Least-cost path analysis for wildlife corridors, explained

If you've sat through a pre-application meeting where the agency biologist asks for "corridor connectivity analysis" without specifying a method, you already know the term covers a few different things. Least-cost path analysis is the one that shows up most in mitigation planning, partly because the inputs are familiar and partly because the output is a single line on a map that's easy to put in a report.

Here's what it does, and where it runs into trouble before you've even picked a site.

What a resistance surface is

Least-cost path analysis needs a resistance surface before it can do anything. That's a raster where every cell gets a value for how hard it is for an animal to move through, based on land cover, slope, road density, or whatever variables your target species responds to. Dense forest might get a resistance value of 1. A four-lane highway might get 1,000. Open pasture lands somewhere in between, and that "somewhere" is usually a judgment call, not a measurement.

This is the part that eats consultancy time. Someone has to classify the study area into cover types, assign resistance values to each class, and defend those values if a reviewer pushes back. Without a current land cover layer for the area, that step starts from a basemap that might be years out of date, which matters when the proposed corridor runs through land that's been cleared or replanted since the last survey.

Once the surface exists, the least-cost path algorithm finds the route between two points, say a known habitat block on one side of a parcel and another on the far side, that accumulates the lowest total resistance. It's a real path, not a buffer or a straight line, and it respects the terrain and cover you fed it. That's genuinely useful for sketching where a corridor might go before anyone's walked the ground.

Where it breaks down for offset screening

The method has two known weak spots, and both matter for sourcing offset sites rather than designing one specific crossing.

First, a least-cost path draws one line. Real animals don't move along a single optimal route. They use a braided set of reasonable options, and a single-path output can make a corridor look narrower and more fragile than it is, or miss a perfectly viable secondary route a few hundred meters off the modeled line. Circuit theory models address this by treating the landscape like an electrical network and showing current flow across many possible paths, which is a better fit when you're trying to understand overall connectivity rather than propose one specific alignment.

Second, and this is the part that actually slows down a shortlisting exercise: least-cost path analysis tells you nothing about patch size, edge density, or how fragmented a candidate habitat block already is. You can run a beautiful corridor model between two patches and still end up proposing a site where the "habitat block" on one end is a sliver too narrow to function as anything, because nobody checked the fragmentation metrics first. Agencies increasingly want to see patch size and connectivity numbers alongside any corridor sketch, not instead of one.

That's the gap that makes commissioning a full bespoke ecology survey before you've even shortlisted sites feel backward. You're paying for resistance-surface precision on a parcel you might not end up recommending.

A faster first pass is to start from an annual, high-resolution land cover classification of the study area and pull the fragmentation metrics, patch size, edge density, connectivity between blocks, directly from that. Habitat Suitability turns that classification into a land cover layer and a connectivity metrics table for whatever area you define, which gives you something concrete to screen candidate sites against before anyone commissions the detailed resistance-surface modeling for the site you end up recommending.

If you're staring down a shortlist of parcels and need patch and connectivity numbers before the ecology survey gets commissioned, that's the gap this is built to close.