Written by Javan Bauder
Connectivity is a very important theme in conservation biology and refers to the ability (or lack thereof) of individuals to move among habitats across a landscape. Good connectivity is important for many reasons. For example, if a population goes extinct other individuals can immigrate to the site and restore the population. Sufficient levels of connectivity can prevent such extinctions in the first place. High connectivity also maintains the genetic diversity of a population and prevents inbreeding.
Flatwoods like these provide prime habitat for Eastern Indigo Snakes in central Florida.
But how do we actually measure connectivity? And once we have measured connectivity, how can we use those measures to identify locations on a real landscape that will either promote or restrict connectivity? Many theses and dissertations have been earned by addressing these questions! As you might expect, there is no one-size-fits-all solution but today I would like to talk about some approaches I am using to answer these questions for Eastern Indigo Snakes in central Florida as part of my dissertation work with Dr. Kevin McGarigal at the University of Massachusetts.
One common way to measure connectivity is to measure genetic connectivity which describes the degree to which genes, rather than individuals, move across the landscape. Why not just measure the movements of individuals rather than genes? First, observing individual movements, particularly dispersals between populations, is very time-consuming and difficult for small, cryptic species (i.e., virtually all reptiles and amphibians). Second, just because an individual moves to a new population doesn’t mean that it successfully reproduces and transfers its genes, which is what helps maintain genetic diversity and prevent inbreeding. Finally, we can get information on genetic connectivity by extracting DNA from scale clips, hair, scat, shed skins, or even the water (see this previous post on eDNA http://www.oriannesociety.org/blog/dna-in-water). While this can still be difficult, it is much easier than monitoring individual movements. If we can collect genetic samples from individuals or populations over some spatial extent, we can then quantify the degree of genetic similarity among samples. The greater the similarity, the greater the “connectedness.”
Shed skins can provide a valuable source of DNA, particularly for difficult to observe species (like Eastern Indigo Snakes!)
Once we have described genetic connectivity we then usually want to know how connectivity varies across space. This can help us identify areas of high connectivity (i.e., corridors) and low connectivity (i.e., barriers). For most wildlife species, connectivity might be greatest in patches of suitable habitat and lowest around features like roads or urban development. Connectivity (not just genetic connectivity) is often represented spatially using resistance surfaces. Resistance surfaces are simply maps where each point on the map describes how “resistant” that point is to movement or gene flow. To think about resistance, imagine you are on a bike trail. Flat, straight stretches are the easiest to pedal along and therefore have low resistance. Flat, curvy stretches might have higher resistance because you may need to slow down for the curves. Very hilly stretches require the hardest pedaling and therefore have the highest resistance. The same idea applies to wildlife resistance surfaces. If you are trying to decide which parts of the landscape to protect to promote connectivity having an accurate resistance surface can be essential.
But how do you get an accurate resistance surface from wildlife genetics data? One of the most common approaches is to assign landscape features a resistance value based on expert knowledge of the species and landscape (see this post for an example using Western Rattlesnakes http://www.oriannesociety.org/blog/keeping-rattlesnakes-connected). For example, forest may get a resistance value of one while roads may get a resistance value of 50 (meaning roads are 50 times more resistant than forest). You can then compare your resistance surfaces to your genetics data to see which surface does the best job of explaining the spatial variation in your genetics data. The problem is that you may come up with hundreds of biologically reasonable resistance surfaces and testing all those surfaces against your genetics data may be very time consuming. Isn’t there an algorithm that can compute the optimal resistance values from your genetics and landscape data?
Yes, there is! Dr. Bill Peterman at Ohio State University has developed software called resistanceGA to empirically optimize resistance surfaces. A user provides their genetics data and one or more surfaces representing different landscape features (e.g., land cover, elevation, forest cover, etc.). The software then finds the optimal resistance values for each landscape feature using a genetic algorithm (hence the “GA” in resistanceGA). A genetic algorithm mimics the process of natural selection whereby individuals with new genotypes are produced through mating and mutation. Individuals with the highest “fitness” are allowed to move to the next generation. In resistanceGA genotypes represent potential resistance values and fitness is defined as a model-based measure of how well a resistance surface using those resistance values explains spatial variation in the genetics data. Resistance values that explain more of this variation are the “fittest” genotypes and individuals with those genotypes are allowed to move on to the next generation. This process repeats until it finds the “best” resistance values.
With the support of The Orianne Society and University of Massachusetts, I have been working with Dr. McGarigal, Dr. Peterman, Dr. Steve Spear from The Wilds at the Columbus Zoo, and Dr. Chris Jenkins to create resistance surfaces for Eastern Indigo Snakes in central Florida. Using resistanceGA, we will not only be able to create accurate resistance surfaces but also identify landscape features that most strongly affect Eastern Indigo Snake genetic connectivity. Our results indicate that the amount of undeveloped upland habitat, which includes scrub, sandhill, flatwoods, and unimproved pasture, has the strongest influence on genetic connectivity. Not surprisingly, we have also found that Eastern Indigo Snakes strongly select these habitats. Here is an example of one of our resistance surfaces. Green areas have lowest resistance/highest connectivity.
This project is still ongoing so stay tuned for more updates! We hope that this project will help identify important areas for promoting genetic connectivity for Eastern Indigo Snakes in central Florida.