The snowshoe hare-coyote-Dall's sheep cycle?

When we teach our students predator-prey relationships, we tend to tell them a very basic story that have come from a few very good studies. It's not that all we have are simple models, but but we like to build up the basics before we dive into the complexities. Because when you get down to the complexities, a lot of counter-intuitive things happen which don't follow "common-sense" relationships.

At the last Alaska chapter of the Wildlife Society meeting, Steve Arthur presented one such study, which I had previously mentioned here by flinging out the abstract. Now the full paper by him and Laura Prugh is out in the Journal of Wildlife Management, and so I want to spend a little more time on the study.

They studied snowshoe hares (Lepus americanus) and Dall's sheep (Ovis dalli) in the central portion of the Alaska Range, illustrated by the figure to the right. It's worth noting that though both species are in the Alaska range, they have differing habitat requirements. Whereas hares are happy in the lower foresty areas, the sheep decidedly prefer the rocky hills where they can can try to avoid predators. However, while adult survival among sheep is fairly high, lamb survival is generally lower, as predators (lynx, coyotes, eagles, wolverines) can still prey on them.

They used road-side counts to index hare abundance, which they double checked against pellet count data - it's little things like that which make a good paper. For the sheep end, they collared a variety of ewes and fitted them with collars, as well as lambs. They re-located them about 3-5 times a week, and the collars were fitted with mortality sensors. On death, they went and determined cause of death of the individuals. One nit-pick here - scavenging is probably represented somewhat in the mortality data, but the two authors admit that. Still, you wouldn't expect scavanging and predation to vary independently of mortality, and survival is what they wanted to get at. Finally, they used surveys to index the total population of Dall's sheep through the period.

When they analysed the adult data, they found that the year really didn't have a major effect on their survival, and that survival was either modeled best as age-specific, or just constant. However, for lambs, the best explanations for the changes in survival were either a three year time lag in hare abundance, positively, or a negative relationship between survival and hare abundance with a one year time lag. The mechanism for changing survival appears to be changes in predation - primarily coyote and golden eagle predation. Accidents and disease seem to remain at a constant low-level throughout all years, whereas predation seems to vary among years. Eagles are best predators early on, whereas coyotes appeared to kill several months after ewes dropped their lambs. As I mentioned before, snowshoe hares and sheep share those predators.

This seems to suggest that although there's a positive relationship between abundance of hares and sheep, it is not, in fact, a mutualistic relationship. If they had just limited their study to abundance, they would have never seen the change in lamb survival. At that point, we would be led to incorrectly believe that snowshoe hares and sheep benefit from each-other's presence. It doesn't appear to be mediated by a functional response in predation - first, a previous study found that coyotes don't switch to sheep when they're low, but secondly, it would mean there would be a positive relationship between hare abundance and sheep survival, not negative.

It's a messy story, one that depends on the sort of data you gather, and, this simplified version I'm presenting here is leaving out further complications itself. The effects of predation can be complex, leading to strange results when all is said and done. Here, the population dynamics of Dall's sheep seem at first blush to be driven by a species that isn't even in the same habitat as them. It's only though diligent data collection the true picture comes out.


Arthur, S., & Prugh, L. (2010). Predator-Mediated Indirect Effects of Snowshoe Hares on Dall's Sheep in Alaska Journal of Wildlife Management, 74 (8), 1709-1721 DOI: 10.2193/2009-322


All figures from Arthur & Prugh 2010, used under fair-use rationale. 

Prodigious Pronghorn Population Projections

Can the presence of wolves be good for prey species? Intuition seems to suggest the answer is no. After all, wolves eat prey, and being eaten is fairly bad for one's health. Wolves are implicated in a number of natural declines of prey species in a number of systems, especially in closed populations where immigration can't bolster floundering populations. The use of wolf control is a controversial tool which I won't discuss at great length here, but one of the key assumptions is that it increases survival of some age classes of prey.  Before I go much further, I've added some entries to my glossary over to the right, for ecological terms.

One of the most vulnerable age classes are neonatal animals and juveniles, or the year's young. While surviving to second or third year doesn't guarantee you'll survive to old age, if you're a deer, it does mean you're fairly less likely to die. Adults are very good at surviving, since evolution has shaped them to be surviving machines, capable of hyper-vigilance, amazing running speeds, and so on. The young, however, enjoy much less of that sort of protection. There tends to be lower survival for the young of the year, and this can be a major factor governing population fate - if you die young, it's hard to do much breeding.

Re-colonizing wolves provide a wonderful mini-experiment to see what the effects of their presence or absence is on various age groups. Kim Berger and Mary Conner, both at Utah State University at the time, studied the effects of re-colonization on pronghorn neonatal survival. Berger and Conner selected two study sites with wolves, and one that was free of wolves in Wyoming. Coyotes were much less abundant at the sites with wolves than they were at the sites with wolves, as they had demonstrated in a previous study.

Fawns were located through a good old fashion leg-work (they must have good vehicle access out there!), and through watching gravid females until they dropped their fawns. They would capture the neonates, weigh them, age them, record the sex, and assess over-all health in a few categories before fitting them with a breakaway radio collar with a mortality sensor. When the animal dies, the radio signal from the collar changes, allowing the researches to swoop in and look for cause of death. Some things they would look for include

1. whether the fawn was alive at the time of the attack
2. tracks, scat, hair, and any caches that might help them identify the predator
3. signs of starvation or other accidents.

The sort of predator who killed it could be classified based off of previously described characteristics. One thing I've always wondered is if there is a better way to ID the cause of death more adequately. However, kill site characteristics have been frequently used before, and will suffice for here.

Ignoring the factors of survival for a moment, there was much greater fawn mortality among wolf free sites than wolf abundant, with as low as 0% survival in one year. The overwhelming majority of the mortality came from either demonstrable coyote predation, or likely coyote predation in those wolf free sites. Coyote predation was still a major factor in wolf abundant sites, but over-all survival was much higher in all years, with the lowest being 24%, higher than the highest survival in any year in the wolf free site. Thus, the presence of wolves appears to beneficially change the survival rate of pronghorn calves through reducing coyote mortality.  Modelling this forward using other vital rates for pronghorn, they find that the wolf-free sites will gradually suffer total collapse, while the wolf abundant sites are quite stable. Clearly, our earlier intuition was not entirely correct about the effects of having wolves in these populations.

This fits nicely into a framework that I've previously mentioned, which is the "Meso-predator release hypothesis." The idea is that wolves or other apex predators (bears, cougars, etc.) keep middling scavenger-predators (such as fox, coyote, racoon, skunk) suppressed when they're around and healthy. They can suppress them through feeding on common prey, attacking meso-predators directly, and so on. This is a top-down effect, where high trophic levels (animals that do bulk of the eating other things) influence the composition of the lower trophic levels (prey; the things that are eaten).

Within the state of Alaska, it begs the question whether there are significant numbers of meso-predators, and whether they are similarly released in times of intensive management. A entirely separate and important question is whether this release negatively impacts prey populations. To my knowledge, no one has done any sort of extensive study of meso-predators in the state. Our knowledge of background fox and coyote density is not very good, and after asking several people, I begin to suspect there are no density estimates for regions of the state. Given that, I am sure no one has looked at meso-predator densities during periods of IM. Anyone who knows otherwise is encouraged to email me! However, the latter most question is probably best identified - there have been studies of calf mortality after bouts of IM, and I don't anyone's identified a shift in mortality towards meso-predation.

---

Berger KM, & Conner MM (2008). Recolonizing wolves and mesopredator suppression of coyotes: impacts on pronghorn population dynamics. Ecological applications : a publication of the Ecological Society of America, 18 (3), 599-612 PMID: 18488620

Cauga una?

Oops! I forgot last week! Sorry! "That person who usually lurks" Got it right - the thing to the left (left side of the screen) is a coyote (Canis latrans). The yup'ik, Kayu, is borrowed from the kassacetun. Or, if it's not from English, it's a huge coincidence!
The coyote isn't endemic to Alaska, and to me, it's low population is somewhat of a mystery. Typically, when you get invasive species, they proceed to explode, frequently drowning out endemic species that fill a similar niche in the enviroment. Curiously, that's not what we've seen. Which makes me wonder a few things. Was there a species filling the coyotes niche that we don't see anymore? Were coyotes around an exceptionally long time ago, and left the niche vacant? Is there some factor limiting coyote population growth?

These are all questions I can't answer - obviously, because I wouldn't be wondering them if I could!

Here's this weeks':

I apologize the first one isn't in better focus. I'm just not fast enough with the camera, I guess! But here's a small hint: It's in Alaska, and it's not a Yellow Shafted Flicker.