simple classifications for the selective identification diagnostic of the areas. 18 Figure 1: Hazus-MH depth damage grid output for the 100 year flood return model of Volusia County, Florida. 19 Figure 2: The modeled U.S. Census Blocks that are considered vulnerable to the 100 year flood return period as assessed by Hazus-MH. Hazus-MH Flood Model of the 100 Year Flood Return Period 20 Figure 3: The vulnerable U.S. Census Blocks that are prioritized for the floodplain mitigation strategy by the amount of “dry” to “wet” landcover change. 21 Percent of Landcover Change Number of Vulnerable Blocks Number of Substantially Damaged Buildings Total Number of Vulnerable Buildings 0 - 20 % 3014 5174 12607 20 - 40 % 298 2395 4131 40 - 60 % 178 1494 2316 60 - 80 % 147 909 1381 80 - 100 % 143 682 1038 100% 53 92 141 Table 4: Tabulated results from the final overlay of the SLAMM with the Hazus-MH results. 22 Figure 4: Sample location of the vulnerable U.S. Census Blocks with the amount of substantially damaged buildings and the landcover change prioritization overlay. 23 Appendix A The landcover dataset that was acquired for the study region is from the Florida Cooperative Landcover Classification Map project and was then reclassified (Florida Fish and Wildlife Conservation Commission and Florida Natural Areas Inventory 2016). FWC and FNAI Cooperative Landcover Codes SLAMM Codes CLC Landcover Codes Landcover Classification Sourced from Glazer (2013) Assigned in this study Final SLAMM Code 1110 Upland Hardwood Forest 2 2 1111 Dry Upland Hardwood Forest 2 2 1112 Mixed Hardwoods 2 2 1120 Mesic Hammock 2 2 1122 Prairie Mesic Hammock 2 2 1123 Live Oak 2 2 1124 Pine - Mesic Oak 2 2 1125 Cabbage Palm 2 2 1130 Rockland Hammock 2 2 1150 Xeric Hammock 2 2 1210 Scrub 2 2 1211 Oak Scrub 2 2 1212 Rosemary Scrub 2 2 1213 Sand Pine Scrub 2 2 1214 Coastal Scrub 2 2 1220 Upland Mixed Woodland 2 2 1230 Upland Coniferous 2 2 1240 Sandhill 2 2 1300 Pine Flatwood and Dry Prairie 2 2 1310 Dry Flatwoods 2 2 1311 Mesic Flatwoods 2 2 1312 Scrubby Flatwoods 2 2 1330 Dry Prairie 2 2 1340 Palmetto Prairie 2 2 1400 Mixed Hardwood-Coniferous 2 2 1410 Successional Hardwood Forest 2 2 1500 Shrub and Brushland 2 2 1510 Other Shrubs and Brush 2 2 1600 Coastal Uplands 2 2 24 1610 Beach Dunes 2 2 1630 Coastal Grassland 2 2 1640 Coastal Strand 2 2 1650 Maritime Hammock 2 2 1660 Shell Mound 2 2 1670 Sand Beach (Dry) 12 12 1710 Sinkhole 2 2 1750 Bare Soil 2 2 1760 Exposed Rock 2 2 1800 Cultural - Terrestrial 1 1 1810 Mowed Grass 1 1 1811 Vegetative Berm 1 1 1812 Highway Rights of Way 1 1 1821 Low Intensity Urban 1 1 1822 High Intensity Urban 1 1 1831 Rural Open 2 2 1832 Rural Structures 1 1 1840 Transportation 1 1 1841 Roads 1 1 1842 Rails 1 1 1850 Communication 1 1 1860 Utilities 1 1 1870 Extractive 1 1 1871 Strip Mines 1 1 1872 Sand & Gravel Pits 1 1 1873 Rock Quarries 1 1 1875 Reclaimed Lands 1 1 1876 Abandoned Mining Lands 1 1 1877 Spoil Area 1 1 1880 Bare Soil/Clear Cut 2 2 2100 Freshwater Non-Forested Wetlands 5 5 2110 Wet Prairie 5 5 2112 Mixed Scrub-Shrub Wetlands 3 3 2114 Seepage Slope 5 5 2120 Marshes 5 5 2121 Isolated Freshwater Marsh 5 5 2133 Coastal Interdunal Swale 3 3 2123 Floodplain Marsh 20 20 2124 Slough Marsh 20 20 2125 Glades Marsh 5 5 2131 Sawgrass 5 5 25 2140 Floating/Emergent Aquatic Vegetation 5 5 2141 Slough 15 15 2142 Water Lettuce 15 15 2145 Duck Weed 15 15 2146 Water Lily 15 15 2150 Submergent Aquatic Vegetation 15 15 2200 Freshwater Forested Wetlands 3 3 2210 Cypress/Tupelo(incl Cy/Tu mixed) 4 4 2211 Cypress 4 4 2212 Tupelo 3 3 2213 Isolated Freshwater Swamp 4 4 2214 Strand Swamp 4 4 2215 Floodplain Swamp 4 4 2220 Other Coniferous Wetlands 3 3 2221 Wet Flatwoods 2 2 2222 Pond Pine 3 3 2230 Other Hardwood Wetlands 3 3 2231 Baygall 3 3 2232 Hydric Hammock 2 2 2233 Mixed Wetland Hardwoods 3 3 2240 Other Wetland Forested Mixed 3 3 2241 Cypress/Hardwood Swamps 3 3 2242 Cypress/Pine/Cabbage Palm 3 3 2300 Non-vegetated Wetland 5 5 2400 Cultural - Palustrine 2 2 2410 Impounded Marsh 5 5 2420 Impounded Swamp 3 3 2430 Grazed Wetlands 5 5 2440 Clearcut Wetland 5 5 2450 Wet Coniferous Plantation 2 2 3000 Lacustrine 15 15 3100 Limnetic 15 15 3110 Clastic Upland Lake 15 15 3113 Flatwoods/Prairie/Marsh Lake 15 15 3114 River Floodplain Lake/Swamp Lake 15 15 3115 Sinkhole Lake 15 15 3117 Sandhill Lake 15 15 3118 Major Springs 15 15 3200 Cultural - Lacustrine 15 15 3210 Artificial/Farm Pond 15 15 3211 Aquacultural Ponds 15 15 26 3220 Artificial Impoundment/Reservoir 15 15 3230 Quarry Pond 15 15 3240 Stormwater Treatment Areas 1 1 3260 Industrial Cooling Pond 1 1 4000 Riverine 18 18 4100 Natural Rivers and Streams 18 18 4120 Blackwater Stream 18 18 4130 Spring-run Stream 18 18 4140 Seepage Stream 18 18 4160 Tidally-influenced Stream 16 16 4170 Riverine Sandbar 10 10 4200 Cultural - Riverine 15 15 4210 Canal 15 15 4220 Ditch/Artificial Intermittent Stream 18 18 5000 Estuarine 17 17 5100 Subtidal 19 19 5200 Intertidal 13 13 5212 Non-vegetated 14 14 5220 Tidal Flat 11 11 5221 Mud 11 11 5222 Sand 11 11 5240 Salt Marsh 8 8 5250 Mangrove Swamp 9 9 5320 Estuarine Artificial Impoundment 8 8 6000 Marine 19 19 7000 Exotic Plants 2 2 7100 Australian Pine 2 2 7200 Melaleuca 2 2 7300 Brazilian Pepper 2 2 7400 Exotic Wetland Hardwoods 3 3 9100 Unconsolidated Substrate 11 11 18211 Urban Open Land 1 1 18212 Residential, Low Density 1 1 18213 Grass 1 1 18214 Trees 1 1 18221 Residential, Med. Density - 2-5 Dwelling Units/AC 1 1 18222 Residential, High Density > 5 Dwelling Units/AC 1 1 18223 Commercial and Services 1 1 18224 Industrial 1 1 18225 Institutional 1 1 18311 Rural Open Forested 2 2 27 18312 Rural Open Pine 2 2 18331 Cropland/Pasture 2 2 18332 Orchards/Groves 2 2 18333 Tree Plantations 2 2 18334 Vineyard and Nurseries 2 2 18335 Other Agriculture 2 2 21112 Cuthroat Seep 5 5 21121 Shrub Bog 3 3 21211 Depression Marsh 5 5 21212 Basin Marsh 5 5 22131 Dome Swamp 5 5 22132 Basin Swamp 3 3 22211 Hydric Pine Flatwoods 3 3 22212 Hydric Pine Savanna 3 3 22311 Bay Swamp 3 3 22312 South Florida Bayhead 3 3 22322 Prairie Hydric Hammock 2 2 22323 Cabbage Palm Hammock 2 2 22331 Bottomland Forest 2 2 22332 Alluvial Forest 3 3 182111 Urban Open Forested 2 2 182112 Urban Open Pine 2 2 182131 Parks and Zoos 1 1 182132 Golf courses 1 1 182133 Ballfields 1 1 182134 Cemeteries 1 1 182135 Community rec. facilities 1 1 183311 Row Crops 2 2 183312 Field Crops 2 2 183313 Improved Pasture 2 2 183314 Unimproved/Woodland Pasture 2 2 183315 Other Open Lands - Rural 2 2 183321 Citrus 2 2 183322 Fruit Orchards 2 2 183323 Pecan 2 2 183324 Fallow Orchards 2 2 183331 Hardwood Plantations 2 2 183332 Coniferous Plantations 2 2 183341 Tree Nurseries 2 2 183342 Sod Farms 2 2 183343 Ornamentals 2 2 28 183344 Vineyards 2 2 183345 Floriculture 2 2 183351 Feeding Operations 1 1 183352 Specialty Farms 2 2 221312 Gum Pond 3 3 222111 Cutthroat Grass Flatwoods 2 2 222112 Cabbage Palm Flatwoods 2 2 183312 Sugarcane 2 2 183315 Fallow Cropland 2 2 29 Works Cited Bosello, F. and E. De Cian. 2014. Climate change, sea level rise, and coastal disasters. A review of modeling practices. Energy Economics 46:593-605. Evans, J., C. Goodison, and P. Zwick. 2015. From the bottom up: Implementing resiliency at the local government level. Florida Sea Grant College Program, Gainesville, Florida, USA. Federal Emergency Management Agency. n. d. Hazus-MH flood model technical manual. Mitigation Division, Department of Homeland Security, Washington D.C., USA. Available from https://www.fema.gov/media-library-data/20130726-1820-250458292/hzmh2_1_fl_tm.pdf. Florida Fish and Wildlife Conservation Commission and Florida Natural Areas Inventory. 2016. Cooperative Land Cover, Version 3.2. Raster. U.S. Fish and Wildlife Service, Tallahassee, Florida. Available from http://myfwc.com/research/gis/applications/articles/Cooperative-Land-Cover. Glazer, R. 2013. Alternative futures under climate change for the Florida Key’s benthic and coral systems. State Wildlife Grant FWC 6242. Florida Fish and Wildlife Conservation Commission, Marathon, Florida, USA. Johnston, A., P. Slovinsky, and K. L. Yates. 2014. Assessing the vulnerability of coastal infrastructure to sea level rise using multi-criteria analysis in Scarborough, Maine (USA). Ocean & Coastal Management 95:176-188. Kousky, C. and M. Walls. 2014. Floodplain conservation as a flood mitigation strategy: Examining costs and benefits. Ecological Economics 104:119-128. Langridge, S. M., E. H. Hartge, R. Clark, K. Arkema, G. M. Vertues, E. E. Prahler, S. Stoner-Duncan, D. L. Revell, M. R. Caldwell, A. D. Guerry, M. Ruckelshaus, A. Abeles, C. Coburn, and K. O’Connor. 2014. Key lessons for incorporating natural infrastructure into regional climate adaptation planning. Ocean & Coastal Management 95:189-197. Mahendra, R. S., P. C. Mohanty, H. Bisoyi, T. S. Kumar, and S. Nayak. 2011. Assessment and management of the coastal multi-hazard vulnerability along the Cuddalore—Villupuram, east coast of India using geospatial techniques. Ocean & Coastal Management 54:302-311. 30 Murali, R. M. and P. K. D. Kumar. 2014. Implications of sea level rise scenarios on land use /land cover classes of the coastal zones of Cochin, India. Journal of Environmental Management 148:124-133. Parkinson, R. W., R. R. DeLaune, C. T. Hutcherson, and J. Stewart. 2006. Turning surface water management and wetland restoration programs with historic sediment accumulation rates: Merritt Island National Wildlife Refuge, East-Central Florida, U.S.A. Journal of Coastal Research 22:1268-1277. Parkinson, R. W., and T. McCue. 2011. Assessing municipal vulnerability to predicted sea level rise: City of Satellite Beach, Florida. Climatic Change 107:203-223. United States Geological Survey. 2013. National Elevation Dataset. Raster. Department of the Interior, Reston, Virginia. Available from: https://viewer.nationalmap.gov/basic/?howTo=true#startUp. Warren Pinnacle. 2010a. Application of the Sea-level Affecting Marshes Model (SLAMM 6) to Archie Carr NWR. U.S. Fish and Wildlife Service, Division of Natural Resources and Conservation Planning, Arlington, Virginia, USA. Available from: http://www.warrenpinnacle.com/prof/SLAMM/USFWS/SLAMM_Archie_Carr.pdf. Warren Pinnacle. 2010b. Application of the Sea-level Affecting Marshes Model (SLAMM 6) to Pelican Island NWR. U.S. Fish and Wildlife Service, Division of Natural Resources and Conservation Planning, Arlington, Virginia, USA. Available from: http://www.warrenpinnacle.com/prof/SLAMM/USFWS/SLAMM_Pelican_Island.pdf. Warren Pinnacle. 2011. Application of the Sea-level Affecting Marshes Model (SLAMM 6) to Merritt Island NWR. U.S. Fish and Wildlife Service, Division of Natural Resources and Conservation Planning, Arlington, Virginia, USA. Available from: http://www.warrenpinnacle.com/prof/SLAMM/USFWS/SLAMM_Merritt_Island_2011.pdf. Warren Pinnacle. 2016. SLAMM 6.7 technical documentation. Warren Pinnacle Consulting, Inc. 31 Available from: http://warrenpinnacle.com/prof/SLAMM6/SLAMM_6.7_Technical_Documentation.pdf.