The Rocky Mountain Parnassius Problem

Lately, I've been thinking about butterflies. I won't subject you to the interesting, if slightly convoluted, train of thought that led me to today's paper (this post is long enough as it is), but suffice it to say that we are back on the topic of butterflies and climate change. If you remember, back in March I wrote about a paper that explored how a single climate parameter can determine population dynamics in a butterfly species, the Mormon Fritillary (Speyeria mormonia) - An Early Spring Isn't Always a Good Thing. In that case, it was how snow melt time in the first year would affect butterfly fecundity through flower abundance.

Along these lines, a preprint in the journal Ecology takes a look at how regional climate, particularly winter and winter extremes, affects annual rates of population change. We know that climate change is causing range shifts in many species. Good examples of this can be seen in high elevation, typically mountainous regions. The idea here is that a warmer climate facilitates growth in areas where a colder climate had previously prevented growth. However, this warming trend is not the only prediction attached to climate change. Variability in climate and weather and the extremes of seasons and events are also expected to have a large impact on ecological processes. This means that not only do species have to respond to general climate warming but also to general and local extremes. Long-lived vertebrate species with overlapping generations may be buffered to this because such these extreme changes act primarily on a single age class or cohort. Short-lived, univoltine (one brood or generation per year) ectothermic species have little to no buffering, meaning the entire population is affected by these extreme events.

The authors of this study use long-term (15 year) estimates of population size for 21 subpopulations of the Rocky Mountain Apollo butterfly (Parnassius smintheus Doubleday) in Alberta, Canada. This species is common in the alpine meadows of the Rocky Mountains of North America. They are known to overwinter as pharate larvae inside the egg, hatching in May, feeding on their obligate host plant (Sedum lanceolatum), pupating in late June, emerging as adults in late July, and the females ovipositing on their host plant through August. Although they are common, they tend to occur in relatively small subpopulations, having limited dispersal, which makes them good for metapopulation studies and studies of local changes. The researchers estimated population size in each subpopulation using mark-recapture data. The climate variable they chose was the Pacific Decadal Oscillation (PDO) index, an index shown to have strong correlations with their chosen study site. This index “contrasts the spatial distribution of sea temperatures between the northeastern and northwestern Pacific Ocean after correction for mean global temperature…providing a single integrative measure of climate across western North America through its strong temporal correlation with both temperature and precipitation.” A positive PDO means that warm water is along the coast and are associated with warm, dry years inland. A negative PDO means that cooler water lies along the coast and are associated with cool, wet years. They used both annual PDO as well as seasonal PDO values corresponding to stages of the life-cycle that were of particular interest. Then they ran some models that I won’t go into (I’ve used up a lot of space and I haven’t even gotten to the results yet!).

These models showed that “more frequent climate extremes pose important consequences or animal population growth affected by climate.” They found that winter values of the PDO were a strong predictor of annual population growth. The effects of climate in these butterflies was found to be curvilinear wherein both extremes (too warm and too cold) result in population decline. This suggests that the variability and extremes predicted by climate change models will greatly affect the population dynamics of species such as this and that there may be less opportunity for them to adapt to general climate warming as the occurrences of these extremes increases. Additionally, the curvilinear nature of these results suggests some complications in the mechanisms involving range shifts. Their data support range shifts (either poleward or elevational) in that climate warming may sustain a positive population growth, although low latitude and low-elevation range margins might be affected more causing negative growth.

Are these results applicable to all species? No. P. smintheus is an alpine species that is naturally subjected to a cold, unpredictable environment, and, as such, they exhibit several behavioral, morphological and physiological adaptations. This means that curvilinear results of the model suggest multiple climate-related factors that need to be teased out (temperature, precipitation, snow cover, snow distribution, etc.) and that the PDO index itself may have a range with extreme values on its edges. Because these are extreme factors rather than just gradual shifts in climate, conservation planning could be more difficult over the long term. The extremes themselves decrease populations and the variability shrinks geographic ranges (depending on event and climate interactions) also causing decreases. Perhaps helping to curtail the effects of the short-term weather extremes may help in the long-term. As yet it is unknown, and, as with most science, needs more investigation.

I encourage you to read the entire paper. There are additional ideas and fleshing out of these conclusions that are particularly interesting.

Roland, J., & Matter, S. (2012). Variability in winter climate, and winter extremes, reduce population growth of an alpine butterfly Ecology DOI: 10.1890/12-0611.1


There are also a couple of articles that have nice interviews with the authors:
From EurekAlert! and the University of Alberta: "Climate change isolates Rocky Mountain butterflies"
ScienceDaily's article: "Climate Change Isolates Rocky Mountain Butterflies" 


(image via GeoLocation)