(For months, the question floating around the office was, “Has Jeff Copeland’s snow paper been published yet?” It came out in March of 2010. This is a summary of what that paper found. I’ve been working on this for weeks and, like some Sysiphian task, I’ve found that the closer I’ve come to something workable, the busier life becomes, and the further from completion the piece gets. So I’m posting it as is, because I have to leave for Canada tomorrow and then Mongolia the week after that. Forgive any imperfections or lack of cohesion.)
The Magic Line
Find a map of the northern hemisphere, and locate the 54th parallel. Trace its arc across the globe – in Siberia, 54N cuts across the southern tip of Kamchatka and the northern tip of Lake Baikal, then passes just south of Moscow. It brushes past the southern edge of Sweden, decapitates the Danish peninsula, and slices England in half. Across the Atlantic, it comes ashore on Newfoundland, sweeps across northern Quebec, and passes through the top of James Bay. In Ontario, it demarcates the southern edge of Polar Bear Provincial Park, and then cuts across a sweep of sparsely inhabited, lake-studded Manitoba, Saskatchewan, Alberta, and British Columbia, reaching the Pacific at Hecate Strait just south of Alaska. North of 54, the world stretches, colder and ever more severe, in a mix of boreal forest and tundra, towards the Arctic Circle at the 80th parallel. The region is underlain by permafrost, discontinuous but widespread closer to 54N, deeper and more continuous to the north. The landscape, though ridged with several mountain ranges in both hemispheres, is primarily flat, and remains covered in snow until late spring. North of 54, the Pleistocene lingers.
54 North is the wolverine’s Magic Line. In the flat, snowbound tundra to the north of this parallel, everything is, essentially, wolverine habitat, and the species is spread in a continuous distribution across much of the landscape. South of 54, boreal habitat grows sparser, migrating uphill as you go further south, until, by the time you reach the US Rockies, it is pinched into the upper reaches of the highest mountains. Wolverine distribution migrates uphill with the habitat, until a creature designed for cruising the vast and primarily flat North becomes, of necessity, a mountaineer. Restricted to these islands of boreal habitat in a sea of sagebrush desert, the wolverine faces a unique set of challenges that center around access to suitable places to live.
Parameter 1: Persistent Spring Snow
Wolverine biologist Jeff Copeland and his colleagues in the wolverine biology world were attempting to determine the parameters of these challenges when they began to look at variables that might influence wolverine habitat selection. What, exactly, did a wolverine need in order to consider a place livable? Working with telemetry locations and confirmed wolverine den sites from research projects in North America and Scandinavia, they began sorting through a list of possibilities – elevation, aspect, vegetation, prey, weather conditions, proximity to human development. Habitat selection models for wildlife can become complicated as factors interact with each other, but in the case of the wolverine, a single overriding variable seemed to be the best predictor of wolverine distribution: persistent spring snow.
At an intuitive level, this made sense. Wolverines den in the snow, so of course they would tend to adhere to regions with adequate snow to protect their babies until the kits were capable of limited independence, in early to mid-May. And the logical and intuitive conclusion would be that wolverines, being dependent on snow through May, are vulnerable to reduced spring snowpack predicted in climate models for the next century. All of this would seem to suggest that the wolverine requires protection and careful management to ensure that it remains on the landscape as declining snow levels and rising temperatures limit reproductive habitat. But in science, intuition and logic are never enough. Someone has to prove that the hypothesis is sound, and proof, in science, involves running the gantlet of the peer-review process, and publishing your results.
Jeff Copeland and his colleagues at the Rocky Mountain Research Station took on the task, and proposed an obligate relationship between wolverines and snow. Obligate relationships, in ecology, are relationships of dependence that are restrictive for the organism in question and that in many cases serve to define the environment on which it relies. So an obligate wolverine-snow relationship implies that you will never find a wolverine living in a place without persistent spring snow and, conversely, that if you see a mother wolverine traveling with two kits sometime in late spring, you know you’re in an area where snow persists until at least mid-May.
To test the model, Copeland and fellow researchers used remote-sensing data to construct a map of late spring snow spanning 7 years, from 2000-2006. To qualify as having persistent spring snowpack, a location (represented as a pixel on the map) had to remain snow covered between April 24 and May 15 – the period in which wolverine kits emerge from the den to begin the freewheeling life of a juvenile gulo – without a single day of bare ground.
Norway and Sweden maintain national wolverine den monitoring programs, which allowed Copeland to access precise den locations during the years for which the snow model was constructed. In North America, den location data are more scattered, so the authors of the paper drew on information spanning 1981- 2007 to obtain an adequate sample. In total, they compiled locations of 562 dens (327 of which were in Norway, 160 of which were in Sweden, 10 of which were in Finland, and 65 of which were in North America, which illustrates how little we know about wolverine reproduction in the US and Canada.) To look at year-round habitat use, Copeland compiled winter and summer telemetry locations of instrumented wolverines from 10 studies in North America and Norway. When the den locations were placed on the snow map, 98% of the dens were located on pixels that were classified as having persistent spring snow cover for at least one of the years, and 69% of the time, female wolverines were selecting for sites with snow cover for six or seven out of seven years. In the 2% of cases in which the den locations fell outside the snow map, the sites were investigated and determined to be snow dens; in these cases, the dens were located in expanses of snow too small to register by way of remote-sensing. Here was conclusive, statistically significant evidence that the relationship between wolverines and snow is, indeed, obligate.
The telemetry points reinforced the hypothesis. During summer, 95% of the telemetry locations adhered to the snow map, and in winter, 86% of the telemetry locations stayed within the bounds of persistent spring snowpack. The discrepancy makes sense if wolverines prefer snow; during winter, a greater portion of the landscape is snow covered and wolverines are therefore better able to travel outside the bounds of spring or summer snowpack. (Even then, however, Copeland determined that the wolverines traveling outside the snow map were primarily males – females maintained a higher fidelity to the snow map in all seasons.) Taken together, this meant that no matter how much snow was actually on the ground in a given season, wolverines predominantly operate in places where there is snow in late May.
Predominantly – but not exclusively. In one study, in the Omineca Mountains of British Columbia, wolverines actually appeared to be avoiding the snow locations during most of the year, but occupying those locations during summer when temperatures were highest. This suggested that there was another factor influencing wolverine habitat selection. To expand the test of whether wolverine distribution is limited by climate, Copeland and his co-authors decided to look at the next logical variable: temperature.
Parameter 2: Upper Thermal Limits
Wolverines are designed for the cold and snow, and early investigations into wolverine metabolism focused on the impressive insulating qualities of wolverine fur. Scientists suggested that a wolverine in a winter coat could tolerate temperatures down to -40º C. But no one had ever tested the upper limits of a wolverine’s thermal tolerance.
For Copeland and his colleagues, there was an easy way to construct a test, without ever handling a wolverine or its pelt. They modeled 50 years of temperature data in the locations where wolverines were hanging out in the telemetry studies, and established that the average maximum August temperature in these areas was 22° C. They then made a global map of 22°C maximum August temperature and laid it over the snow layer map. The temperature layer painted a cool swath across the boreal north and then, further south, splintered into peninsulas and islands that corresponded to mountains or cool maritime regions. In these insular and peninsular southern regions, the August temperature layer and the snow layer corresponded, but further to the north, where the entire landscape had maximum August temperatures of 22°C, there were regions with no snow in mid-May. The point of divergence between the snow layer and the August temperature layer was just south of the Omineca study area, at approximately 54N latitude.
The wolverine studies that Copeland included in the analysis crossed 24 degrees of latitude, with average temperature variations of about 10°C between the southernmost and northernmost locations. To the north, where average August temperatures are consistently below 22°C, wolverines operated across the landscape in a more general way. To the south, where average August temperature varies with latitude, they were selecting for locations with lower temperatures – generally, at higher elevations. Year-round, wolverines adhere to swaths or stretches or, further south, scraps and slivers, of boreal habitat.
‘Whales in the Desert’
The paper that Copeland and his colleagues wrote wields the weighty title The bioclimatic envelope of the wolverine (Gulo gulo): do climatic constraints limit its geographic distribution?, and was published in the Canadian Journal of Zoology in March, 2010. The paper answers the question posed in its title with a decided ‘yes.’ In a statistically verified way, wolverines have been shown to select for cold and snow, and to avoid areas that possess neither of these characteristics.
The paper and the science behind it are simple and elegant, and Copeland, in a phone conversation this morning, mentioned that the paper was ‘fun to write,’ because it dealt with three simple factors: areas of known wolverine presence, areas with persistent spring snowpack, and areas with a limited maximum August temperature. The first factor defined the geographic limits of the study, and this, in turn, increases the confidence that the results are valid. Referring to what he and his co-author Kevin McKelvey call a ‘whales in the desert effect,’ Copeland pointed out that in creating a habitat selection model for whales, you will find that whales are avoiding selecting deserts if you include deserts in your analysis. The problem with many habitat selection models is the fact that they are so broad and encompass so much territory that, by default, the ‘deserts’ are included, and you end up showing avoidance of areas or factors that might not actually be relevant to the needs of the species, but that might appear statistically significant in an analysis.
People have speculated about whether wolverines might have been widespread in the past, and were driven up into their current mountain habitat by expanding human activity. Copeland says that if he and his co-authors had included all of North America in their analysis, the results would have shown a statistically significant avoidance of human development and lowland areas. Because they confined their analysis to the known current habitat of the wolverine, and were able to show that within that habitat wolverines are making fine-scale selections for specific factors – snow, and low August temperatures – they were able to elucidate something truly significant for the gulo-curious and for managers: wolverines need cold and snow, and they can’t live in places that don’t have it. Post-Pleistocene, wolverines in the Lower 48 have probably always been confined to the mountains.
As Copeland explained the analogy over the phone, the idea of whales in the desert fused neatly with the vision that I hold in my head of the fragmented peninsulas and outlying islands of boreal habitat in the Rockies. I’ve always thought of it in terms of a land mammal having to swim across non-habitat between those islands – exhausting, but possible. But something about the idea of a whale trying to cross dry land seems apt, too. North of 54, the wolverine is swimming in the ocean for which it evolved. South of 54, the wolverine is a whale in puddle in a desert, contemplating survival in a landscape in which it survives by the grace of a cold and snowy climate, but to which it otherwise does not belong.