Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Fear or food – abundance of red fox in relation to occurrence of lynx and wolf


Apex predators may affect mesopredators through intraguild predation and/or supply of carrion from their prey, causing a trade-off between avoidance and attractiveness. We used wildlife triangle snow-tracking data to investigate the abundance of red fox (Vulpes vulpes) in relation to lynx (Lynx lynx) and wolf (Canis lupus) occurrence as well as land composition and vole (Microtus spp.) density. Data from the Swedish wolf-monitoring system and VHF/GPS-collared wolves were used to study the effect of wolf pack size and time since wolf territory establishment on fox abundance. Bottom-up processes were more influential than top-down effects as the proportion of arable land was the key indicator of fox abundance at the landscape level. At this spatial scale, there was no effect of wolf abundance on fox abundance, whereas lynx abundance had a positive effect. In contrast, at the wolf territory level there was a negative effect of wolves on fox abundance when including detailed information of pack size and time since territory establishment, whereas there was no effect of lynx abundance. This study shows that different apex predator species may affect mesopredator abundance in different ways and that the results may be dependent on the spatiotemporal scale and resolution of the data.


Apex predators may have multiple top-down effects on lower trophic levels in the ecosystem. Classic ecosystem theory predicts that predators control herbivores, restricting their grazing effect on the vegetation and thereby keeping “the world green”1. However, the discovery that predators kill, and even consume, not only their herbivore prey, but also members of other competing predator species within their own guild2,3,4, introduced more complexity to ecosystem theory, and supported the Mesopredator Release Hypothesis5. According to the Mesopredator Release Hypothesis, the human-caused worldwide reduction of apex predators has not only reduced their impact on herbivores, but also “released” mesopredators from top-down control, allowing them to thrive. The empirical evidence for negative effects of apex predators on mesopredator abundance is strong in both aquatic6, 7 and terrestrial systems8,9,10,11,12. Although there is ample anecdotal evidence of apex predators killing mesopredators4, 13 there are few studies that have been able to demonstrate that this mortality is additive and strong enough to limit mesopredator population growth14. It has been suggested that fear of being killed makes mesopredators avoid not only direct proximity to apex predators, but also the habitats used by them which in turn may influence the demography of mesopredator populations14.

There may also be reasons for mesopredators to be attracted to apex predators. Most vertebrate carnivores are not only predators, but also facultative scavengers, and scavenging is a prevalent phenomenon in most terrestrial systems15. There is an increasing amount of literature on the role and importance of scavenging15,16,17,18, some even suggesting that scavenging may buffer the effects of climate change19. Much of the carrion available to scavengers is provided by apex predators because of their habit of killing large prey, which they often cannot completely consume by themselves20,21,22,23,24. Thus, mesopredators are confronted with a dilemma: should they avoid these larger guild members to avoid being killed, even if this means missing scavenging opportunities, or should they take the risk to obtain a free meal? We studied this question in a boreal ecosystem in Sweden where red fox (Vulpes vulpes, hereafter fox) was the dominant mesopredator, and Eurasian lynx (Lynx lynx, hereafter lynx) and grey wolf (Canis lupus, hereafter wolf) were the apex predators.

Both lynx and wolves have been extirpated from most of Europe for several centuries, but recently they have followed a world-wide pattern of large carnivore recovery25. The lynx returned to south-central Sweden in the 1950’s after 50 years of absence26, and the wolf returned in the 1980’s after having been absent for more than a century27. During this time, the fox has remained ubiquitous throughout Europe, including Sweden28. Fox abundance and dynamics are linked to the amount of arable land and to the abundance of their main prey, voles (Microtus spp.)28, 29. For the fox, ungulate remains are also an important food source, especially when accessibility of other food sources is low30, 31. When vole abundance is low, foxes may increase their use of ungulate carrion in forest habitats32. The consumption of remains from wolf-kills is highest during late winter and spring33, 34. Increased food availability during winter has been shown to positively impact the reproductive success of fox35. In Sweden, the fox is the most frequent scavenger of wolf-killed moose34 (Alces alces) and also commonly scavenges remains from lynx-killed roe deer (Capreolus capreolus)36.

We tested for the two alternative effects of lynx and wolf occurrence on fox abundance at the landscape level (i.e., negative due to higher risk of intra-guild predation or positive due to more scavenging opportunities), while accounting for the known effect of land composition37 and vole density35. In the next step, we limited our analyses to the wolf territory level and predicted that fox abundance would be negatively related to wolf pack size because more wolves may result in less availability of carrion biomass38 and higher predation risk. Finally, we predicted that the negative effect of wolves should be most pronounced during the first years after wolf establishment, because wolves may constitute a novel source of mortality for the fox.


Landscape Level

Fox abundance increased with increasing proportion of arable land, and this variable was the most informative parameter explaining the variation in fox abundance at the landscape level (RVI = 1.0, Fig. 1a and Tables 1 and 2). Fox abundance at the landscape level was also positively related to lynx abundance (RVI = 0.78, Fig. 1b and Tables 1 and 2). There was no effect of wolf occurrence (independent of method used to classify wolf occurrence) on fox abundance (Fig. 2a and Table 1) at the landscape level, nor was there any support for models including the vole density index (Table 1).

Figure 1

Abundance of red fox (crossings 10 km−1 24 h−1) in relation to (a) arable land (%) and (b) lynx abundance (crossings 10 km−1 24 h−1). Model averaged predictions derived from the highest ranked models (Table 1) are shown (solid lines) with associated 95% confidence intervals (grey shading). Predictions for each explanatory variable were made while holding the other variables constant at mean values (arable land 8% and lynx abundance 1.5 crossings 10 km−1 24 h−1). Red fox abundance predictions were back-transformed to their normal scale for the figure. For model averaged parameter estimates see Table 2. Red fox and lynx abundance were surveyed by snow-tracking in wildlife triangles in south-central Sweden from 2001–2003.

Table 1 Highest ranked candidate models relating red fox abundance (crossings 10 km−1 24 h−1) to arable land (%), vole density index (low or high), lynx abundance (crossings 10 km−1 24 h−1), and wolf occurrence.
Table 2 Model-averaged parameter estimates with standard error (SE) for each variable retained in the best models (ΔAICc ≤ 2) in Table 1.
Figure 2

Upper panel (a) shows mean (±SE) red fox abundance (crossings 10 km−1 24 h−1) in relation to a 4-level wolf category: (I) inside an observed wolf territory, (II) inside an average wolf territory, (III) inside a maximum wolf territory, and (IV) outside a wolf territory. Lower panel (b) shows mean (±SE) red fox abundance from wildlife triangles located inside estimated wolf territory borders (i.e., wolf category I) in relation to time since territory establishment (years) where year 0 is the year of territory establishment by a wolf pair. Red fox abundance was surveyed by snow-tracking in wildlife triangles in south-central Sweden from 2001–2003.

Wolf Territory Level

At the wolf territory level there was a weak negative relationship between fox abundance and wolf pack size (RVI = 0.29) in combination with the proportion of arable land, although arable land and vole density index were the best predictors of fox abundance (RVI = 0.56 and 0.53, respectively, Fig. 3 and Tables 1 and 2). Including time since territory establishment in the analyses showed that there was a negative temporal effect of wolves on fox abundance (RVI: time since territory establishment = 0.57, vole density index = 0.49, arable land = 0.47). Fox abundance decreased 1–2 years after wolf territory establishment and after these first two years fox abundance rebounded again, although to a slightly lower level than at the year of wolf territory establishment (Fig. 2b). In contrast to the landscape level, there was no relationship between fox abundance and lynx abundance at the wolf territory level (Tables 1 and 2).

Figure 3

Red fox abundance (crossings 10 km−1 24 h−1) in relation to wolf pack size, arable land (%) and vole density index (low or high). Model averaged predictions derived from the highest ranked models (shown in Table 1) are shown for different wolf pack sizes at (a) high and (b) low vole density index, while holding the arable land constant at the mean value (2% within wolf territory borders, solid line). Dashed lines indicate red fox abundance when arable land varies with ±1 SE (±0.3). Red fox abundance predictions were back-transformed to their normal scale for the figure. Red fox abundance was surveyed by snow-tracking in wildlife triangles in south-central Sweden from 2001–2003.


Although both wolves and lynx had a top-down effect on fox abundance in our study area during winter, bottom-up processes were more influential, as the proportion of arable land was the key indicator of fox abundance. The almost complete loss of apex predators during the 19th century, in combination with geographical variation in ecosystem productivity, has been shown to be a key factor in the fox population increase in Scandinavia8. The agricultural landscape holds the highest densities of foxes28. Also, lynx in Scandinavia are attracted to agricultural fields because their main prey, roe deer, are clumped around this habitat39,40,41. Wolf territories, on the other hand, are composed by a small amount of open areas, such as agricultural fields, suggesting an adaptation to avoid humans42. Thus, in our study area, fox and lynx likely spatially overlap more than fox and wolves. The avoidance of open areas by wolves is further supported by the fact that the proportion of arable land in wildlife triangles inside wolf territories averaged 2% (range 0–20%), compared to 8% (range 0–76%) in the whole study area. The importance of vole density on fox abundance may be more pronounced in less favorable habitats29. Thus, the lower proportion of arable land within wolf territories may explain why the vole density index affected fox abundance at the wolf territory level, but not at the landscape level. Consequently, the lowest fox abundance was inside wolf territories that had a low proportion of arable land and low vole density index. This supports the conclusion that bottom-up processes were an important factor affecting fox abundance.

At the landscape level, there was a positive relationship between fox and lynx abundance. Simulations where we added observation uncertainty to the lynx abundance measurements (i.e., coefficient of variation ranging from 0.2 to 3) revealed that our results were robust; it was a high probability (i.e., ˃97%) of a positive relationship between fox abundance and lynx abundance (see Supplementary Methods for further details). This result was surprising since several studies have shown a negative relationship between these two species8, 9, 28, 43. For example, Elmhagen & Rushton8 used historic data to demonstrate a mesopredator release of fox after lynx and wolves were exterminated from Sweden, indicating that foxes had been top-down limited before the extermination of these two apex predators, and Elmhagen et al.9 found that this process was reversed when lynx recolonized parts of Finland in recent time. Our data did not allow us to determine whether the positive effect of lynx abundance on fox abundance in our study was a demographic or a behavioral response. Although previous studies have suggested a negative demographic effect of lynx on fox populations8, 9, 13, we suggest that the effect we observed was a behavioral response by fox to lynx abundance. Foxes are attracted to sites with olfactory cues from lynx, a behavior that should increase the possibility to access carrion from lynx killed prey44. However, from lynx scent marks, a fox might also gain information about the existence of lynx in its home range, a knowledge that may be used to adjust behavior in order to reduce the risk of encountering lynx44. Studies of multiple predator sympatry in African and Asian ecosystems have shown fine-scale behavioral tactics and trade-offs associated with resource availability45,46,47. Hence, we suggest that the positive relation between fox and lynx abundance indicates that foxes are attracted to lynx to increase their chances to find carrion, possibly facilitated by using a fine-scale spatiotemporal segregation strategy to minimize the risk of being killed. Two alternative hypotheses are that 1) both lynx and fox are attracted to a third, currently unknown, factor, or 2) species within the same mammalian family are likely to compete more than species in different families48, which causes fox to avoid wolves and not lynx. In any case, our findings contradict the hypothesis that the avoidance of apex predators could suppress mesopredator abundance more than predator-induced mortality14, at least with respect to fox and lynx in our study area.

Whereas we found no effect of wolves on fox abundance at the landscape level, we show a negative temporal effect of wolf occurrence at the wolf territory level. This effect was most pronounced the first two years immediately after wolf territory establishment. We suggest that this negative effect is explained by a combination of the novelty of wolf occurrence and their local abundance, i.e., pack size. Zimmermann et al.38 showed that large wolf packs (6–9 wolves) in Scandinavia had a kill rate that provided just enough, or even less, prey biomass than required, while small packs (2–3 wolves) had a kill rate that provided more prey biomass than required, thereby leaving a larger surplus of carrion. Furthermore, the probability that foxes encountered wolves increased in large pack territories due to the tendency of large wolf packs to split up in sub-groups during winter49. The establishment of a wolf territory usually starts with a breeding pair, followed by an increased number of wolves the consecutive winter after the pair’s first reproduction. Fox abundance one and two years after wolf establishment was reduced by 50% compared to the establishment year and with triangles outside wolf territories (wolf category III and IV in Fig. 2a). The reason that we did not see any negative or positive effect during the establishment year could be either that the pack consisted only of the establishing pair, i.e., the smallest possible pack size, and/or that the pair was formed late in the winter. Consequently, we suggest that the pronounced negative effect on fox abundance the years immediately following wolf establishment was due to an avoidance reaction by foxes to the novelty of wolf presence which faded over time, or the increased mortality of naïve foxes. We suggest that a demographic effect is less likely than a behavioral effect as wolf scat analyses indicate that intra-guild predation by wolves on foxes is low in Scandinavia (<0.2% of wolf scats (n = 2091) contained hair from fox)50, 51. However, if wolves kill foxes primarily to reduce competition (i.e., foxes are not consumed), wolf scat analysis might underestimate intra-guild predation. But studies of wolf predation in Scandinavia show that few wolf-killed foxes are found nearby the remains of wolf-killed ungulates52, 53 which also indicates low intra-guild predation.

We argue that food competition is not a plausible cause for the negative effect of wolves on fox abundance. Wolf occurrence results in increased carrion availability, a food source frequently utilized by foxes in our study area34. Although wolves occasionally kill smaller prey species (hares (Lepus timidus, Lepus europeus) and forest grouse (Tetrao tetrix, Tetrao urogallus))53, that are also prey for foxes, wolf predation on these species is likely not high enough to affect their abundance. After three or more years since the wolf territory was established, fox abundance rebounded. Unfortunately, our data set did not allow us to assess whether or when fox abundance rebounded starting at pre-wolf numbers. We were unable to conduct a similar analysis of possible time effects on the response of foxes to lynx, since lynx had already established in the area several years before our study started13. It is possible that the fox population underwent a learning process, comparable to what appeared to have occurred with wolves, which enabled them to co-exist with lynx. This process is likely more advanced with the lynx population when compared to the wolf populations, due to the longer time of co-existence between fox and lynx in our study area.

In our study, foxes were positively affected by lynx abundance, while wolf occurrence had a negative temporal effect on fox abundance, although support for the latter effect was weak. The observed response of foxes to these two apex predator species contradicts the results from Pasanen-Mortensen et al.43 and Pasanen-Mortensen & Elmhagen28, who found a strong negative effect of lynx, and no effect of wolves, on fox abundance at the Eurasian scale. This is even more puzzling considering wolves, which prey mainly on moose, likely leave more carrion than lynx, which mainly rely on smaller roe deer34, 41. Also, because the risk of intra-guild predation is greater when two species are of similar size, there should be greater risk for foxes in lynx areas than in wolf areas48. However, the extent of top-down control may depend on the density of the apex predators11, 54. Unfortunately, a comparison between apex predator density in our small-scaled, short-term study (stable or slightly decreasing lynx population and a recolonizing wolf population) and previous studies conducted in Scandinavia is not straightforward due to different measurements of apex predators occurrence (presence/absence28, 43, harvest statistics8), and variation in spatial scales (country8, 9, continent28, 43), time span (long-term9), and population phase (recolonized lynx population9). Our study suggests that the relationship between fox, lynx and wolves might be more complex than observed at a larger scale, or by only using presence/absence data. Actually, Guillaumet et al.55 found that by reducing the spatial scale of analysis there was a positive synchrony between Canadian lynx and coyote (Canis latrans) numbers, contrary to the negative pattern observed at a larger spatial scale. Furthermore, temporal partitioning of activity may facilitate the coexistence of foxes and apex predators56. The importance of spatiotemporal scale is further illustrated by our results that at a larger spatial scale (the landscape level) fox abundance and wolf occurrence were not correlated (c.f., Pasanen-Mortensen et al.43), whereas at a smaller spatial scale (the wolf territory level) we found a negative temporal effect of wolf occurrence on fox abundance. Our results emphasize the importance of considering multiple spatiotemporal scales when investigating species interactions. Further studies at an even smaller scale than used in our study (e.g., locations of tracks in space and time within wildlife triangles or using GPS-collared individuals) are warranted to gain further insight into the impact of lynx and wolf on fox, and determine how this is influenced by habitat.

Materials and Methods

Study System

The study was conducted from 2001–2003 in the south-central part of Sweden within the counties of Värmland and Örebro (referred to as the landscape level, 58°50′–60°N, 12°–17°E, approximately 26 000 km2). The area is dominated by boreal forest interspersed with numerous lakes, rivers and mires that contribute to a more variable landscape than is found in other areas within the boreal forest zone. The proportion of arable land increases from north-northeast to the south-southwest. About 74% of the total land area of Värmland, and 67% of Örebro, is forested and the corresponding area of arable land is 6% and 16%57. The number of days with snow ranges from approximately 75/year in the south to approximately 175/year in the north58. For a detailed description of the study area see Wallgren et al.59.

In the 1980s, the density of the fox population decreased dramatically due to an epizootic outbreak of sarcoptic mange60, and subsequently recovered in the 1990s61. Lynx reoccupied south-central Sweden in the beginning of the 1950s26 and, by 2002–2003, the total Swedish lynx population was estimated to 1300–1600 individuals62. During the study period the number of lynx family groups (i.e., female with kittens, the monitoring unit within the national lynx monitoring program)62 in the study area was 69 (2001) and 49–55 (2003). Roe deer are the main prey for lynx within the study area41. Wolves returned to the study area in the early 1980s through natural re-colonization and, in 2002–2003, the total Swedish wolf population was estimated at 84–100 individuals63. During the study period (winters of 2000/2001–2002/2003), the study area had 8, 10, and 10 wolf territories, including scent marking pairs or family groups (≥3 wolves)63,64,65. Moose are the main prey for wolves in Scandinavia throughout the year38, 53.

Wildlife Triangles and Classification of Wolf Territories

Snow-tracking data on fox, lynx, and wolves were collected between January 15 and March 20, 2001–2003, along 182 equilateral triangular routes66. The triangles had a side length of 4 km and were regularly spaced at predetermined positions about 10 km apart. Intersecting tracks were recorded during 1–4 visits per triangle each winter, each triangle visit occurred ≤4 days (mean 2 days) after snowfall (n2001 = 205, n2002 = 184, n2003 = 106 triangle visits) resulting in a total of 495 triangle visits during the three study years (all triangles were not visited all years, Fig. 4). For each triangle visit, we calculated the relative abundance of each species as crossings 10 km−1 24 h−1, following Kurki et al.37. Kurki et al.37 demonstrated that abundance based on snow-tracking was significantly correlated with hunting bags of two mesopredators (fox and pine marten (Martes martes)) although other factors also contribute to variation in bag. The use of snow-tracking as an index of abundance is further exemplified by Lyly et al.67, but may not work for all species or areas68.

Figure 4

Location of wildlife triangles in south-central Sweden (county of Värmland to the left and county of Örebro to the right) where red fox, lynx, and wolf abundances were surveyed by snow-tracking from 2001–2003, in relation to arable land (grayed, data source land use: 2.1 Arable land and 2.3 Pastures, Corine Land Cover (CLC) 2006, version 18.5 ( Triangles were classified according to a 4-level wolf category: (I) inside an observed wolf territory (grey polygons according to monitoring data63,64,65), (II) inside an average wolf territory (smallest circles), (III) inside a maximum wolf territory (largest circles), and (IV) outside a wolf territory (if not included in category I–III). Maps created using ArcGIS 10.4 (–4).

Fox tracks were recorded at almost all triangle visits, i.e., during 493 of the 495 visits (mean = 15 ± 0.03 SE crossings 10 km−1 24 h−1, range 0.2–139) while lynx tracks were found during 172 visits (mean = 1.5 ± 0.01 SE crossings 10 km−1 24 h−1, range 0.2–12). Wolf tracks were only recorded in 49 triangle visits (mean = 1.9 ± 0.05 SE crossings 10 km−1 24 h−1, range 0.2–15). To analyse the potential effect of wolves on fox abundance, we used two different data sets: (a) wolf abundance (crossings 10 km−1 24 h−1) from the wildlife triangles as described above, and (b) wolf category (I–IV, see below) based on the distance between the triangle and nearest wolf territory with scent-marking wolf pairs or family groups (Fig. 4), available from the national wolf-monitoring system conducted annually by the County Administrative Boards49. Wolf territory borders were estimated based on the monitoring data63,64,65 and locations from VHF- or GPS-collared wolves (available in 22 annual territories). All procedures including capture, handling, and collaring of wolves69 was in accordance with ethical requirements and approved by the Swedish Animal Welfare Agency. This dataset also includes estimates of the minimum number of wolves per wolf territory (range: 2–8 individuals) and time since territory establishment (range: 0–22 years where 0 equals the year of wolf establishment).

As available data from snow-tracking and locations from VHF-collared wolves (two locations per week on average) did not give a full representation of wolf territory borders, we measured distances between the centre of each wildlife triangle to the centre of the nearest wolf territory (using ArcView 3.0, ESRI Corporation, Redlands, CA) and classified wolf occurrence for each wildlife triangle according to the following four categories: (I) inside an observed wolf territory if the triangle centre was located within the confirmed borders of a wolf territory, (II) inside an average wolf territory if the triangle centre was not located inside an observed wolf territory but within a distance corresponding to the average wolf territory radius from the nearest territory centre (18.0 km, average wolf territory size = 1017 km2)70, (III) inside a maximum wolf territory if the triangle centre was not located inside an observed or an average wolf territory but within a distance corresponding to the maximum wolf territory radius from the nearest territory centre (23.0 km, maximum wolf territory size = 1676 km2)70, or (IV) outside a wolf territory if the distance from nearest territory centre was longer than the radius used in category III. Thus, a triangle belongs to only one category.

Habitat Composition and Vole Density Index

As arable and pasture land are known to have a positive impact on fox density due to food abundance8, 37, we calculated the percentage of these habitats (pooled and referred to as arable land), using the Swedish Land Cover Data71 (similar to the European Union classification system Corine Land Cover (CLC)). This was done within a 1 km buffer zone around each wildlife triangle (mean = 8%, range: 0–76%) in order to account for the effect of land composition on fox abundance (Fig. 4). For a detailed description of the methodology and classification see Wallgren et al.59. Furthermore, the fox population tends to fluctuate following the variation in vole density with a time lag of one year35, 72. To account for the effect of vole density, we used the annual estimates of vole density index from Grimsö Wildlife Research Area (located within the study area; 59°40′N, 15°25′E), where bank vole (Clethrionomys Glareolus) and field vole (Microtus agrestis) have been monitored by capture in snap traps in May each year since 1973. In years 2000, 2001, and 2002, the vole density index (both species pooled) was 0.50, 0.49, and 0.81 voles captured per 100 trap nights73. During the period 1990–2010, the vole density index varied between 0 and 1.51. Consequently, the vole density index was classified as low in 2001 and 2002 and high in 2003 for our study period.

Statistical Analyses

We conducted all statistical analyses in R version 3.2.274 using General Linear Mixed Models in the lme4 package75 with fox abundance (crossings 10 km−1 24 h−1) as a response variable. Since fox abundance is a continuous variable, and its variance increased with increasing mean (i.e., similar to a log-normal distribution where σ = μ2), we used log-transformed fox abundance and a Gaussian error distribution in all models to fulfil the assumptions of linearity. We used wildlife triangle identity and year as random effects in all models to account for repeated measurements and year effects. To test the effect of apex predators on fox abundance at the landscape level, we included four explanatory variables: wolves (classified according to wolf abundance or wolf category, as described in (a) and (b) above, exclusive of each other), lynx abundance (crossings 10 km−1 24 h−1), proportion of arable land (%), and vole density index (2-level categorical: low or high). We also included pairwise interactions between habitat and all other explanatory variables as well as an interaction between wolf and lynx abundance.

To assess the effect of wolf pack size and time since territory establishment on fox abundance, we restricted our analyses to data from wildlife triangles that were exclusively inside the observed borders of a confirmed wolf territory (wolf category I, n = 77, referred to as the wolf territory level). We used wolf pack size and time since territory establishment (3-level categorical: year of establishment (referred to as 0), 1–2 years after establishment, or ≥3 years after establishment) together with lynx abundance, proportion of arable land, and vole density index as explanatory variables in the models.

For each analysis, we compared candidate models using the sample-size corrected Akaike Information Criterion (AICc) and AIC weights (w i ) from the ‘MuMIn’ package76 in R. Models with ΔAICc <2 were used to generate model-averaged parameter estimates77. We used a bootstrap method implemented in R using the ‘ez’ package78 to calculate predictions and 95% confidence intervals for mixed models. Additionally, we used AICc weights on the full candidate model set to generate Relative Variable Importance weights (RVI) for each explanatory variable.

Data Availability

All data analysed during this study are included in the Supplementary Data file.


  1. 1.

    Hairston, N. G., Smith, F. E. & Slobodkin, L. B. Community structure, population control, and competition. Am. Nat. 94, 421–425 (1960).

    Article  Google Scholar 

  2. 2.

    Polis, G. A., Myers, C. A. & Holt, R. D. The ecology and evolution of intraguild predation: potential competitors that eat each other. Annu. Rev. Ecol. Syst. 20, 297–330 (1989).

    Article  Google Scholar 

  3. 3.

    Polis, G. A. & Holt, R. D. Intraguild predation: The dynamics of complex trophic interactions. Trends Ecol. Evol. 7, 151–154 (1992).

    CAS  Article  PubMed  Google Scholar 

  4. 4.

    Palomares, F. & Caro, T. M. Interspecific killing among mammalian carnivores. Am. Nat. 153, 492–508 (1999).

    Article  Google Scholar 

  5. 5.

    Soulé, M. E. et al. Reconstructed dynamics of rapid extinctions of chaparral-requiring birds in urban habitat islands. Conserv. Biol. 2, 75–92 (1988).

    Article  Google Scholar 

  6. 6.

    Carpenter, S. R., Kitchell, J. F. & Hodgson, J. R. Cascading trophic interactions and lake productivity. Bioscience 35, 634–639 (1985).

    Article  Google Scholar 

  7. 7.

    Estes, J. A., Tinker, M. T., Williams, T. M. & Doak, D. F. Killer whale predation on sea otters linking oceanic and nearshore ecosystems. Science 282, 473–476 (1998).

    ADS  CAS  Article  PubMed  Google Scholar 

  8. 8.

    Elmhagen, B. & Rushton, S. P. Trophic control of mesopredators in terrestrial ecosystems: Top-down or bottom-up? Ecol. Lett. 10, 197–206 (2007).

    Article  PubMed  Google Scholar 

  9. 9.

    Elmhagen, B., Ludwig, G., Rushton, S. P., Helle, P. & Lindén, H. Top predators, mesopredators and their prey: interference ecosystems along bioclimatic productivity gradients. J. Anim. Ecol. 79, 785–794 (2010).

    CAS  PubMed  Google Scholar 

  10. 10.

    Berger, K. M., Gese, E. M. & Berger, J. Indirect effects and traditional trophic cascades: a test involving wolves, coyotes, and pronghorn. Ecology 89, 818–828 (2008).

    Article  PubMed  Google Scholar 

  11. 11.

    Levi, T. & Wilmers, C. Wolves – coyotes – foxes: a cascade among carnivores. Ecology 93, 921–929 (2012).

    Article  PubMed  Google Scholar 

  12. 12.

    Ripple, W. J. et al. Status and ecological effects of the world’s largest carnivores. Science 343, 1241484 (2014).

    Article  PubMed  Google Scholar 

  13. 13.

    Helldin, J. O., Liberg, O. & Glöersen, G. Lynx (Lynx lynx) killing red foxes (Vulpes vulpes) in boreal Sweden - frequency and population effects. J. Zool. 270, 657–663 (2006).

    Article  Google Scholar 

  14. 14.

    Ritchie, E. G. & Johnson, C. N. Predator interactions, mesopredator release and biodiversity conservation. Ecol. Lett. 12, 982–998 (2009).

    Article  PubMed  Google Scholar 

  15. 15.

    DeVault, T. L., Rhodes, O. E. Jr. & Shivik, J. A. Scavenging by vertebrates: behavioral, ecological, and evolutionary perspectives on an important energy transfer pathway in terrestrial ecosystems. Oikos 102, 225–234 (2003).

    Article  Google Scholar 

  16. 16.

    Selva, N., Jędrzejewska, B., Jędrzejewski, W. & Wajrak, A. Factors affecting carcass use by a guild of scavengers in European temperate woodland. Can. J. Zool. 83, 1590–1601 (2005).

    Article  Google Scholar 

  17. 17.

    Olson, Z. H., Beasley, J. C., DeVault, T. L. & Rhodes, O. E. Scavenger community response to the removal of a dominant scavenger. Oikos 121, 77–84 (2012).

    Article  Google Scholar 

  18. 18.

    Wilson, E. E. & Wolkovich, E. M. Scavenging: how carnivores and carrion structure communities. Trends Ecol. Evol. 26, 129–35 (2011).

    Article  PubMed  Google Scholar 

  19. 19.

    Wilmers, C. C. & Getz, W. M. Gray wolves as climate change buffers in Yellowstone. PLoS Biol. 3, e92 (2005).

    Article  PubMed  PubMed Central  Google Scholar 

  20. 20.

    Wilmers, C. C., Crabtree, R. L., Smith, D. W., Murphy, K. M. & Getz, W. M. Trophic facilitation by introduced top predators: grey wolf subsidies to scavengers in Yellowstone National Park. J. Anim. Ecol. 72, 909–916 (2003).

    Article  Google Scholar 

  21. 21.

    van Dijk, J. et al. Diet shift of a facultative scavenger, the wolverine, following recolonization of wolves. J. Anim. Ecol. 77, 1183–1190 (2008).

    Article  PubMed  Google Scholar 

  22. 22.

    Margalida, A., Colomer, M. À. & Sanuy, D. Can wild ungulate carcasses provide enough biomass to maintain avian scavenger populations? An empirical assessment using a bio-inspired computational model. PLoS One 6, e20248 (2011).

    ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

  23. 23.

    Mattisson, J., Andrén, H., Persson, J. & Segerström, P. Influence of intraguild interactions on resource use by wolverines and Eurasian lynx. J. Mammal. 92, 1321–1330 (2011).

    Article  Google Scholar 

  24. 24.

    Elbroch, L. M. & Wittmer, H. U. Table scraps: inter-trophic food provisioning by pumas. Biol. Lett. 8, 776–779 (2012).

    Article  PubMed  PubMed Central  Google Scholar 

  25. 25.

    Chapron, G. et al. Recovery of large carnivores in Europe’s modern human-dominated landscapes. Science. 346, 1517–1519 (2014).

    ADS  CAS  Article  PubMed  Google Scholar 

  26. 26.

    Liberg, O. Lodjuret: viltet, ekologin och människan. (Almqvist och Wiksell Tryckeri, 1997).

  27. 27.

    Wabakken, P., Sand, H., Liberg, O. & Bjärvall, A. The recovery, distribution, and population dynamics of wolves on the Scandinavian peninsula, 1978–1998. Can. J. Zool. 79, 710–725 (2001).

    Article  Google Scholar 

  28. 28.

    Pasanen-Mortensen, M. & Elmhagen, B. Land cover effects on mesopredator abundance in the presence and absence of apex predators. Acta Oecologica 67, 40–48 (2015).

    Article  Google Scholar 

  29. 29.

    Angelstam, P., Lindström, E. & Widén, P. Role of predation in short-term population fluctuations of some birds and mammals in Fennoscandia. Oecologia 62, 199–208 (1984).

    ADS  CAS  Article  PubMed  Google Scholar 

  30. 30.

    Bartoń, K. A. & Zalewski, A. Winter severity limits red fox populations in Eurasia. Glob. Ecol. Biogeogr. 16, 281–289 (2007).

    Article  Google Scholar 

  31. 31.

    Needham, R., Odden, M., Lundstadsveen, S. K. & Wegge, P. Seasonal diets of red foxes in a boreal forest with a dense population of moose: The importance of winter scavenging. Acta Theriol. 59, 391–398 (2014).

    Article  Google Scholar 

  32. 32.

    Kidawa, D. & Kowalczyk, R. The effects of sex, age, season and habitat on diet of the red fox Vulpes vulpes in northeastern Poland. Acta Theriol. 56, 209–218 (2011).

    Article  PubMed  PubMed Central  Google Scholar 

  33. 33.

    Jedrzejewski, W. & Jedrzejewska, B. Foraging and diet of the red fox Vulpes vulpes in relation to variable food resources in Biatowieza National Park, Poland. Ecography 15, 212–220 (1992).

    Article  Google Scholar 

  34. 34.

    Wikenros, C., Sand, H., Ahlqvist, P. & Liberg, O. Biomass flow and scavengers use of carcasses after re-colonization of an apex predator. PLoS One 8, e77373 (2013).

    ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

  35. 35.

    Lindström, E. Food limitation and social regulation in a red fox population. Holarct. Ecol. 12, 70–79 (1989).

    Google Scholar 

  36. 36.

    Helldin, J.-O. & Danielsson, A. V. Changes in red fox Vulpes vulpes diet due to colonisation by lynx Lynx lynx. Wildlife Biol. 13, 475–480 (2007).

    Article  Google Scholar 

  37. 37.

    Kurki, S., Nikula, A., Helle, P. & Linden, H. Abundances of red fox and pine marten in relation to the composition of boreal forest landscapes. J. Anim. Ecol. 67, 874–886 (1998).

    CAS  Article  PubMed  Google Scholar 

  38. 38.

    Zimmermann, B., Sand, H., Wabakken, P., Liberg, O. & Andreassen, H. P. Predator-dependent functional response in wolves: from food limitation to surplus killing. J. Anim. Ecol. 84, 102–112 (2015).

    Article  PubMed  Google Scholar 

  39. 39.

    Bunnefeld, N., Linnell, J. D. C., Odden, J., van Duijn, M. A. J. & Andersen, R. Risk taking by Eurasian lynx (Lynx lynx) in a human-dominated landscape: effects of sex and reproductive status. J. Zool. 270, 31–39 (2006).

    Google Scholar 

  40. 40.

    Basille, M. et al. What shapes Eurasian lynx distribution in human dominated landscapes: selecting prey or avoiding people? Ecography 32, 683–691 (2009).

    Article  Google Scholar 

  41. 41.

    Andrén, H. & Liberg, O. Large impact of Eurasian lynx predation on roe deer population dynamics. PLoS One 10, e0120570 (2015).

    Article  PubMed  PubMed Central  Google Scholar 

  42. 42.

    Karlsson, J., Brøseth, H., Sand, H. & Andrén, H. Predicting occurrence of wolf territories in Scandinavia. J. Zool. 272, 276–283 (2007).

    Article  Google Scholar 

  43. 43.

    Pasanen-Mortensen, M., Pyykönen, M. & Elmhagen, B. Where lynx prevail, foxes will fail - limitation of a mesopredator in Eurasia. Glob. Ecol. Biogeogr. 22, 868–877 (2013).

    Article  Google Scholar 

  44. 44.

    Wikenros, C., Jarnemo, A., Frisén, M., Kuijper, D. P. J. & Schmidt, K. Mesopredator behavioral response to olfactory signals of an apex predator. J. Ethol. 35, 161–168 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

  45. 45.

    Karanth, K. U. et al. Spatio-temporal interactions facilitate large carnivore sympatry across a resource gradient. Proc. R. Soc. B Biol. Sci. 284, 20161860 (2017).

    Article  Google Scholar 

  46. 46.

    Swanson, A., Arnold, T., Kosmala, M., Forester, J. & Packer, C. In the absence of a ‘landscape of fear’: How lions, hyenas, and cheetahs coexist. Ecol. Evol. 6, 8534–8545 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

  47. 47.

    Cozzi, G. et al. Fear of the dark or dinner by moonlight? Reduced temporal partitioning among Africa’s large carnivores. Ecology 93, 2590–2599 (2012).

    Article  PubMed  Google Scholar 

  48. 48.

    Donadio, E. & Buskirk, S. W. Diet, morphology, and interspecific killing in carnivora. Am. Nat. 167, 524–36 (2006).

    Article  PubMed  Google Scholar 

  49. 49.

    Liberg, O. et al. Monitoring of wolves in Scandinavia. Hystrix 23, 29–34 (2012).

    Google Scholar 

  50. 50.

    Knappwost, U. Das nahrungsspektrum des wolfes (Canis lupus) in Schweden im territorialen vergleich. (Diploma Thesis. School of Forest Science and Resource Management, Technical University of Munich, Germany, 2006).

  51. 51.

    Müller, S. Saisonale variation im nahrungsspektrum des wolfes in Skandinavien. Oder: Was ist wirklich drin im wolfskot? (Diploma Thesis. School of Forest Science and Resource Management, Technical University of Munich, Germany, 2006).

  52. 52.

    Sand, H., Zimmermann, B., Wabakken, P., Andren, H. & Pedersen, H. C. Using GPS technology and GIS cluster analyses to estimate kill rates in wolf-ungulate ecosystems. Wildl. Soc. Bull. 33, 914–925 (2005).

    Article  Google Scholar 

  53. 53.

    Sand, H. et al. Summer kill rates and predation pattern in a wolf-moose system: Can we rely on winter estimates? Oecologia 156, 53–64 (2008).

    ADS  Article  PubMed  Google Scholar 

  54. 54.

    Newsome, T. M. & Ripple, W. J. A continental scale trophic cascade from wolves through coyotes to foxes. J. Anim. Ecol. 84, 49–59 (2015).

    Article  PubMed  Google Scholar 

  55. 55.

    Guillaumet, A., Bowman, J., Thornton, D. & Murray, D. L. The influence of coyote on Canada lynx populations assessed at two different spatial scales. Community Ecol. 16, 135–146 (2015).

    Article  Google Scholar 

  56. 56.

    Lesmeister, D. B., Nielsen, C. K., Schauber, E. M. & Hellgren, E. C. Spatial and temporal structure of a mesocarnivore guild in midwestern north America. Wildl. Monogr. 191, 1–61 (2015).

    Article  Google Scholar 

  57. 57.

    Skogsstyrelsen. Statistical Yearbook of Forestry 2007. (Swedish Forest Agency, Sweden, 2007).

  58. 58.

    Swedish Meteorological and Hydrological Institute. Normalt antal dygn med snötäcke per år. (2015).

  59. 59.

    Wallgren, M., Bergström, R., Danell, K. & Skarpe, C. Wildlife community patterns in relation to landscape structure and environmental gradients in a Swedish boreal ecosystem. Wildlife Biol. 15, 310–318 (2009).

    Article  Google Scholar 

  60. 60.

    Lindström, E. R. et al. Disease reveals the predator: sarcoptic mange, red fox predation, and prey populations. Ecology 75, 1042–1049 (1994).

    Article  Google Scholar 

  61. 61.

    Kjellander, P. & Nordstrom, J. Cyclic voles, prey switching in red fox, and roe deer dynamics - a test of the alternative prey hypothesis. Oikos 101, 338–344 (2003).

    Article  Google Scholar 

  62. 62.

    Liberg, O. & Andrén, H. The lynx population in Sweden 1994 – 2004: an evaluation of census data and methods. (Swedish University of Agricultural Sciences, 2006).

  63. 63.

    Wabakken, P., Aronson, Å., Sand, H., Strømseth, T. H. & Kojola, I. Ulv i Skandinavia: Statusrapport for vinteren 2002–2003. (Høgskolen i Hedmark, 2004).

  64. 64.

    Wabakken, P., Aronson, Å., Sand, H., Steinset, O. K. & Kojola, I. Ulv i Skandinavia: Statusrapport for vinteren 2000–2001. (Høgskolen i Hedmark, 2001).

  65. 65.

    Wabakken, P., Aronson, Å., Sand, H., Steinset, O. K. & Kojola, I. Ulv i Skandinavia: Statusrapport for vinteren 2001–2002. (Høgskolen i Hedmark, 2002).

  66. 66.

    Lindén, H., Helle, E., Helle, P. & Wikman, M. Wildlife triangle scheme in Finland: methods and aims for monitoring wildlife populations. Finnish Game Res. 49, 4–11 (1996).

    Google Scholar 

  67. 67.

    Lyly, M. S. et al. Guardian or threat: does golden eagle predation risk have cascading effects on forest grouse? Oecologia 182, 487–498 (2016).

    ADS  Article  PubMed  Google Scholar 

  68. 68.

    Kojola, I. et al. Tracks in snow and population size estimation: the wolf Canis lupus in Finland. Wildlife Biol. 20, 279–284 (2014).

    Article  Google Scholar 

  69. 69.

    Sand, H., Wikenros, C., Wabakken, P. & Liberg, O. Effects of hunting group size, snow depth and age on the success of wolves hunting moose. Anim. Behav. 72, 781–789 (2006).

    Article  Google Scholar 

  70. 70.

    Mattisson, J. et al. Home range size variation in a recovering wolf population: Evaluating the effect of environmental, demographic, and social factors. Oecologia 173, 813–825 (2013).

    ADS  Article  PubMed  Google Scholar 

  71. 71.

    National Land Survey of Sweden. Kartor och geografisk information. (2008).

  72. 72.

    Lindström, E. R. & Hörnfeldt, B. Vole cycles, snow depth and fox predation. Oikos 70, 156 (1994).

    Article  Google Scholar 

  73. 73.

    Swedish Infrastructure for Ecosystem Science. (2015). Data och publikationer. (2015).

  74. 74.

    R Development Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing. (2015).

  75. 75.

    Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using. J. Stat. Softw. 67, lme4 (2015).

    Article  Google Scholar 

  76. 76.

    Bartón, K. MuMIn: Multi-model inference. (2013).

  77. 77.

    Burnham, K. P. & Anderson, D. R. Model selection and inference: a practical information-theoretic approach. (Springer-Verlag, 2002).

  78. 78.

    Lawrence, M. A. ez: Easy analysis and visualization of factorial experiments. R package version 4.2–2 (2013).

Download references


We thank the Swedish Forest Agency in Värmland and Örebro for collecting the wildlife triangle data and preparing the databases, as well as all those who have participated in the monitoring. Especially we thank Kjell Hagström who led all the field workers with enthusiasm. We acknowledge the support by Harto Lindén and Marcus Wikman. We thank the National Land Survey of Sweden for generously providing us with GIS vegetation maps as well as technical support. We are indebted to Per Ahlqvist, Jon M. Arnemo, and Ulf Grinde who captured and handled the wolves as well as all people that participated in wolf monitoring. The annual estimate of vole density index was made in cooperation with the Swedish Infrastructure for Ecosystem Science (SITES), in this case at Grimsö Wildlife Research Station, SLU. We thank Henrik Andrén, Bodil Elmhagen, and Aimee Tallian who provided valuable comments on earlier drafts of this paper, and Marie Chevallier and Morgane Gicquel for GIS work. The study was funded by the Norwegian Environment Agency, Swedish Environmental Protection Agency, Swedish Research Council Formas, Swedish Association for Hunting and Wildlife Management, WWF (Sweden), Swedish Carnivore Association, and Olle and Signhild Engkvists Foundations.

Author information




R.B., H.S. and O.L. collected the data. C.W., R.B. and M.W. compiled the data. C.W., H.S., O.L., R.B. and M.A. designed the study. This article derives in part from a bachelor thesis conducted by J.H. and supervised by A.J. and C.W. C.W. and M.A. analyzed the data. C.W., M.A., O.L., H.S. and A.J. wrote the article with input from M.W., J.H. and R.B.

Corresponding author

Correspondence to Camilla Wikenros.

Ethics declarations

Competing Interests

The authors declare that they have no competing interests.

Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Wikenros, C., Aronsson, M., Liberg, O. et al. Fear or food – abundance of red fox in relation to occurrence of lynx and wolf. Sci Rep 7, 9059 (2017).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links