Incorporating climate change and morphological uncertainty into coastal change hazard assessments
Documented and forecasted trends in rising sea levels and changes in storminess patterns have the potential to increase the frequency, magnitude, and spatial extent of coastal change hazards. To develop realistic adaptation strategies, coastal planners need information about coastal change hazards that recognizes the dynamic temporal and spatial scales of beach morphology, the climate controls on coastal change hazards, and the uncertainties surrounding the drivers and impacts of climate change. We present a probabilistic approach for quantifying and mapping coastal change hazards that incorporates the uncertainty associated with both climate change and morphological variability. To demonstrate the approach, coastal change hazard zones of arbitrary confidence levels are developed for the Tillamook County (State of Oregon, USA) coastline using a suite of simple models and a range of possible climate futures related to wave climate, sea-level rise projections, and the frequency of major El Niño events. Extreme total water levels are more influenced by wave height variability, whereas the magnitude of erosion is more influenced by sea-level rise scenarios. Morphological variability has a stronger influence on the width of coastal hazard zones than the uncertainty associated with the range of climate change scenarios.
KeywordsClimate change El Niño Exposure Increasing storminess Probabilistic coastal hazard zones Sea-level rise Uncertainty
Sea-level rise (SLR), increasing storminess, and development pressures all contribute to making coastal communities vulnerable to coastal change hazards (e.g., Bindoff et al. 2007; Young et al. 2011; Strauss et al. 2012). A significant challenge for coastal planners in reducing this vulnerability is the lack of tools or information that recognize the inherent uncertainty of climate change and its influence on physical processes that shape the coastal zone. Although they are the basis for coastal policies, regulations, and adaptation planning efforts, many coastal change hazard mapping efforts (e.g., Allan and Priest 2001) effectively ignore the non-stationarity of coastal processes, such as accelerating SLR and changes in storminess.
Forecasts of potential impacts to coastal areas due to climate change and/or extreme events have been performed on the US West Coast using qualitative indices (Thieler and Hammar-Klose 1999); quantitative, event-selection or benchmark event approaches (e.g., Allan and Priest 2001); and response-based or structural function approaches (Revell et al. 2011). Fully probabilistic, full temporal simulation approaches that account for conditional dependencies between relevant variables (e.g., Callaghan et al. 2008) are considered the most robust and involve randomly sampling from fitted probability distributions in a Monte Carlo sense. Both the response function and full simulation approaches have the potential to include the non-stationarity associated with climate change.
In addition to the variability in physical processes, community exposure to coastal change hazards varies depending on how communities currently use or plan to use hazard-prone areas. There have been various efforts in recent years to estimate societal exposure to various SLR scenarios throughout the world (e.g., Wu et al. 2009; Strauss et al. 2012), as well as efforts to estimate the influence of SLR in increasing societal exposure to sudden-onset coastal hazards, such as hurricane storm surge (e.g., Frazier et al. 2010). Lacking from these studies are attempts to characterize community vulnerability to coastal change hazards, such as chronic erosion associated with winter storms.
Within this context of hazard uncertainty and societal relevance, the objective of this paper is to summarize a multi-scale, probabilistic methodology that incorporates the impacts of projected climate changes and variability, as well as morphological variability, into coastal change hazard assessments. We first develop a suite of climate change scenarios that reflect various assumptions regarding SLR, storminess, and major El Niño occurrences and their impact on future extreme water levels. Simple models are then used to quantify potential coastal change hazards and generate a series of probabilistic hazard zones for select events over a range of timescales (event to decadal) that account for both climate and morphological uncertainties. We integrate these hazard zones with data on local infrastructure to examine spatial variations in community exposure to coastal change hazards. Finally, we examine the relative contribution of sea level, wave climate, and the frequency of occurrence of major El Niños to the magnitude and spatial extent of coastal change hazard zones.
To demonstrate this integrated approach, we focus on the multiple coastal communities and littoral cells along dune-backed shorelines in Tillamook County on the northern coast of Oregon (USA). Although we focus on one county and its littoral cells, the methodology is applicable to all dune-backed shorelines in the US Pacific Northwest [covering approximately 45 % of the outer coasts of Oregon and Washington (Cooper 1958)], as well as other coastal regions with similar physical settings. Methods presented here further the dialogue on understanding community exposure to coastal change hazards that incorporate climate change uncertainty and can be used by coastal planners in their efforts to balance community growth and increased adaptive capacity to natural hazards over the next several decades.
2 Study area
The dynamic beaches of Tillamook County, OR, are exposed to mesotidal conditions and a relatively extreme wave climate. Monthly mean significant wave heights off the coast of Oregon are on the order of 1.5 m in the summer with periods of ~8 s, while wave heights in the winter are typically double in height, averaging ~3 m with periods on the order of 12–13 s (Ruggiero et al. 2005). It is typical for winter storms to annually generate wave heights >10 m, while some of the strongest storms on record have generated waves up to 14 m (Allan and Komar 2002, 2006). Peak periods associated with these large wave heights tend to be about 15–17 s, but can be as large as 20 s (Allan and Komar 2002). The strong seasonal variability in water levels, ~20 cm (Komar et al. 2011), is in phase with the seasonal variability in the wave climate, resulting in significantly higher total water levels (TWLs) during the winter season. During major El Niño events, water levels, wave heights, and wave direction are all anomalous, resulting in regional and hotspot erosion throughout the region.
3 Climate change scenario development
Climate change scenarios serve as the foundation for the coastal change hazard zones and are based on the projections for SLR, the wave climate, and the probability of occurrence of major El Niño events through the year 2100. Due to the inherent uncertainty in climate projections (e.g., Hemer et al. 2013), our approach here is to broadly explore the range of variability documented in the literature rather than dynamically or statistically downscaling the various processes from models (Wilby and Dessai 2012).
The extreme wave height climatology has been documented to be increasing over the last few decades (Allan and Komar 2000, 2006; Méndez et al. 2006; Menéndez et al. 2008; Ruggiero et al. 2010; Young et al. 2011), and this increase has been potentially more responsible than changes in sea level alone for increasing the frequency of extreme events along the PNW coast (Ruggiero 2013). Moreover, wave heights of different magnitudes (i.e., exceedance percentiles) are increasing at different rates such that larger waves are getting bigger faster (Ruggiero et al. 2010). Therefore, wave heights in our future wave climate scenarios were allowed to change by quartile, depending on their exceedance percentile of the wave height cumulative distribution function. We developed three simple future wave climate scenarios: (1) wave heights increase at their present rate of 1.0, 1.5, 2.0, or 4.0 cm/year, depending on quartile, as documented by Ruggiero et al. (2010) until 2030; (2) wave heights remain at their present levels indefinitely; and (3) wave heights decrease until the year 2030 at the same rate as the wave height increase scenario (Fig. 2b). We extend our wave scenarios only for two decades due to the lack of a detailed understanding of the causation of the existing trends, and the uncertainty associated with downscaling winds and waves from model projections (e.g., Hemer et al. 2013).
Major El Niños significantly affect both water levels and wave heights in the PNW and have been associated with the most severe erosion and flooding hazards in the region documented over the last few decades (Kaminsky et al. 1998). Over the period of wave buoy measurements, there have been two major El Niños—one in 1982–1983 and another in 1997–1998. Because little is certain about how El Niño frequency will change in the future, we include one scenario in which the frequency of occurrence of major El Niños doubles to approximately two per 15 years. Recent modeling work (Cai et al. 2014) has suggested that a scenario of increased frequency of major El Niños due to greenhouse warming is plausible. The additional El Niños are added by increasing water levels and wave heights over the course of entire winters, to match the scale of the event during the winter of 1997–1998. For comparison, we also develop a scenario in which the frequency of El Niños remains the same as during the last several decades. Combining these SLR, wave climate, and El Niño projections, we generate 20 climate change scenarios, covering a broad range of climate uncertainty (Fig. 2c). In this application, we assume that each climate change scenario has an equal probability of occurrence (i.e., a rectangular probability density function); however, it is straightforward to assess the impacts of an imposed occurrence probability distribution function if warranted.
4 Defining coastal change hazard zones
To statistically determine extreme TWLs for the annual and 100-year return level at each time period of interest, we use the peak-over-threshold method of extreme value theory (Coles 2001) that requires a high threshold and models exceedances over this threshold. We assume that the number of exceedances in a given year follows a Poisson distribution and that the threshold excesses are modeled using the generalized Pareto distribution. The threshold value is set such that, on average, five TWL events per year are analyzed, a value equal to about the 97–98th exceedance percentile of the overall TWL climatology. While this application of extreme value theory assumes stationarity in the TWL time series, climate change-induced trends in the TWL are accounted for by calculating extreme values separately for each time period of interest.
5 Morphometric and hydrodynamic inputs
Shoreface slopes, tan βsf, which are inputs to Bruun Rule calculations, were estimated by calculating the slope between the approximate multidecadal depth of closure and the MHW shoreline (2.1 m contour, relative to NAVD88), derived from lidar data. Nearshore bathymetry was extracted at 500-m alongshore intervals from a 6 arc-second digital elevation map (NOAA Center for Tsunami Research 2004). The 20-m isobath was chosen for estimating shoreface slopes; however, results were fairly insensitive to a range of reasonable values (15–25 m). The 500-m resolution of alongshore varying shoreface slopes was interpolated to match the more highly resolved- and variable-beach morphometrics. The remaining morphometric parameters were derived from high-resolution lidar data collected in September 2002 (NOAA Coastal Services Center 2002) and are summarized in Mull and Ruggiero (2014), where dune toe elevation and backshore beach slope were defined for thousands of cross-shore profiles every few meters along the southwest Washington and Oregon coastlines.
We account for beach and dune morphometric variability by randomly sampling dune toe elevations and beach slopes from distributions. Using the highly resolved spatial variability in the lidar dataset as a proxy for temporal variability [following the approach of Ruggiero and List (2009)], the backshore beach slope is allowed to vary by randomly sampling from a normal distribution defined by the median and standard deviation of 1-km shoreline segments in each littoral cell. Beach level variations due to rip current embayments are indirectly accounted for by incorporating the spatial variability within a 1-km stretch of shoreline. Dune toe elevations are normally distributed where the mean is the lidar-derived value, and the standard deviation is defined by the vertical error associated with its selection and interpolation, mean total RMSE = 0.66 m (Mull and Ruggiero 2014). One hundred random dune toe and beach slope configurations were combined with each of the 20 climate change scenarios, resulting in 2,000 computations of storm-induced coastal change at each of the lidar-derived profiles.
Mean and standard deviations (SD) of beach and foredune parameters for the littoral cells of Tillamook County, Oregon, relative to NAVD88, as extracted from 2002 lidar data (Mull and Ruggiero 2014)
Dune toe elev. (m)
Dune crest elev. (m)
After characterizing the local morphology, wave and water-level time series are generated so that extreme TWLs can be calculated as input conditions to the coastal change models. A combined time series is first produced that synthesizes data from wave buoys in the region (NDBC 2007) to develop as complete a significant wave height and spectral peak period record as possible (following Allan et al. 2012). The wave record is then joined with the Newport, Oregon tide gauge record, which dates back to the mid-1960s (NOAA 2007). Following the methods described in Ruggiero et al. (2001), we generate an hourly TWL time series by combining hourly estimates of run-up with hourly tide gauge measurements (Eq. 3). Since the wave run-up model depends on local beach slope, separate TWL time series are developed for all possible beach slopes relevant to our study region. For our north-central Oregon coast study site, the joint time series extends from 1976 to 2009 and is approximately 85 % complete due to data gaps.
Extreme value theory is then applied to the joint time series to develop design conditions relevant for present-day conditions, i.e., the annual and 100-year return level TWL. By definition, the annual event has on a 100 % chance of occurring every year, while the 100-year event, representing a much larger, and more rare, storm with a higher TWL, has a 1 % chance of occurring every year. We construct 20 projected TWL time series (per representative beach slope)—one for each of the 20 climate change scenarios. The projected TWL time series are extended through the year 2100 by repeating the observed time series approximately three times and incorporating non-stationarity by applying the climate change scenarios to the wave height and water-level components of the TWL. These synthetic TWL time series allow us to compute extreme design conditions at any time over the next century (e.g., 2030, 2050, and 2100) for each scenario, covering a broad range of climate change variability.
Because the extreme TWLs vary smoothly across the range of beach slopes, lookup tables of the extreme values through time are developed to interpolate extreme TWL values for any slope. For the 20 climate change scenarios and 100 combinations of slope and dune toe parameters, 2,000 sets of extreme TWLs for each event at each time period of interest are calculated by interpolating from the lookup tables. The resulting extreme TWLs were used in conjunction with the derived morphometrics to quantify future coastal change hazards.
6 Magnitude of coastal change
Though it varies slightly between littoral cells, the difference in magnitude of total coastal change forecasted between the annual and 100-year TWL events for any time period of interest is ~17 m (Eq. 6). The Neskowin littoral cell is forecasted to have the least amount of shoreline retreat due to increasing sea levels, as it has the steepest shoreface slopes (tan βsf) of the county (Eq. 2). However, because there are many areas of relatively flat beaches backed by foredunes with low dune toes, it is estimated to experience the greatest amount of event-based coastal dune erosion throughout the century.
7 Coastal change hazard zones
The raw coastal change estimates predicted from the simple models are highly variable in the alongshore due to the high-resolution of the lidar dataset from which the calculations are based. Presenting the data in this way could exhibit too much variation for coastal planning purposes; therefore, the alongshore varying coastal change lines are smoothed with a quadratic loess filter to eliminate variability less than 250 m (Plant et al. 2002).
In addition to the increase in magnitude of the total coastal change hazards throughout the century, there is also an increase in the uncertainty associated with the calculations of expected coastal change. This increase has direct implications for the estimated impacts to coastal communities and shoreline properties because the hazard zones are ultimately defined based on these statistics. From the smoothed, alongshore varying coastal change distances for the mean and 95 % confidence interval, a suite of coastal change hazard zones were defined based on the probability of retreat exceeding a certain distance. In other words, there is a 98 % probability that the actual retreat will be greater than the mean −2σ, a 50 % probability it will be greater than the mean, and only a 2 % probability that the magnitude of retreat will be greater than the mean +2σ (Fig. 8). The two landward most zones, therefore, compose the 95 % confidence interval within which the retreat for a given year and TWL event is most likely to achieve. The seaward edge of the 98 % exceedance probability zone is the alongshore-smoothed position of the 2002 lidar-derived dune toe.
The zones delineating various probabilities of coastal change exceedance provide decision makers with the ability to select probabilities that best match their risk tolerance. For example, an emergency manager may have a low-risk tolerance for life loss or other impacts to individuals and therefore use the most conservative zone delineating the 2 % exceedance. In contrast, a public works official deciding on the placement or upkeep of coastal infrastructure may have a higher risk tolerance than the emergency manager and may therefore be content using only the 50 % exceedance zone.
With these considerations in mind, coastal change hazard zones were developed for all dune-backed beaches of the four littoral cells in Tillamook County, for the annual and 100-year return level TWL events in 2009, 2030, 2050, and 2100. Mapped coastal change hazard zones take into account local geology (i.e., retreat was not allowed to cut through bluffs and erosion-resistant uplands), but do not take into account existing shore protection structures (e.g., riprap). Both the magnitude of predicted coastal change and the width of the hazard zones increase through time as the hazard become more severe and the uncertainty increases.
8 Variations in community exposure
The Rockaway littoral cell contains the greatest number of structures that are in coastal change hazard zone, both today and in the future. In 2009, there were 281 and 534 structures in coastal change hazard zones associated with the annual and 100-year TWL events, respectively. By the year 2050, the number of potentially exposed structures are more than doubles to 536 (793) structures due to the annual (100-year) TWL event and then to 968 (1116) structures by 2100. The Neskowin and Sand Lake littoral cells have far fewer structures within the hazard zones; however, the communities in these cells are much smaller than the City of Rockaway Beach and may sustain more significant losses as a whole. The fewest impacts to structures will be in the Netarts littoral cell because there is no coastal development along Netarts Spit, except for campground facilities in Cape Lookout State Park.
Trends of increased exposure through time will be a common theme throughout coastal areas, especially those that are heavily developed close to the shoreline. As the amount of forecasted coastal change increases so will the impact to adjacent infrastructure, depending on the community’s spatial distribution and density. The relative impact of these hazards will depend on the community’s size and resilience, with smaller communities potentially being more sensitive to coastal infrastructure damage or loss. Methods outlined here to characterize the magnitude, uncertainty, and spatial variability of coastal hazard zones, as well as potential changes in structures exposure to these hazards, provide land-use planners and emergency managers with insight on potential changes in and implications of societal vulnerability over time. This information can serve as a critical input in the development of land-use scenarios, comprehensive land-use plans, resource management plans, and capital-improvement plans that each seeks to balance community development with long-term sustainability and resilience.
9 Influence of model components on coastal change hazards
We first assessed the relative contribution of sea level, wave climate, and the frequency of occurrence of major El Niños to the hazard zones by isolating the impact of the 20 individual scenarios using extreme TWLs in the Neskowin littoral cell in 2050. Standard deviations for each of the scenarios are small (less than ~0.5 m), implying that there is relatively minimal variability in the TWLs that result from the 100 different morphological configurations. We conducted a sensitivity analysis by investigating the relative impact of each one of the climate controls on the extreme TWLs by holding two of the three climate controls constant and comparing the extreme TWL values between the low and high projections of the third component. Based on the differences for all sets of projections and averaging the results, variability in the TWL based on the wave height scenarios has more of an impact on the extreme TWLs (0.4 m) than does the variability in the SLR scenarios (0.2 m) or the El Niño scenarios (<0.1 m).
Relative impact of SLR, wave climate, and El Niño scenario ranges alone, with the other two components held constant, on the total predicted coastal change, expressed in terms of magnitude and percent increase (%) from the lowest to highest projection
El Niño frequency
100 % annual chance
1 % annual chance
For example, scenarios involving the high SLR projection result in 15.9 m of additional erosion, on average, associated with the 100-year TWL event in 2050 over those involving the low SLR projection (Eq. 1). This corresponds to a 25 % increase in the magnitude of retreat between scenarios involving the low versus high SLR projections when wave height and El Niño scenarios are held constant (Table 2). For the same year and TWL event, the total coastline retreat predicted for scenarios involving the high wave climate projection, where wave heights are allowed to increase through the year 2030, is only 7.5 m greater (equivalent to a 10 % increase) than the scenarios with the low projection where wave heights decrease over the same time period. Because the total predicted retreat increases significantly between 2050 and 2100 but the range in the wave height scenarios stays relatively the same (since wave heights were only allowed to increase until 2030), the relative percentage decreases over time.
Doubling the frequency of El Niño events in the future, however, does not have a significant impact on the width of the hazard zones. The magnitude of erosion predicted for the 100-year return TWL event is insensitive to a doubling in the frequency of El Niño, and the annual return TWL event only increases the total predicted retreat by 2 %. These results may be due to the fact that while this scenario brings the total number of major El Niños to four every 30 years, it modifies waves and water levels for only a few months approximately every 15 years. Therefore, the TWLs between the climate change scenarios with and without a doubling of El Niños are quite similar.
Our probabilistic method of quantifying and mapping coastal change hazards for a variety of dune-backed environments incorporates a range of projections for SLR, wave climate, and El Niño frequency, recognizes the uncertainty and variability associated with the impact of future climate change on extreme water levels and coastal change, and accounts for morphological variability. Rather than imposing one set of hazard zones on a local government, the approach presented here allows for decision makers to choose the level of risk they are willing to accept as well as what time period and design event for which they would like to develop adaptation plans. Results demonstrate that, despite significant uncertainties, coastal change hazards are likely to increase in the future and that community exposure to these hazards varies substantially. In order to most accurately portray the potential impacts of climate change on coastlines, one must consider not only SLR but also changes in storminess which can play a large role in the extreme water levels experienced at the shoreline and can cause increased episodic erosion of protective foredunes.
The authors gratefully acknowledge the support of NOAA’s Climate Program Office Sectoral Applications Research Program (SARP) under NOAA Grant #NA08OAR4310693, NOAA’s Climate Program Office Coastal and Ocean Climate Applications Program under NOAA Grant #NA12OAR4310109, NOAA’s National Sea Grant College Program under NOAA Grant #NA06OAR4170010, the Climate Impacts Research Consortium (CIRC) funded under NOAA Grant #NA10OAR4310218, and the US Geological Survey Land Change Science Program.
- Allan JC, Komar PD (2002) Extreme storms on the Pacific Northwest coast during the 1997–98 El Niño and 1998–99 La Niña. J Coast Res 18(1):175–193Google Scholar
- Allan JC, Priest GR (2001) Evaluation of coastal erosion hazard zones along dune and bluff backed shorelines in Tillamook County, Oregon: Cascade Head to Cape Falcon Rep. Open File Report O-01-03, Oregon Department of Geology and Mineral Industries, Newport, ORGoogle Scholar
- Allan JC, Ruggiero P, Roberts JT (2012) Coastal flood insurance study, Coos County, Oregon. Oregon Department of Geology and Mineral Industries Special Paper 44, 119 ppGoogle Scholar
- Bindoff NL, Willebrand J, Artale V, Cazenave A, Gregory J, Gulev S, Hanawa K, Le Q, Levitus S, Nojiri Y, Shum CK, Talley LD, Unnikrishnan A (2007) Observations: oceanic climate change and sea level. In: Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL (eds) Climate change 2007: the physical science basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate. Cambridge University Press, CambridgeGoogle Scholar
- Bruun P (1962) Sea-level rise as a cause of shore erosion. J Waterw Harbors Div 88:117–130Google Scholar
- Cooper WS (1958) Coastal Sand Dunes of Oregon and Washington. Memoirs of the Geological Society of America (Boulder, Colorado). Memoir 72, pp 1–6Google Scholar
- Kaminsky GM, Ruggiero P, Gelfenbaum G (1998) Monitoring coastal change in southwest Washington and northwest Oregon during the 1997/98 El Niño. Shore Beach 66(3):42–51Google Scholar
- Komar PD, McDougall WG, Marra JJ, Ruggiero P (1999) The rational analysis of setback distances: applications to the Oregon coast. Shore Beach 67(1):41–49Google Scholar
- Mull J, Ruggiero P (2014) Estimating storm-induced dune erosion and overtopping along U.S. West Coast beaches. J Coast Res. doi:10.2112/JCOASTRES-D-13-00178.1
- NAIP (2009) Oregon 2009 half-meter color orthoimagery. National Agriculture Imagery Program, U.S. Department of Agriculture. http://imagery.oregonexplorer.info/. Accessed 1 June 2010
- National Research Council (2012) Sea-level rise for the coasts of California, Oregon, and Washington: Past, Present, and Future. Committee on Sea Level Rise in California, Oregon, and Washington; Board on Earth Sciences and Resources; Ocean Studies Board; Division on Earth and Life StudiesGoogle Scholar
- NDBC (2007) National Data Buoy Center, National Oceanographic and Atmospheric Administration. http://seaboard.ndbc.noaa.gov/. Accessed 1 June 2010
- NOAA (2007) National Oceanographic and Atmospheric Administration http://tidesandcurrents.noaa.gov/sltrends/sltrends.shtml. Accessed 1 June 2010
- NOAA Center for Tsunami Research (2004) Central Oregon, OR 6 arc-second MHW Tsunami Inundation DEM. NOAA’s National Geodetic Data Center (NGDC). http://www.ngdc.noaa.gov/mgg/inundation/.82. Accessed 18 Nov 2010
- NOAA Coastal Services Center (2002) 2002 NASA/USGS Airborne LiDAR Assessment of Coastal Erosion (ALACE) Project for California, Oregon, and Washington Coastlines. NOAA’s Ocean Service, Coastal Services Center (CSC). http://www.csc.noaa.gov/ldart
- Rahmstorf S (2010) A new view on sea level rise. Nature 4:44–45Google Scholar
- Ruggiero P, Kaminsky GM, Gelfenbaum G, Voigt B (2005) Seasonal to interannual morphodynamics along a high-energy dissipative littoral cell. J Coast Res 21(3):553–578Google Scholar
- Ruggiero P, Komar PD, McDougal WG, Marra JJ, Beach RA (2001) Wave runup, extreme water levels and the erosion of properties backing beaches. J Coast Res 17(2):407–419Google Scholar
- Ruggiero P, Kratzmann MA, Himmelstoss EG, Reid D, Allan J, Kaminsky G (2013) National assessment of shoreline change: historical shoreline change along the Pacific Northwest Coast. U.S. Geological Survey Open-File Report 2012–1007, 55 pGoogle Scholar
- Sallenger AH (2000) Storm impact scale for barrier islands. J Coast Res 16(3):890–895Google Scholar
- Sleeter BM, Sohl TL, Bouchard MA, Reker RR, Soulard CE, Acevedo W, Griffith GE, Sleeter RR, Auch RF, Sayler KL, Prisley S, Zhu Z (2012) Scenarios of land use and land cover change in the conterminous Unites States: utilizing the special report on emission scenarios at ecoregional scales. Glob Environ Change 22:896–914CrossRefGoogle Scholar
- Thieler ER, Hammar-Klose ES (1999) National assessment of coastal vulnerability to sea-level rise, U.S. Atlantic coast. U.S. Geological Survey Open-File Report 99-593Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.