Global warming strains at species interactions

Climate change is like a stalking predator, a threat that first crept up, and then swiftly leapt out at the ecological science community. There is no doubt it was an issue around which there was a simmering awareness for decades. However, recent detailed multidisciplinary studies, which have pored over numerous long-term datasets (most compiled for reasons unconnected to climate change monitoring), have forced a re-appraisal of the magnitude and pace of the challenge that global warming represents to both geophysical systems (e.g., ice sheets, mountain glaciers, sea level) and to biodiversity in already stressed environments.

One of the clearest ‘attribution fingerprints’ of global warming on biological systems is the advance of reproductive events to earlier in the breeding season. Species that take their cue to mate or migrate from climatic phenomenon (e.g., average or extreme temperature, or rainfall), rather than non-climatic triggers (e.g., length of day), are predicted to respond most readily to climate change.

For most species, this is exactly what is seen, with the most commonly observed changes manifested in the advance of spring events, delay of autumn events, shortening of life-stages and changes in migration patterns.

However, this is not a universal response — which raises a number of interesting possibilities. Perhaps some (mostly larger-bodied) species are unable to adapt quickly enough, due to long generation lengths? Others, especially those threatened by other human-induced factors such as habitat loss or over-exploitation, may lack sufficient standing genetic diversity or behavioural flexibility to respond effectively. In other cases, there may be advantages in quite different types of adaptation (e.g., breeding earlier to keep up with the food supply, or breeding later and having less food, but avoiding the breeding season of the predator).

I recently wrote an In Focus review for Journal of Animal Ecology, which took a reflective look at a recent paper published in the journal, which considers some of these issues of ecosystem adaptation to climate change (the above and below are edited excerpts from it). The excellent paper I describe is by some Dutch ecologists who have a long-established track record studying the ecosystem of Hoge Veluwe National Park in the Gelderland province of the Netherlands.

The system is a European mixed woodland food web. Oak and pine dominate the study site. During bud burst, the deciduous oak trees are food for moth caterpillars, which are, in turn, predated upon by a number of bird species. These songbirds, and especially their fledglings, are targeted as prey by Eurasian sparrowhawks. Two decades of data (1985 to 2005) have been collected from this community, monitoring everything from flowering and breeding times to the birth and death rates of individual species. Previously published work has shown that the component species are individual reacting to climate change.

The recent work is particularly interesting. It goes beyond a simplified one- or two-species response (whereby, for instance, the timing of peak caterpillar numbers induces songbirds such as pied flycatchers and blue tits, to breed earlier to keep up the food supply). Instead they consider adaptation pressures and inter-species relationships, from caterpillars grazing on oak, songbirds eating caterpillars, and predatory sparrowhawks picking off the songbirds. Quite a complex set of interactions, as is typical of ecosystem studies.

The researchers showed that all components of this system: tree budburst, peak caterpillar numbers, songbird hatching dates, and sparrowhawk hatching dates, are happening earlier and earlier over time. This statistical analysis shows the smallest response was for the sparrowhawk (0.9 days advance in breeding per decade), then budburst (almost twice that rate, at 1.7 days), a big change for songbirds (3.5 days for coal and great tits and 5 days for flycatchers and blue tits) and a whopping 7.5 days per decade for the caterpillars. At that rate, the peak of caterpillar numbers will occur a full month earlier within the space of 40 years – and that’s at the current pace of warming, which is expected to accelerate considerably without emissions reductions.

What would a decade worth of change do? Well, the peak of caterpillar numbers would have advanced almost 8 days, the songbirds which feed on them only half that, and the sparrowhawks a mere 1 day. Over time, this would put strain on the species interactions, and would inexorably pull the species interactions apart, due to mismatches in the timing of breeding and maximum food availability. This could, in turn lead to outbreaks in caterpillar numbers, because they are being fed to fewer songbird chicks. More caterpillars would mean more of the oaks are heavily grazed, which could damage the forest structure. The ecosystem starts to unravel…

So what might these different rates among the different component species of this food web indicate? It’s difficult to be sure, but the authors have some interesting ideas.

One possibility is that the larger species, because of their longer generations, aren’t able to adapt to the changes in climate as quickly as the fast breeding species. Another is what’s termed ‘weak predictability’ — animal’s decision-making (when the birds choose to lay their clutch) is simply too far separated in time from the key selection environment (when the caterpillars peak – driven by climate warming). So they can’t react to cues because they’d have to have some window into the future (the build-up of caterpillar eggs, or perhaps the size of the previous season’s crop of grubs might do it, but this type of forecasting would take considerable time to evolve).

Or perhaps there is a strong pressure from natural selection for prey to outpace their predators by breeding earlier (since climate conditions now allow this). This would explain why the caterpillars have advanced more rapidly than the tree budburst (they trade-off having less available food because this also reduces the number of them that end up getting eaten by songbird chicks). The songbirds are probably trying to both keep up with their caterpillars and escape being eaten by sparrowhawks, and so are adapting as fast as they can (which is slower than the caterpillars, who have shorter generation times and enhanced growth rates under warmer conditions). The sparrowhawks eat many species besides songbirds, and so the pressure on them to respond to indirect changes in caterpillars and budburst is quite weak. It’s an intriguing idea.

As I’ve noted previously on BNC, the pressures faced by biodiversity are multiple and they tend to reinforce each other to make the whole situation that much worse. For instance, heatwaves and droughts tend to exacerbate forest fires, and habitat fragmentation can mean that species find it difficult to move through urban and agricultural landscapes to reach new areas that are become climatically suitable to them as the average temperature warms. They get anchored in patchy environments that become increasingly unsuitable for them over time.

We need more studies like this, to help us better understand the variety of responses ecosystems are likely to experience under global change (climate warming and other human impacts such as deforestation and the spread of invasive species). Certainly there are still large gaps in knowledge, especially in areas such as southern hemisphere biological communities (yes, even Australia). But I suspect the answers are out there, buried in the literature and in government and research organisation databases, just waiting to be properly analysed to tease out the climate change signals. My team of postdocs and PhD students at University of Adelaide are hard at this task.

Stay tuned, I’ll certainly be reporting back here in the future about what we found!