Results - the story so far

Research profile

My research centres on species interactions, population dynamics and community ecology. I investigate how top-down and bottom-up processes determine the abundance, dynamics and long-term trends of populations over large spatio-temporal scales. This includes top-down cascading impacts of changed species interactions, bottom-up cascading impacts of climate and land-use change, interactions between these processes, as well as resulting shifts between community states.

My research has revolved around the two fox species present in Fennoscandia; the red fox and the arctic fox. The red fox is a key predator in the Fennoscandian wildlife community, because it can limit and regulate a number of prey species and smaller predators. In contrast, the arctic fox is critically endangered, and red fox expansion in the alpine tundra is one of the main threats against the arctic fox.



Top-down or bottom-up? Cascades and ecosystem structure and function

In the role of top predators, large carnivores can exert top-down suppression of herbivore prey and smaller mesopredators. Hence, when large carnivores decline, herbivores and mesopredators might increase in abundance and this can have negative cascading effects on other species (Ripple et al. 2014). We have tested the hypothesis that nineteenth century persecution of Eurasian lynx and wolf in Sweden favoured the red fox through relaxed top-down control. Statistical modelling of hunting bags in 1828-1917 confirmed that high rates of decline in wolf and lynx were associated with high rates of increase in red fox. However, the results suggested that it was only in productive southern Sweden that top predators suppressed foxes to the extent that top predator decline led to mesopredator release of red foxes (Elmhagen & Rushton 2007). In a collaboration with Finnish scientists, we explored the relationship between lynx and red fox further using recent monitoring data (1989-2005). Classical ecological theory hypothesises that large carnivore suppression of herbivores should strengthen with productivity, but also that the resulting ecosystem structure - where herbivore biomass is limited top-down - should not be affected by interactions within the predator guild. We showed that lynx recolonisation in Finland caused a shift from bottom-up to top-down limitation of fox. Moreover, and in contrast to classical theory, this caused a shift from top-down to bottom-up limitation of shared herbivore prey (mountain hare). The top-down impacts were mediated by productivity. Hence, we suggest there are 'interference ecosystems' where predator interference interactions and productivity determine ecosystem structure (Elmhagen et al. 2010). 

A continental-scale analysis of red fox abundance in northern Eurasia showed that presence/absence of lynx is associated with shifts between top-down and bottom-up limitation of red fox also at this large scale. This suggests that in the absence of lynx, fox density is more prone to respond to relaxed bottom-up limitation e.g. due to climate warming (Pasanen-Mortensen et al. 2013). In the absence of lynx, the red fox is also favoured by croplands, most likely because they offer a productive habitat with high prey availability (Pasanen-Mortensen & Elmhagen 2015). It has been suggested that top-down effects of large carnivores can buffer bottom-up driven change. Returning to south-central Sweden, we explored this hypothesis using present-day associations between red fox abundances and its drivers to project changes in fox abundance over 200 years. The results indicated that lynx restoration to nineteenth century levels would not be sufficient to restore fox abundances, as the positive effects of observed climate warming and land-use change remained. Similarly, future projections indicated that lynx abundance would have to increase by 79% by 2050 to compensate for expected climate-driven increases in fox abundance. We conclude that top-down interactions in theory can buffer bottom-up change, but compensatory changes in apex predator abundance are required. Thus, top predator restoration to historic levels would not be sufficient to compensate for the widespread relaxation in resource constraints which has favoured many now abundant herbivores and mesopredators (Pasanen-Mortensen et al. 2017).

Our research on lynx and red fox is part of an increasing number of studies which show that top predators can carry out important top-down ecosystem functions. However, it also highlights the complexity of ecological interactions, where the effect of top-down limitation can vary with e.g. ecosystem productivity (Ritchie et al. 2012; Ripple et al. 2014) and the degree of human land-use (Kuijper et al. 2016). To understand interactions between changes in top-down and bottom-up processes associated with e.g. climate and land use is a challenge for the scientific community and biodiversity conservation (Elmhagen et al. 2015 in Ecology and Society; Kuijper et al. 2016). The context dependence of top-down effects should also be taken into account in ecosystem restoration programmes, and there is a need to direct the conservation focus more on predator functionality in specific systems. However, these ecosystem restoration programmes also need to take into account the economic and social impacts of large carnivores (Ritchie et al. 2012). From a socio-economic perspective it is important to find reasonable trade-offs between the pros and cons of large carnivores, as well as to find solutions for human coexistence with large carnivores (Ripple et al. 2014).

In unproductive northern ecosystems, large carnivores are naturally rare and cannot suppress smaller predators. The abundance and dynamics of smaller predators may therefore depend primarily on bottom-up processes (Elmhagen & Rushton 2007). The scavenging wolverine may benefit from ungulate carcasses provided by larger carnivores as long as interference levels are low (Khalil et al. 2014). Several medium-sized predators depend on the bottom-up resource pulse provided by high-abundance peaks of cyclic rodent populations for reproduction. These peaks generally occur every 3-5 years (Elmhagen et al. 2011), and the Norwegian lemming for example, then become highly abundant also outside its optimal habitats (Le Vaillant et al. 2018). The rodent cycles have periodically faded in the 20th century, likely as an indirect effect of climate change (Elmhagen et al. 2011; 2015 in Ambio). Furthermore, several northern species have retreated north or decreased, while several southern species (including red fox) have advanced north or increased. Thus, northern ecosystems are changing, but not only due to climate change. There appear to be several potential drivers of change, e.g. climate, land-use, infrastructure and hunting, which all seems to have favoured southern species more often than northern species (Elmhagen et al. 2015 in Ambio). One example is observed advances in red fox distribution and abundance in parts of the Arctic tundra, which can be driven by climate warming in productive tundra, but locally also by anthropogenic subsidies. The latter can allow red foxes to establish north it its climate-imposed distribution limit. We suggest that such synergies between anthropogenic subsidies and climate warming can speed up Arctic ecosystem change (Elmhagen et al. 2017).



Fox ecology in alpine tundra

The arctic fox is a small (3-4 kg) carnivore with a circumpolar distribution in the Arctic. The arctic fox was abundant in Fennoscandian alpine tundra in the 19th century, but declined drastically in the early 20th century. It was protected in Sweden in 1928, in Norway in 1930 and in Finland in 1940, but never recovered. A number of possible reasons have been suggested. For example, that the arctic fox has suffered from food shortage, increased competition from an expanding red fox population and Allee effects.

In the early 2000s, the distribution of arctic foxes in Scandinavia had been fragmented into four genetically distinct populations (Dalén et al. 2006), which indicates that the arctic fox population in Scandinavia cannot function as a metapopulation, as previously suggested (Elmhagen & Angerbjörn 2001). The arctic fox also appeared to suffer from negative Allee-like effects of low connectivity between remaining habitat fragments (Herfindal et al. 2010). 

The arctic fox breed in pairs or groups of related females (Norén et al. 2012, Elmhagen et al. 2014). Its primary prey are small rodents (lemmings and voles) which generally display cyclic dynamics with high-abundance peaks every 3-5 years (Elmhagen et al. 2000; 2011). The arctic fox is adapted to maximise its reproductive output during the increase phase of the rodent cycle, when cub survival is highest (Meijer et al. 2013). However, the rodent cycles have shown periodic disruptions in the twentieth century, with long periods without high-abundance peaks (Elmhagen et al. 2011; 2015 in Ambio). These fading cycles has had a negative effect on the arctic fox (Angerbjörn et al. 2013), and several other rodents specialist predators (Elmhagen et al. 2015 in Ambio).

The red fox has increased in northern Sweden and the alpine tundra since the nineteenth century (Elmhagen et al. 2015). Arctic and red foxes have similar food niches, which suggest they compete for territories (Elmhagen et al. 2002). Spatio-temporal patterns in habitat use show that arctic and red foxes overlap in winter, but not in summer (Dalén et al. 2004). Spatio-temporal patterns of den use in summer show that arctic foxes rarely breed close to dens occupied by reproducing red foxes. On the few occasions when arctic foxes do breed in the vicinity of red foxes, a high incidence of red fox predation on arctic fox cubs suggests that arctic foxes should actively avoid establishment in red fox territories (Tannerfeldt et al. 2002). This can explain observed changes in arctic fox distribution over the last century. Arctic fox reproduction became restricted to high-altitude dens in the least productive parts of the species’ former range, whilst red foxes now occupy the more productive tundra in lower altitudes (Dalerum et al. 2002, Herfindal et al. 2010). Simulations of arctic and red fox interactions in a spatially explicit population model confirm that arctic fox avoidance of red fox territories during reproduction is sufficient to have a negative impact on arctic fox population trends (Shirley et al. 2009).

A review on interactions between arctic and red foxes in the Arctic show that red foxes, when present, exclude arctic foxes from dens, habitat and food resources, and that red foxes sometimes kill arctic foxes. Where red foxes reach ecologically effective densities, e.g. in Fennoscandia, they can cause arctic fox decline and extirpation. Red fox establishment and increase can be driven climate change and/or anthropogenic subsidies. In Alaska, subsidies alone has led to red fox establishment and a consequent decline in the arctic fox population around Prudhoe Bay, suggesting that subsidies can allow red foxes to establish north of its climate-imposed distribution limit (Elmhagen et al. 2017).

The arctic fox population in Sweden, Norway and Finland has been small since the early-to-mid 1990s, and declined further during a period of disrupted lemming cycles from the early 1980s until 2000. The last reproduction in Finland occurred in 1996 when the Finnish breeding population went extinct. The Swedish-Norwegian population was estimated to only 40-60 adults in 2000. However, the population increased fourfold in 2001-2010. The increase was driven by intense conservation actions, red fox control and supplementary feeding, in two subpopulations (Angerbjörn et al. 2013).

In the Helags mountain region in Sweden, there are guided tours to an inhabited arctic fox den. A survey show that participation in the tour improves the knowledge on the status of the arctic fox, as well as how one avoids disturbing the foxes. The revenues from the tours are donated to local conservation actions, and because the current level of disturbance on the foxes by the tours seems low, we suggest the net impact on the local arctic fox population is positive (Larm et al. 2018).

Read more: the web page of the Arctic Fox Project.