Florida Natural Areas Inventory (FNAI) developed a conservation planning dataset that prioritizes Florida's coastal landscapes based on sea-level rise adaptation potential. This dataset will be used to evaluate state land acquisition projects based on their potential to lessen the impacts of sea-level rise on Florida's coastal biota. It will likely also inform other conservation planning efforts such as Florida's State Wildlife Action Plan and the Peninsular Florida Landscape Conservation Cooperative.
Context
Florida's location and geography make the state especially vulnerable to sea level rise (SLR), often projected at 1-3 meters by the end of this century. Five to 15 percent of the state could be inundated within 90 years (Figure 1). Hundreds of species and associated ecological communities will be affected, though scientists cannot predict with confidence exactly how biota will respond. Protecting the "ecological stage" is an adaptation objective recommended in the Yale framework to protect current and future patterns of biodiversity. This coarse-filter approach recognizes that predicting the responses of individual species and communities to sea level rise involves impractical degrees of complexity and uncertainty.
Protecting the ecological stage is not a new concept (e.g., Anderson and Ferree 2010), but applying this approach to a fine-scale, coastal analysis with a time horizon of only a hundred or so years is new. These specifications raise the issue of whether ecological stages should be defined by "enduring features", such as elevation, slope, soils, and geologic substrate, or by existing natural communities, which may represent ecological stages more accurately in the short run. Defining stages is also complicated in this case by the dynamic nature of Florida's coast--the elevation, slope, soils, and natural communities present along our coast can be changed very quickly by storms. Given these concerns, we have approached stage-definition from both the geophysical and natural-community perspectives, though we have found only one approach to be practical in this context.
Figure 1: Florida's coastal elevation gradients
Project Objectives
- Identify coastal areas with high habitat heterogeneity ("hotspots") and evaluate probable SLR impacts on these hotspots.
- Identify and prioritize potential linkages between impacted and non-impacted hotspots.
- Evaluate the role of geophysical data in SLR-focused conservation planning in Florida.
- Consider applications of SLR-focused conservation planning, especially with regard to Florida's land acquisition program.
Geographic Location
Florida
Principal Investigators
Gary Knight, Michael O'Brien, Jon Oetting,
Amy Knight, Tom Hoctor
Ecosystem Type
Terrestrial
Framework focus
We evaluated habitat heterogeneity and available geophysical data to define and prioritize ecological stages, also known as land facets or arenas, in portions of Florida likely to be affected by sea level rise (areas of up to three meters elevation, plus a one kilometer buffer, with some small gaps filled in). Because of multiple issues surrounding the use of geophysical data in this assessment, habitat types alone were ultimately used to define ecological stages. Areas of high habitat heterogeneity were considered best-suited to sea level rise adaptation, as these heterogeneous locations contain more ecological stages per unit area than do homogenous locales. We therefore identified habitat heterogeneity "hotspots" and evaluated them in the context of connectivity, habitat fragmentation, and ecological integrity. This analysis resulted in a conservation planning dataset that prioritizes coastal landscapes based on sea level rise adaptation potential. This dataset is being used to inform statewide conservation planning efforts in Florida.
Habitat Heterogeneity Hotspots
Our analysis was conducted entirely within ArcGIS. The Focal (or Neighborhood) Statistics tool was used to calculate the variety of natural land cover classes (habitats) within a 25 hectare neighborhood of each cell in the study area (Figure 2). From the resulting variety surface, cells with six or more natural land cover classes within their neighborhood were extracted and grouped. These groups, or "hotspots", were then sorted into two classes based on their elevations: Groups intersecting areas of less than 1 meter elevation but not intersecting areas of 3 or more meters elevation were classified as sources (at-risk hotspots); groups intersecting areas of 3 or more meters elevation were classified as refuges (safer hotspots).
Figure 2: Number of unique natural land cover classes per 25 hectare neighborhood
Linkages Between Impacted and Non-Impacted Hotspots
A cost surface was then developed to represent the resistance biota might experience when moving between sources and refuges. This cost surface incorporated elevation, land-use intensity, and land-cover heterogeneity. Cost-distance surfaces were then calculated to represent the resistance-weighted distances from each source and refuge. These cost-distance surfaces were combined using a least-cost-corridor tool to yield a surface representing each cell's suitability for linking sources with refuges (Figure 3). Corridors with high linkage suitability were then identified.
Figure 3: Classified corridor suitability surface (1 = highest suitability)
Additional Corridor Processes
Two additional processes were performed to demonstrate options for further refining the corridor surface. In the first, the Tabulate Area function was used to prioritize sources in terms of their degree of overlap with other natural resource data layers. Corridors "rescuing" these high-priority hotspots were then identified. In the second, the Cost Path tool was used to determine the most efficient refuge for and escape route from a sample source.
Geophysical Data
We also evaluated the potential for existing geophysical datasets to contribute to SLR-focused conservation planning in Florida. We consulted with Florida NRCS and the Florida Geologic Survey regarding appropriate datasets and relevant attributes. Datasets we considered included NRCS's SSURGO Soils Database, digital elevation models (and derived slope and aspect datasets), and stratigraphic maps. We used various tools, including scatterplots, R-squared values, and the ArcGIS Tabulate Area tool, to determine whether any attributes of these geophysical datasets could serve as reliable predictors of natural community type (which we considered the best representation of the ecological stage in our study area).
Our work was exploratory, serving more to evaluate methods, datasets, and the Yale Framework than to generate quantitative results. This discussion therefore focuses on the former but also offers several maps and figures.
Priority Coastal Landscapes
We found several useful tools for prioritizing coastal landscapes based on SLR adaptation potential. A zonal (or neighborhood) statistical evaluation of a natural land cover dataset proved a straightforward and reliable way to identify coastal areas of high ecological value. This method quickly provided us with a set of 6585 coastal habitat heterogeneity hotspots, which corresponded well with other assessments of conservation value, such as FNAI's Rare Species Habitat Conservation Priorities model (Figure 4). Our comparison of these hotspots with a digital elevation model divided them into two classes, those likely to be impacted by SLR (n=1429) and those likely to remain unimpacted (n=5156). A prioritization of the sources based on size and degree of overlap with other conservation value layers yielded 232 high-priority sources.
Figure 4: Habitat heterogeneity hotspots (divided into sources and refuges) atop FNAI's Rare Species Habitat Conservation Priorities model
The corridor suitability analysis, which identified and prioritized corridors linking high-priority source hotspots with refuge hotspots, resulted in the map shown in Figure 5. We feel that this analysis, executed with the ArcGIS Corridor tool, effectively and intuitively established least-cost linkages between sources and refuges. However, we caution that, because the Corridor tool's outputs are entirely dependent on the two cost distance surfaces fed into it, these surfaces must accurately reflect the resistance biota are likely to encounter while migrating upland. Our choice of cost distance factors--elevation, land cover, and natural land cover heterogeneity--reflects reasonable logic but is not definitive.
Furthermore, we feel that the corridor suitability surface created by the Corridor tool requires refinement to provide much utility. We first classified our corridor suitability surface into rough quantiles, where each class contained approximately 5% of the surface's area. In this classification scheme, the highest quantile (class 1) still contained some 300,000 acres, leading us to explore least-cost path (distinct from least-cost corridor) as an option for further corridor refinement.
Figure 5: Complete classified corridor suitability surface (1 = highest suitability)
We tested the ArcGIS Cost Path tool's effectiveness at refining corridors on one source hotspot. The results, shown in Figure 6, do serve as a refinement to the corridor suitability surface, the most-suitable 20% of which appears in green in Figure 6. Protecting this path and a buffer thereof would obviously be more feasible than protecting the whole green polygon. However, identifying a least-cost path for each source would have required the creation of both a cost distance surface and a cost back link surface for each source, making this an impractical method for developing a refined corridor for each source hotspot. The least-cost-path approach to corridor refinement therefore shows potential but should be regarded as labor- and processing-intensive.
Figure 6: The least-cost path between one SLR-impacted hotspot and a range of non-SLR-impacted hotspots.
Evaluation of Geophysical Data for SLR-Focused Planning In Florida
Our evaluation of available geophysical data in terms of its value for SLR-focused conservation planning in Florida yielded less-encouraging results. As mentioned in the Methods section, we evaluated a number of attributes in a range of geophysical datasets, hoping to use these data to define ecological stages . These stages might be useful for long-term SLR- and climate-focused planning in that they represent relatively enduring features, proper combinations of which might preserve many biotic suites. Through this process we identified several potential hindrances to applying such data to SLR-focused biotic analyses, at least in Florida.
First, Florida's relatively continuous landscape is not as easily divided into geophysical stages as are regions with more dramatic landforms and microclimates. Florida's highest point is only about 105 meters, and slope and aspect are similarly muted. This limited topographic relief also precludes the existence of discrete microclimates, which are important in defining ecological stages. As such, the whole concept of using geophysical data to define ecological stages in Florida might be flawed, a conclusion (among others) that led us to use land cover types as a representation of stages.
In addition, the temporal scale of this project suggested that geophysically defined stages might have less validity than natural-community defined stages. Given that SLR is expected to be a relatively immediate manifestation of climate change, our time horizon for this analysis was about 90 years. Natural communities in existence today should mostly persist through this period (although perhaps not along high energy shorelines), making them a more relevant representation of ecological stages for our time frame.
Other hindrances arise from the datasets themselves. Large spatial gaps exist in the SSURGO dataset, and many of its attributes appear to have been estimated based on other attributes. The best available digital elevation model, created by the Florida Fish and Wildlife Conservation Commission, contained clear inaccuracies and is probably not of sufficient resolution to create a truly accurate representation of slope or aspect. Likewise, the geologic datasets we evaluated were of very coarse resolution relative to the scale of our analysis.
Preliminary Applications
Our intent was to develop a dataset that could be used to help prioritize land acquisitions and inform other conservation planning analyses. Our applications of the hotspot and corridor suitability datasets developed for this project have thus far been preliminary, but these trials have produced interesting results. For example, at least six of Florida's land acquisition projects intersect both a high-priority source and a refuge, suggesting that these projects could serve as (or at least contain) corridors between the two. Furthermore, our corridor suitability surface identifies several high value east-west corridors, especially in southwest Florida. Other conservation planning efforts in this part of the state have focused on north-south corridors (Figure 7), so the identification of east-west options may expand corridor-design thinking in Florida.
Figure 7: North-south corridors proposed by the Cooperative Conservation Blueprint Pilot project (stippled) versus an east-west corridor identified by this project (color).
Evaluating the Yale Framework
The Yale Framework primarily helped us identify key gaps in FNAI's and Florida's conservation planning efforts. Seven of the Yale framework matrix's 17 cells remained largely unaddressed in Florida prior to this assessment: protecting future patterns of biodiversity and protecting climate refugia at all three levels, and protecting the ecological stage at the ecosystem level. In addition, the Framework brought to our attention the concept of protecting the ecological stage, which we now see as an important component of planning in the context of climate change. On the other hand, most of the data sources identified by the Framework were too general or too coarse to be helpful in our analysis. We used our own data and created new data to complete the analysis. Furthermore, while the Framework provided the initial inspiration for the project, it did not serve as a day-to-day guide. In sum, we found the Yale Framework most useful as a means of identifying gaps in Florida's conservation planning efforts, gaps made critical by impending sea level rise.
This exploratory analysis has proven fruitful for establishing a sound method for identifying conservation priorities along coastal areas potentially subject to sea level rise, as well as revealing potential issues that must be considered.
This analysis assumes that high habitat heterogeneity is a desirable surrogate value for protecting biodiversity and associated natural systems. More research is needed to establish whether habitat heterogeneity, as measured by remotely sensed land cover classes, is appropriate for both the spatial and temporal scales associated with sea level rise impacts, which are likely to be much more immediate and on a more local scale than other climate change-related impacts. Regardless, the corridor prioritization method outlined here would prove useful with conservation "hotspots" identified through any number of approaches. Any locations that could serve as sources and refuges in relation to sea level rise would be amenable to this method.
We have also been able to demonstrate potential for the application of this analysis to practical conservation problems, such as the prioritization of existing and future land acquisition projects along Florida's coast, as well as informing current efforts to identify landscape-level networks of ecological connectivity throughout the state.
Additional future directions include:
- Further refinement of corridors linking sources and refuges, using tools such as Least Cost Path analysis or other prioritization approaches.
- More assessment of predicted patterns of biodiversity response to sea level rise.
- Further exploration of geophysical data to define ecological stages in coastal landscapes.
- Further sensitivity testing and scenario comparison to refine the corridor modeling method.
- Expert review of prioritization methods and applications to land acquisition and other conservation efforts.
