Jenness Enterprises
Wildlife and Forestry Tools for ArcGIS Pro — Tool Reference
jennessent.com
Tool Reference › Corridor Designer Tools › About Land Facet Corridors

About Land Facet Corridors

Corridor Designer Tools · background and theory · by Jeff Jenness

The land facet approach designs wildlife linkages that stay useful in the face of climate change — by conserving the arenas rather than the actors. A conventional corridor (see About Corridor Design) is built on today's vegetation map, and today's vegetation map is exactly what climate change will redraw. Land facets — landscape units defined by enduring topographic and soil traits, like high-elevation north-facing slopes on rocky soils or low-elevation flats with deep soils — will still be there when the vegetation has moved. Future plant communities (and, indirectly, animal assemblages) will be determined by the interaction of these enduring physical stages with whatever the future climate brings, so a linkage that keeps continuous strands of each facet should keep continuous strands of future vegetation, whatever that vegetation turns out to be. The concept is Beier and Brost (2010); the procedures and validation are Brian Brost's NAU thesis work; and this add-in's Land Facet tools modernize the ArcGIS 9.3 + R toolkit that Jeff and Brian built for the original project (Brian wrote the R functions, Jeff the ArcGIS extension).

What a land facet linkage looks like A multistranded design: one least-cost corridor per major land facet (typically 5–15 of them), plus one corridor maximizing local interspersion and diversity of facets (to support range shifts and species turnover), plus a hand-drawn riverine strand for aquatic species and wetland processes. The union of the strands is the linkage design — no global circulation model, emission scenario, or species-specific climate envelope anywhere in the chain, and it works at the full resolution of your DEM, even where no vegetation map exists at all.

The five Major Steps

Step 1 — define and map the land facets. Using only pixels inside the wildland blocks (the facets should represent what the linkage connects), each pixel is first assigned a topographic position — canyon, ridge, or slope — and then, within each position, classified on a few continuous variables (the original work used elevation, steepness, and — for slopes — annual solar insolation; soil variables join in wherever decent soil maps exist). Multivariate outliers are removed, fuzzy c-means clustering finds the natural groupings, the analyst chooses the cluster count that matches the landscape's real “lumpiness,” and poorly classified pixels are dropped via a confusion index. The result is typically 8–16 interpretable facets. In the original toolkit this step required exporting to R; the Land Facet Clustering tool now runs the whole procedure — outlier screening, fuzzy c-means, and the confusion screen — inside ArcGIS Pro.

Step 2 — develop resistance maps. For a focal species, resistance means difficulty of movement; for a land facet, resistance is how far a pixel departs from the prototypical cell of the facet — measured as Mahalanobis distance, the number of “multivariate standard deviations” between the pixel's attributes and the facet's characteristic values, with one deliberate twist: the facet-density variable's ideal is pinned at 100% (a perfect pixel is surrounded by its own facet). That is precisely what the Land Facet Density and Land Facet Mahalanobis tools compute. For the interspersion corridor, resistance is instead built from Shannon diversity of facets in a neighborhood (resistance = 1 / (H′ + 0.1); the Diversity Indices tool provides the H′ surface). Urbanized areas get NoData — infinite resistance — though the original manual cautions against wholesale exclusion of agricultural land that could be restored.

Step 3 — least-cost corridor modeling. Exactly as in the focal-species workflow: termini where the facet is most concentrated in each block, cost-weighted distance summed from both ends, and a slice chosen as the corridor — the original guidance suggests a minimum width of roughly 1 km for corridors under 10 km, up to about 2 km for longer ones. The Identify Termini Polygons and Least-Cost Corridor tools cover this step.

Step 4 — add a riparian corridor if needed. Rivers and drainages span elevation gradients, concentrate moist soils, and carry animals, water, sediment and nutrients — and a human expert usually draws the important riverine connection more sensibly than any algorithm. The manual recommends hand-digitizing it.

Step 5 — join the strands. The union of the facet corridors, the interspersion corridor, and the riverine strand is the preliminary linkage design (the Union Corridors tool). Where focal-species corridors can also be modeled, start from the land facet design and expand it to serve any species the facets miss.

Does it work?

Brost tested the approach in three Arizona landscapes where focal-species linkage designs already existed. Linkages designed purely from land facets — with no biological input at all — served 25 of 28 focal species as well as or better than the linkages designed expressly for those species; the three exceptions were the species with the most narrowly distributed habitat. The reverse was not true: the focal-species designs covered only about half the land facets comparably well. The recommendation that falls out is to use land facets to complement, not replace, the focal-species approach — start from the (typically larger) land facet design and graft on whatever the habitat-specialist species need.

The Major Steps, as Pro tools

Major StepTools in this add-in
1. Define facets Topographic Position Index tools (canyon / ridge / slope), then Land Facet Clustering (outliers + fuzzy c-means + confusion screen, all internal — no R required).
2. Resistance Land Facet Density, then Land Facet Mahalanobis (one cost surface per facet); Diversity Indices + Invert Raster for the interspersion corridor.
3. Corridors Identify Termini Polygons, Least-Cost Corridor.
4–5. Join Union Corridors, Fill Corridor Holes, Cumulative Surface (how many strands cover each cell), and the Corridor Analysis Data Report, which reads the provenance chain the tools stamp on every output.

The tools validate the chain as they go — each step records what produced its inputs, so a density raster built from the wrong clustering run, or a Mahalanobis surface fed a mismatched density band, is caught rather than silently accepted.

Credits and references

The land facet methodology is by Paul Beier and Brian Brost; the original ArcGIS tools and these ArcGIS Pro ports are by Jeff Jenness, Jenness Enterprises (www.jennessent.com), with Brost and Beier. Produced originally with the support of the USDA Forest Service Rocky Mountain Research Station and the Arizona Board of Forest Research / McIntire-Stennis Cooperative Forestry Program.