Landscapes = Circuits?

Here's an interesting story from Northern Arizona U. A researcher named Brian McRae, a recent forestry graduate who used to be an electrical engineer, used his knowledge of circuit theory to craft a new model for gene flow through landscape corridors:

McRae had been struggling with how to predict genetic effects of landscape pattern while working with Beier on a study of cougars in the southwest United States. "We had maps of cougar habitat and genetic samples spread across four states," he said, "but no way to predict how habitat pattern was driving gene flow across the region."

Using experience from his previous career as an electrical engineer, he reasoned that gene flow across a complex landscape should follow the same rules as electrical conductance in a complex circuit board.

The result was what McRae calls the Isolation by Resistance model, so named because it's similar to standard "isolation by distance" models used by geneticists, but incorporates landscape resistance into gene flow predictions. The model represents patches of habitat as nodes in an electrical circuit and the genes of animals and plants as the current that flows between the nodes. Flow occurs across multiple pathways, encountering more resistance in some areas--poor habitats or human-made barriers--and flowing preferentially through better habitats.

The resistance model incorporates multiple pathways, instead of just the most obvious one. It represents the landscape as a conductive surface, calculating all possible pathways connecting the patches. The PNAS article tested how well the new model explained genetic patterns across 12 wolverine populations across the United States and Canada, and eight big-leaf mahogany populations in Central America.

McRae and collaborators are now using the model to pinpoint critical linkages in landscapes and aid in conservation planning. "If you can imagine current flowing across a landscape, areas where it concentrates--bottlenecks or pinch points in the flow--typically correspond to important areas to maintain connectivity," McRae said. "If you can distribute that current across multiple corridors, you'll get greater connectivity, greater gene flow and greater robustness to climate change or catastrophes like wildfires."

Interesting. I'm a bit rusty on circuit theory, but it certainly seems like a fresh approach, especially since "isolation by distance" assumes certain ideal conditions, as McRae notes in his paper describing IBR:

...such analyses assume homogeneous, unbounded populations, ignoring effects of range boundaries and of variation in demographic parameters within species' ranges. As Slatkin noted, most real populations are neither homogeneous nor unbounded, and when population densities or migration rates vary across space, the predictive abilities of these models may suffer.