Hubs & Corridors 2021

The Blueprint uses a least-cost path connectivity analysis to identify connections between priority areas. A program called LinkageMapper defines corridors that link hubs across the shortest distance possible, while also routing through as much Blueprint priority as possible. Inland corridors connect large patches of highest priority Blueprint areas and/or protected lands, within broad areas of established conservation interest for connectivity. Marine and estuarine corridors connect large estuaries and/or large patches of highest priority Blueprint areas, within broad marine mammal movement areas.

Inland Hubs & Corridors

Inland Resistance Raster

This is the resistance raster or cost surface used in the Linkage Mapper-based connectivity analysis for the inland portion of the South Atlantic area in Blueprint 2021.

Input Data

• South Atlantic 2021 subregions
• 2021 ecosystem integrity scores (rebalanced, mosaiced Zonation results)
• The Nature Conservancy’s (TNC) Resilient Land Project (Anderson et al. 2016)
• TNC’s Secured Lands Database (External Eastern Division 2018)
• U.S. FWS National Wetlands Inventory, accessed 3-4-2021
• Select Eastern North Carolina Wildlife Road crossings (provided by Gary Jordan at the Raleigh, NC U.S. Fish and Wildlife Service Ecological Services field office on 7-22-2021; Gary received these data from Travis Wilson at the North Carolina Wildlife Resources Commission)
• OpenStreetMap data “roads” layer, accessed 3-11-2021
• South Atlantic 2020 Corridor Analysis - Hubs
• South Atlantic 2020 Corridor Analysis - full corridor output from Linkage Mapper (available on request by emailing Amy Keister at amy_keister@fws.gov)
• Eastern Wildways Hubs and Corridor map (available on request by emailing Ron Sutherland at Wildways Network, ron@wildlandsnetwork.org)
• 2016 National Land Cover Database (NLCD)

Mapping Steps

Define Inland/Nearshore Extent of Resistance Raster

• Dissolve all the inland/nearshore polygons from the South Atlantic 2021 subregions into one polygon.
• Buffer the inland/nearshore polygons by 26.1 km. The distance of 26.1 km is based on dispersal distances of subadult black bears (White et al. 2000), a species that disperses from within the South Atlantic into all other adjacent areas.
• Combine the TNC Local Connectedness and Ecosystem Integrity Scores
• Use the local connectedness layer from TNC’s Resilient Land Project to fill in a buffer area around the South Atlantic geography. This allows corridors to connect from inside the South Atlantic geography into the neighboring landscapes.
• Clip TNC local connectedness to the buffered inland/nearshore area.
• Using a linear function, rescale the local connectedness layer to assign values ranging from 1 to 100. Also, flip the high and low values so that areas with high local connectedness are easier to move through (i.e., lower resistance) than areas of low local connectedness.
• Similarly, flip the ecosystem integrity scores so that areas with high Zonation scores are easier to move through.
• Mosaic together the local connectedness layer and the ecosystem integrity scores, only using the local connectedness in places where there are no Zonation results.

Separate Inland Areas from Estuarine/Marine

• Use the edited Estuarine and Marine Deepwater NWI layer created for the estuarine coastal condition indicator to separate inland areas from estuarine/marine areas for the purposes of the connectivity analysis. Note: Some hand editing of the Estuarine and Marine Deepwater class was needed to extend it to the full extent of the South Atlantic Blueprint in the Gulf portion. In the Atlantic, the NWI has the same marine side extent as the HUC2 boundary, but in the Gulf portion, it has a blocky marine side extent that did not match our boundary, which was based on the HUC2 watershed boundary.
• Using the erase tool, remove areas from the above buffered inland areas layer that intersect the edited Esturarine and Marine Deepwater NWI class. This removes nearshore areas from the inland/nearshore extent created above, defining an inland corridor extent.
• Keep the remaining inland areas and use as the extent of the inland corridor analysis.
• Clip the mosaiced TNC local connectedness and ecosystem integrity scores layer to the inland corridor extent.
• Burn NLCD Urban into Resistance Raster
• To get the Linkage Mapper connectivity analysis to successfully run across the whole area, we had to resample the resistance raster from 30 meters to 90 meters. During this process, developed areas often get resampled out. To make sure that the connectivity analysis has sufficient information to move around developed areas, select the High and Medium Developed classes from the NLCD, resample that to 90 meters, and burn it into the resistance raster with a high resistance value.

Make Known Wildlife Road Crossings Easy to Move Through

We obtained a small sample of known wildlife road crossings for eastern North Carolina. This served as a test to route corridors through these crossings. We would like to get more spatial data for wildlife road crossings in the future.

• Buffer known wildlife road crossings by 150 m and convert them to raster with a low resistance value (1) that makes it easy for corridors to route through.
• Select roads surrounding known wildlife road crossings from the Open Streets map data. We do not want corridors to go across these roads except at the designated crossings, so we need to make the resistance very high in these areas. Buffer the roads by 180 m and convert to raster with a very high resistance value (10000).
• In testing, the methods above weren’t enough to encourage corridors to route through the crossings, so we buffered the crossings again, this time by 900 meters, erased the road areas from the larger buffered crossings (so the roads remain a high value) and converted to a raster with value that is easier to move through (10).
• Combine the roads with the mosaiced Zonation results, using a maximum value so that these roads are very hard to cross.
• Combine the above raster with the two buffered road crossing layers, using a minimum value so that the areas with wildlife road crossings get the lower value from the buffered layers or Zonation results.

Limit Resistance Raster to Inland Corridors Mask

At the 2021 South Atlantic Blueprint workshops, we solicited feedback on how stable the corridors in the Blueprint should be over time. Overall, workshop participants strongly preferred a “hybrid” approach in which corridors remain somewhat flexible to improved data and changing landscape conditions, but maintain the same general connections as much as possible. To respond to this feedback, we created an inland corridors mask using both hubs and corridors from the 2020 Blueprint and the Eastern Wildway analysis. By limiting the resistance raster to this mask, we effectively constrained the 2021 Blueprint connectivity analysis, allowing Linkage Mapper to model least-cost paths only within those 2020 Blueprint and Eastern Wildway corridors. In the future, we plan to test adding to the mask corridors from older versions of the Blueprint and other well-vetted datasets.

• This was the first year testing these methods, and we found that the final 2020 corridors were narrower than what would be optimal to implement workshop feedback. We went back to the original Linkage Mapper corridor output from 2020, and selected a deeper cut of the corridors to constrain the 2021 analysis. Select all values <500,000 from the original Linkage Mapper corridor output. In the future, we would like to refine this cutoff.
• We also used the generalized cores and corridors from the Eastern Wildway map to constrain the 2021 corridor analysis. These are very broad and somewhat hand drawn, but they capture many corridors and are a good test case for the approach of constraining the corridor analysis to broad corridors identified previously, while allowing them to fluctuate inside those corridors based on new Zonation results.
• Combine the previous corridor layers. Expand and shrink by 2 cells to eliminate artifacts. Save as a raster, assigning a value of 1 to the previous generalized corridors and a value of 0 to everything else.
• Limit the resistance raster above to the area covered by the previous generalized corridors.

Resample to 90 meters

• Resample the resistance layer from 30 m to 90 m resolution, to allow it to run successfully in Linkage Mapper.

Note: For more details on the mapping steps, code used to create this layer is available in the Ancillary folder of the South Atlantic Blueprint 2021 data download package.

Inland Hubs

These are the hubs used in the Linkage Mapper-based connectivity analysis for the inland portion of the South Atlantic area in Blueprint 2021.

Input Data

• South Atlantic 2021 subregions
• 2021 ecosystem integrity scores (rebalanced, mosaiced Zonation results): all inland subregions
• The Nature Conservancy’s (TNC) Resilient Land Project (Anderson et al. 2016)
• TNC’s Secured Lands Database (External Eastern Division 2018)
• U.S. FWS National Wetlands Inventory, accessed 3-4-2021

Mapping Steps

Use Ecosystem Integrity Scores and TNC Local Connectedness

• Start with the rebalanced ecosystem integrity scores for each inland subregion, mosaiced together into one raster. Identify as potential hubs areas in the highest scoring 10% of the inland 2021 South Atlantic Blueprint extent (based on ecosystem integrity scores).
• Outside of the 2021 South Atlantic Blueprint extent, use local connectedness from TNC’s Resilient Land Project. Identify as potential hubs areas ≥1 standard deviation above average.
• Select polygons from TNC’s Secured Lands Database (External Eastern Division 2018) that are in GAP Status 1 - Permanent Protection for Biodiversity, 2 - Permanent Protection to Maintain a Primarily Natural State, 3 - Permanently Secured for Multiple Uses, or 9 - Unknown GAP status.
• Since military-owned lands are often protected for training or defense rather than conservation, remove any polygons that listed the Fee Owner as U.S. Department of the Army, U.S. Department of Defense, U.S. Air Force, or U.S. Department of the Navy.
• Combine potential hubs from the ecosystem integrity scores, TNC Resilient Land, and Secured Lands.
• Remove from the hubs areas identified as reservoirs to match our approach to reservoirs in the rest of Blueprint 2021.
• Convert this polygon layer to a raster with a cell size of 90 m using the default Esri method (nearest neighbor). This was a quick fix to address a new issue that arose after the introduction of the new Piedmont prairie indicator, which resulted in long, narrow features (powerline rights-of-way) ranking in the top 10% of the Zonation results. Some of these features are up to 15 miles long and only 30 m wide. As a part of its analysis process, Linkage Mapper converts the hubs back to a raster. Since we have to run Linkage Mapper at 90 m to overcome computational limitations, when Linkage Mapper converts the 30 m hubs to a 90 m raster, it breaks up the long linear features into small pieces, creating many small hubs. This produced errors that prevented Linkage Mapper from running successfully. By converting the hubs to 90 m before running Linkage Mapper, we removed those long narrow protuberances from the cores.
• Keep as hubs contiguous patches greater than 2,000 ha (5,000 acres) in size. Identify these patches using the ArcGIS Spatial Analyst-Region Group function (8-neighbor). This size threshold from Hoctor et. al (2000) is also used in the Florida Conservation Blueprint to the south for connectivity purposes.
• Convert the raster patches to polygons.
• Use as inland hubs in the Linkage Mapper analysis all resulting polygons either within the 2021 South Atlantic Blueprint extent or within 26.1 km of the 2021 South Atlantic Blueprint extent. The distance of 26.1 km is based on dispersal distances of subadult black bears (White et al. 2000), a species that disperses from within the South Atlantic into all other adjacent areas.

Inland Corridors

This is the corridor raster used in the inland portion of the South Atlantic area in Blueprint 2021.

Input Data

• South Atlantic 2021 inland corridor resistance raster (created above)
• South Atlantic 2021 inland hubs (created above)

Mapping Steps

• Perform a Linkage Mapper corridor analysis in ArcMap V 10.7.1 using the following settings as pictured in the Blueprint Development Process.
• Linkage Mapper outputs a corridor raster with a continuous surface. We tested different cutoffs and found that keeping all pixels <78,000 created corridors that cover approximately 5% of the area of the inland South Atlantic region (that were not already covered by highest, high, and medium).
• We then removed corridors that ran through reservoirs, to be consistent with how we treated reservoirs in the rest of the 2021 Blueprint. This result is the full inland corridor layer in South Atlantic Blueprint 2021. Inland corridor pixels not already identified as highest, high or medium priority were incorporated into Blueprint 2021 priority connections class. We chose the 78,000 cutoff because it resulted in a priority connections area of approximately 5% of the South Atlantic inland area (not already covered by highest, high, or medium priority pixels).

Note: For more details on the mapping steps, code used to create this layer is available in the Ancillary folder of the South Atlantic Blueprint 2021 data download package.

Marine & Estuarine Hubs & Corridors

Marine & Estuarine Corridors Mask

Input Data

• Inland corridors extent from inland corridors analysis (created above)
• South Atlantic 2021 extent
• U.S. FWS National Wetlands Inventory, accessed 3-4-2021
• Habitat-based marine mammal density models from the Duke Marine Lab covering the entire U.S. Atlantic. They use aggregated survey information, distance-sampling, and oceanographic variables to predict cetacean density throughout the region. Species with monthly models used to create the marine/estuarine corridors mask: North Atlantic right whale, sperm whale, and humpback whale. For additional information on the species models, please visit the Northeast Ocean Data Portal under the heading “Marine Mammals & Sea Turtles”.

Mapping Steps

Define Marine & Estuarine Extent of Resistance Raster

• Reclassify the inland area extent created for the inland corridor analysis to extract marine/estuarine areas.
• Clip the resulting layer to the 2021 South Atlantic extent. This raster allows us to separate inland areas from estuarine/marine areas for the purposes of the connectivity analysis.

Define Open Water Estuaries

• To identify open water estuaries, export the NWI “Estuarine and Marine Deepwater” class as a shapefile. Note: Some hand editing of the Estuarine and Marine Deepwater class was needed to extend it to the full extent of the South Atlantic Blueprint in the Gulf portion. In the Atlantic, the NWI has the same marine side extent as the HUC2 boundary, but in the Gulf portion, it has a blocky marine side extent that did not match our boundary, which was based on the HUC2 watershed boundary.

Identify Marine Mammal Movement Areas

At the 2021 South Atlantic Blueprint workshops, we solicited feedback on how stable the corridors in the Blueprint should be over time. Overall, workshop participants strongly preferred a “hybrid” approach in which corridors remain somewhat flexible to improved data and changing landscape conditions, but maintain the same general connections as much as possible. We have also received repeated feedback, from workshop attendees and from Blueprint users, that they would like marine corridors to better incorporate connectivity for marine mammals. To respond to this feedback, we created a marine corridors mask using all open water estuaries and movement areas for marine mammals. By limiting the resistance raster to this mask, we effectively constrained the 2021 Blueprint connectivity analysis, allowing Linkage Mapper to model least-cost paths only within those estuaries and marine mammal movement areas.

• We selected 3 marine mammal species from the Duke Marine Lab data with both sufficient data to create movement areas and significant interest in ensuring connectivity for the species. The 3 different movement areas used by these species also cover most of the major movement areas used by other marine mammals in the South Atlantic.
• For each of the 3 marine mammal species, sum the monthly abundance rasters across all months.
• For right and sperm whales, reclassify the summed abundance to retain only the top 70% of values.
• The above approach did not create a reasonable connected corridor for humpback whales. For humpback whales, reclassify the summed abundance based on a manually identified threshold (0.04) to create a fully connected corridor for the species.
• Combine all three species movement areas into a single layer identifying whether a pixel is in at least one marine mammal corridor.

Combine Marine Mammal Movement Areas and Estuaries

• Mosaic together the combined marine mammal movement areas and the hand-edited NWI estuarine and marine deepwater class, resulting in a mask combining open water estuaries and the marine mammal movement areas. The resulting layer serves as the marine/estuarine corridors mask.

Note: For more details on the mapping steps, code used to create this layer is available in the Ancillary folder of the South Atlantic Blueprint 2021 data download package.

Marine & Estuarine Resistance Raster

This is the resistance raster or cost surface used in the Linkage Mapper-based connectivity analysis for the marine and estuarine portion of the South Atlantic area in Blueprint 2021.

Input Data

• U.S. FWS National Wetlands Inventory, accessed 3-4-2021
• Rebalanced, mosaiced 2021 ecosystem integrity scores
• Flipped Zonation results from inland corridors analysis
• Inland vs. estuarine/marine extent (created above)
• Marine/estuarine corridors mask (created above)

Mapping Steps

Make Resistance Raster

• Clip the flipped Zonation results to the estuarine/marine extent.
• Clip the resulting raster to the marine/estuarine corridors mask.
• Resample to 90 m using the bilinear function.

Note: For more details on the mapping steps, code used to create this layer is available in the Ancillary folder of the South Atlantic Blueprint 2021 data download package.

Marine & Estuarine Hubs

These are the hubs used in the Linkage Mapper-based connectivity analysis for the marine/estuarine portion of the South Atlantic area in Blueprint 2021.

Input Data

• U.S. FWS National Wetlands Inventory, accessed 3-4-2021
• Rebalanced, mosaiced 2021 ecosystem integrity scores
• Inland vs. estuarine/marine extent (created above)

Mapping Steps

• To identify open water estuaries, select estuarine and marine deepwater areas from the NWI and remove areas in the buffer that extend into the marine environment. (WETLAND_TY = ‘Estuarine and Marine Deepwater’ And ATTRIBUTE <> ‘M1UBL’). Note: Some hand editing of the Estuarine and Marine Deepwater class was needed to extend it to the full extent of the South Atlantic Blueprint in the Gulf portion. In the Atlantic, the NWI has the same marine side extent as the HUC2 boundary, but in the Gulf portion, it has a blocky marine side extent that did not match our boundary, which was based on the HUC2 watershed boundary.
• Add and calculate a field to use to convert the resulting layer to raster.
• Convert to raster to create an open water estuaries layer. All open water estuaries will be considered potential hubs.
• Identify the highest scoring 10% of the marine area (based on ecosystem integrity scores). These will also be considered potential hubs in the marine ecosystem.
• Combine the open water estuarine hubs and the marine Zonation-based hubs using a cell statistics maximum function.
• Convert the resulting layer to a 90 meter raster. This step was necessary in the inland hubs to deal with impacts from the Piedmont prairie indicator that created long, linear protrusions from the hubs. We kept this step it in the marine connectivity analysis as a precaution.
• Use the ArcGIS Spatial Analyst-Region Group function (8-neighbor) to identify patches of marine/estuarine hubs. Contiguous patches greater than 2000 ha were kept as hubs. This is consistent with terrestrial hubs and also within the 2-10 km width recommended by Krueck et al (2017).
• Convert the raster patches to polygons for use in Linkage Mapper.

Note: For more details on the mapping steps, code used to create this layer is available in the Ancillary folder of the South Atlantic Blueprint 2021 data download package.

Marine & Estuarine Corridors

This is the corridor raster used in the marine portion of the South Atlantic area in Blueprint 2021.

Input Data

• Marine resistance raster (created above)
• Marine hubs (created above)

Mapping Steps

• We performed a Linkage Mapper corridor analysis using the following settings as pictured in the Blueprint Development Process.
• Linkage Mapper outputs a corridor raster with a continuous surface. We tested different cutoffs and found that keeping all pixels <750,000 created corridors that cover approximately 5% of the area of the estuarine/marine South Atlantic region (that were not already covered by highest, high, and medium).

Note: For more details on the mapping steps, code used to create this layer is available in the Ancillary folder of the South Atlantic Blueprint 2021 data download package.

Known Issues

• When creating the marine hubs, we identified the top 10% of the marine Zonation runs using values >90. Later, when creating the final Blueprint, we realized that using ≥90 did a better job of getting at the top 10%. As a result, the marine hubs cover slightly less area than the marine highest priority class. If we had used ≥90 to calculate marine hubs, we would have also had slightly different marine corridors and marine priority connections.

Literature Cited

Anderson, M.G., A. Barnett, M. Clark, C. Ferree, A. Olivero Sheldon, J. Prince. 2016. Resilient Sites for Terrestrial Conservation in Eastern North America. The Nature Conservancy, Eastern Conservation Science.

Hoctor, T. S., M. H. Carr, P. D. Zwick. 2000. Identifying a linked reserve system using a regional landscape approach: the Florida ecological network. Conservation Biology 14:984-1000.

Homer, Collin G., Dewitz, Jon A., Jin, Suming, Xian, George, Costello, C., Danielson, Patrick, Gass, L., Funk, M., Wickham, J., Stehman, S., Auch, Roger F., Riitters, K. H., Conterminous United States land cover change patterns 2001–2016 from the 2016 National Land Cover Database: ISPRS Journal of Photogrammetry and Remote Sensing, v. 162, p. 184–199. https://doi.org/10.1016/j.isprsjprs.2020.02.019.

Krueck, N. C., Legrand, C., Ahmadia, G. N., Estradivari, E., Green, A., Jones, G. P., Riginos, C., Treml, E. A. and Mumby, P. J. 2017. Reserve Sizes Needed to Protect Coral Reef Fishes. Conservation Letters. Accepted Author Manuscript. doi:10.1111/conl.12415.

McRae, B. and Kavanagh, D. Linkage Mapper User Guide: Version 1.0. 2014. http://www.circuitscape.org/linkagemapper.

Moilanen, A., Pouzols, F. M, Meller, L., Veach, V., Arponen, A., Leppänen, J. Kujala, H. Zonation Version 4 User Manual: Spatial conservation planning methods and software. https://github.com/cbig/zonation-core/releases/download/4.0.0/zonation_manual_v4_0.pdf.

The Nature Conservancy. 2018. Secured Lands dataset. The Nature Conservancy, Eastern Conservation Science. Boston, MA. http://www.conservationgateway.org/ConservationByGeography/NorthAmerica/UnitedStates/edc/reportsdata/terrestrial/secured/Pages/default.aspx.

White, T.H., JR., J.L. Bowman, B.D. Leopold, H.A. Jacobson, W.P. Smith, and F.J. Vilella. 2000. Influence of Mississippi alluvial valley rivers on black bear movements and dispersal: Implications for Louisiana black bear recovery. Biological Conservation 95:323–331.