If pop culture is any indication, we hold a special place in our heart for penguins compared to the many other denizens of the animal kingdom.
Maybe it is something about the close-knit familial communities that have been portrayed in documentaries such as March of the Penguins, illuminating the long treks penguins make to breeding grounds and the gritty survival through gusty winters to keep their eggs safe and warm. Or maybe it’s their endearing waddle and playful belly slides into the ocean. Or some combination of this bravery humor that makes us see a little of ourselves in an animal that could not be more outwardly different.
I am sure we would like to think these birds are surviving happily in their wintry tundra to continue to admire from afar for generations to come. They are a top predator in Anartactica after all, the occasional leopard seal an occasional threat and the harsh winters their central nemesis. But new research in Nature Climate Change has exposed a new enemy, using stochastic modeling to predict how climate change and sea ice loss will affect emperor penguin populations. The news is not good. This story has been picked up by the mainstream media – as seen here, and at NPR, for example – but I thought I’d go into a little more detail about what the modeling and results actually involved.
First, why is sea ice availability so important to emperor penguins? The reasons are many, including ice cover impacting the entire food web, but most important is the penguins’ use of ice as a breeding ground. Previous research has already shown that low sea ice concentration is correlated with lower survival due to lower food availability. Interestingly, too much sea ice also worsens survival due to longer foraging trips during breeding season, and reduced chick feeding. So there is a Goldilocks zone that is just right for penguins – another good example of the delicacy of ecosystems, even in frigid Antarctica.
So we know sea ice is important, but no research (until now!) has yet been done to predict how decreased sea ice levels from climate change will affect penguins. This is crucial information because species vulnerability, and hence conservation funding and efforts, is determined by predictive population dynamics. To remedy this, Jenouvrier et al gathered information about sea ice around 45 different emperor penguin colonies on the Antarctic continent – most of these have not been visited by humans but are known to exist from satellite data. Most direct research has only been done on one penguin colony in East Antarctica, so the research team used previous population dynamical data from that colony, combined it with climate change models included in the IPCC report, and developed a predictive model looking at populations of each of the 45 colonies separately.
Here’s how the model generally works: detailed data from the East colony provides robust information about how levels of sea ice affect that penguin population, The climate change model data (technically known as an atmosphere-ocean general circulation model) predicts how much sea ice will change near each colony over the next century. This climate information is then fed into the model relating sea ice loss/gain to penguin populations – the output then is a prediction of population change for each individual colony! So the researchers plugged all this data in and predicted population changes by 2100. (One note – these types of predictive models of climate deal with uncertainties that increase the farther ahead you wish to forecast. To deal with this, the researchers calculated median population changes from many different simulations with varied initial conditions – this type of stochastic method provides a confidence interval for results as an indication of the variability in the predicted estimate. You can see this in the graph of breeding pair changes a little farther below.)The results from the model, shown above, are extremely worrisome and are hopefully a call for new forms of conservation efforts. Each circle overlaid on the Antarctic map depicts a colony, numbered 1 to 45. The scale on the right and the corresponding colors on the map indicate the percent change in sea ice projected by 2100 – whiter colors indicate much greater change, up to 20 percent. Red colonies are predicted to be near extinct, orange are endangered, yellow are vulnerable, and green are not threatened (remember, this is 2100, so there’s time to do something). A couple take-home points…
1)Notice the correlation between sea ice loss in the Indian Ocean and Weddell Sea areas and the number of near-extinct and endangered colonies. The researchers found a clear connection between variability in sea ice loss and colony population loss.
2) 75% of colonies are expected to be vulnerable to sea ice change, and 20% will be near-extinct.3) Breeding pairs stay stable until around 2050, at which point the number of pairs begins to decrease and growth rates decline. This gives a general timetable for when climate effects become severe enough to affect population dynamics. It’s important to note the uncertainty, however – the red line is the median and the gray area is the 90% confidence interval This means that 90% of simulations predicted breeding pair numbers within this shaded region. Notice also that the gray area expands with time – this is the inevitable effect of entropy! As stated above, the farther from the present we predict in time, the less we can know with great certainty, leading instead to great variability in predictions. But the median does seem to follow the curve of the lower bound of the confidence interval, indicating that half of the simulations fell in that narrower regions between the median and the bottom bound. This indicates that there’s much more variability in the simulations above the median.
Also remember that these predictions are based on the relationship between sea ice and population levels for one colony, extrapolated to the rest. This does introduce some error, which the researchers admit, but they note that it errs on the side of underestimating population loss and overestimating growth. This is mainly because the base penguin colony used is in an optimal region that has not experienced drastic changes in sea ice levels. Therefore, their population has been relatively stable with respect to sea ice variability and may not elucidate stronger, nonlinear relationships that occur once sea ice loss passes a certain threshold. Thus, these reported predictions are conservative estimates that still are quite shocking and need a response.
All in all, the simulations predict a population decline of 78% across three penguin generations. Endangered species labeling is based on population dynamics, mainly, so this is a great first step to call attention to emperor penguins as a vulnerable species.
Jenouvrier, S., Holland, M., Stroeve, J., Serreze, M., Barbraud, C., Weimerskirch, H., & Caswell, H. (2014). Projected continent-wide declines of the emperor penguin under climate change Nature Climate Change DOI: 10.1038/nclimate2280