Numerical response of predators to large variations of grassland vole abundance, long-term community change and prey switches

Voles can reach high densities with multi-annual population fluctuations of large amplitude, and they are at the base of large and rich communities of predators in temperate and arctic food webs. This status places them at the heart of management conflicts wherein crop protection and health concerns are often raised against conservation issues. Here, a 20-year survey describes the effects of large variations in grassland vole populations on the densities and the daily theoretical food intakes (TFI) of vole predators based on roadside counts. Our results show how the predator community responds to prey variations of large amplitude and how it reorganized with the increase in a dominant predator, here the red fox, which likely negatively impacted hare, European wildcat and domestic cat populations. They also indicate which subset of predator species might have a role in vole population control in the critical phase of a low density of grassland voles. Our study provides empirical support for more timely and better focused actions in wildlife management and vole population control, and it supports an evidence-based and constructive dialogue about management targets and options between all stakeholders of such socio-ecosystems.

The relationship between people and rodents is an old one. Early accounts clearly show 2 that rodents were a destructive agent for crops and a source of disease for many ancient 3 and current societies [1][2][3]. Voles can reach high densities with multi-annual population 4 fluctuations of large amplitude, and they are often considered as pests in temperate 5 farmland [4,5]. However persecuted for this reason [4,6], their effects on biodiversity are 6 crucial. They are at the base of temperate and arctic food webs, maintaining large and population dynamics has also been reported for a long period in Newfoundland, where 48 lynx (Lynx lynx ), prey on snowshoe hares (Lepus americanus), until the hare 49 population crashes. Then, lynx switch to caribou calves (Rangifer tarandus), and the 50 cycle continues [24]. As a whole, those multiple and complex interactions can hardly be 51 investigated in depth by simple modelling [25] or by small-scale experiments that cannot 52 technically take into account all the relevant space-time scales and species communities 53 involved in the real world and, thus, be generalized. many species of carnivorous mammals and birds in grassland and by contrast low 68 densities of secondary prey-resources that are less accessible (vegetation and/or 69 anti-predation behaviour) such as forest, marsh and fallow small mammals (maximum 70 about 3kg.ha −1 ) (e.g. bank vole, wood mice, (Apodemus sp.), field vole, etc.), with 71 periodic (5-6 years) concomitant low densities in every habitats. 72 The variation in this predator community structure over the time span of large 73 fluctuations of prey abundance has not been documented yet in this system, limiting 74 both comparisons with ecosystems described in other part of the world where small 75 mammal outbreaks occur [4] or with more simple food webs of northern ecosystems. 76 Moreover, a large scale inadvertent experiment was offered by chemical control of vole 77 populations in the 1990s, leading to a dramatic decrease in the fox population due to 78 indirect poisoning, and its gradual recovery the following years after a shift in vole 79 control practices [8]. 80 The aim of this 20-year study is to describe the effects of large variations of grassland 81 vole populations on their predator communities and of the long term increase in the fox 82 population in such system. The aims were to (i) describe how a predator community 83 responds to prey variations of large amplitude, (ii) describe how this community 84 reorganizes over the long term with increases in a dominant predator, here the red fox, 85 (iii) attempt to quantify the prey consumption of this predator community. The possible 86 impact on the grassland prey population dynamics will be discussed on this basis. 87 Material and methods 88 Study area 89 The study was carried out around the Pissenavache hamlet (46.95°N, 6.29°E) in 90 Franche-Comté, France, in an area of 3425 ha (2646 ha of farmland, 1094 ha of forest, 91 167 ha of buildings), at an average altitude of 850-900 m above sea level ( Fig. 1 and 2). 92 There, 100% of the farmland was permanent grassland used for pasture and (high grass) 93 meadow for cattle feeding in winter (minimum of 5 months, November-March), with a 94 productivity ranging from 5-6.5 tonnes of dry matter.ha −1 .an −1 under the specifications 95 of the European Protected Geographical Indication of the locally produced Comté 96 cheese. A KML file (S1 kml file) with the bounding box of the study area is provided 97 with the data. 98 Roadside counts 99 Predator and hare (Lepus europeus) populations have been monitored from June 1999 100 to September 2018 (20 years) using night and day road-side counts. Each sampling Fig 1. Location of the study area. a, general location in France; b, study area (red square) and communes it includes; c, land cover, road side counts and the small mammal transect, P1 and P2 indicate the directions of Fig. 2 photos. Until 2009, a road side count segment was driven straight along the dotted line, but in 2010 mud prevented the use of this bypass and slightly changed the itinerary (n-shaped solid line around the dotted line). Commune boundaries were derived from OpenStreetMap and land use from 'BD Carto' provided freely for research by the Institut Géographique National, modified based on field observations.

Fig 2.
General views of the study area. Top, from the road-side count road at P1 (see Fig. 1); bottom, from P2 with the Pissenavache hamlet, a segment of the road-side count road can be seen in the background (photos PG, 20/02/2020).  [33] with small mammals considered as prey. The average body 123 mass of predators, when missing in [33], was estimated based on the Encyclopédie des 124 carnivores de France [34][35][36][37], the Handbook of Birds of Europe, the Middle East and 125 North Africa [38] and the Encyclopedia of Life (https://eol.org).  road-side count itinerary (Fig. 1). Assessments were made in autumn since 1989. The 141 FREDON assessment uses a ranking system that ranges from 0 to 5: 0 -no A. terrestris 142 sign in any parcel within the commune; 1 -low or no A. terrestris tumuli, voles and 143 moles (T. europea) cohabiting the same tunnel systems; 2 -A. terrestris tumuli present 144 in some parcels within the commune and mole burrow systems still present in some 145 parcels; 3 -A. terrestris tumuli present in some parcels within the commune, few or no 146 mole burrow systems present in the commune; 4 -A. terrestris colonies established in 147 the majority of meadows and within pastures; 5 -all of the commune colonized by A.

148
terrestris. The FREDON index not directly translates to transect-based indices, partly 149 because it is applied at the commune scale and not the parcel scale, but Giraudoux et 150 al. [41] found that levels 0-1 correspond to densities < 100 voles.ha −1 , level 2 to 100-200 151 voles.ha −1 , and levels 3-5 to > 200 voles.ha −1 . For a given year, the median score of 152 the 7 communes was taken as a score of abundance.

153
Grassland prey resource relative abundance 154 The dynamics of prey resource abundance in grassland have been estimated (i) over the 155 time span when transects were carried out, summing the relative abundance of A. 156 terrestris and M. arvalis divided by four, divided by the maximum of this sum over the 157 series and (ii) before this time span, when no transect was present, by dividing the into account that the M. arvalis body mass is four times smaller than A. terrestris's on 160 average [43] and helped to better visualize grassland rodent populations variation on the 161 same scale and fill the gap when transect data were lacking. The amplitude of the high 162 density phase is biased to an unknown extent with this method (e.g. arbitrarily 163 summing weighted relative abundances, chained with standardized FREDON scores), 164 but not the time-locations of the low density phases. Thus, the alternation between 165 high density and low density phases, which are always very large (ranging from 0-1000 166 voles.ha −1 ), was robustly and correctly represented over the time series as an 167 abundance index, in the best possible way given the data, for further comparisons.

168
Rodenticide use 169 In France, bromadiolone, an anticoagulant rodenticide, has been used to control water 170 vole populations since the 1980s, with deleterious effects on non-target wildlife including 171 vole predators [9]. In the early 2000s, the development of an integrated pest 172 management (IPM) approach [44] led to a dramatic decreasein the quantity of 173 bromadiolone applied by farmers and their non-intentional effects [8,9]. By law, the  replicates [52]. The grassland prey resource index corresponding to each road-side count 185 was linearly interpolated over time between the two bracketing abundance index 186 estimates.

187
Response of predators to prey abundance. We used generalized linear models 188 with a Poisson error distribution of the form n = a 0 + a 1 ln(x 1 ) + a 2 x 2 + a 3 x 3 + , with 189 n, the number of observations, x 1 , the length of the itinerary, x 2 , the season, x 3 , the  Predator and hare population density estimates. To obtain density estimates, 201 the distance to the itinerary data were analysed using conventional distance sampling 202 with a truncation distance [53][54][55] including 90% of the observations for each species at 203 the minimum. As avoidance behaviour along the road was detected for most species, we 204 used hazard-rate detection functions fitted to the data. This function type has a more 205 pronounced shoulder that compensates for the bias due to avoidance [47]. Models with 206 a seasonal effect as a covariate were compared with concurrent models with no covariate 207 using the Akaike Index Criterion [56].

237
For instance, common buzzard KAI was highly significantly correlated to grassland 238 prey index, with KAI 2.2 times higher in autumn than that in spring. In spring, during 239 the breeding season, KAI was 4.3 times larger in the peak phase than that in the low 240 density phase of grassland vole populations. Red kite's correlation p-value was equal to 241 and kestrel and hen harrier's above but not far from the critical threshold generally 242 accepted of p(Ho) ≤ 0.05. This lack of significance for the latter two species held from 243 one outlier, when prey estimates were derived from the FREDON scores on a communal 244 scale only. Dropping this observation from the data set would lead to reject Ho at p = 245 0.01 and p = 0.02, respectively, and to conclude formally on a correlation between the 246 number of observations of those species and grassland prey abundance. Small mammal population dynamics. Numbers with arrows indicate high density peaks in the communes including the study area; a, dotted grey line, A. terrestris FREDON scores; red line and red scale, quantity of bromadiolone (g) applied for A. terrestris control in the communes of the study; b, abundance index based on transects, vertical bars are 95% confidence intervals (grey scale and dotted line are related to the A. terrestris FREDON scores for comparison); c, estimated variations of the grassland prey resource, the rug on the x axis represents roadside count events.
index but to seasons, with lower counts in winter. Hare and wildcat KAIs were 251 significantly correlated to grassland prey index but seasonal variations could not be 252 detected (Table 1 and Fig. 7). Fox and hare KAIs were highly and negatively correlated 253 to each other (p < 0.001). Furthermore, a model of hare abundance as response variable 254 Day roadside counts. Black circles at the bar top identify autumn counts. The grey line in the background shows the variations of grassland prey abundance (the scale is the same in every plot). The letters above identify the sessions available and selected to estimate densities based on distance sampling during high (ˆ) or low (o) abundance period.
including grassland prey index and fox KAI as independent variables showed that 255 controlling for grassland prey, hare abundance did not significantly correlate to fox KAI 256 at a probability ≤ 0.05 (however with an observed p-value of 0.07).

257
Red fox and badger showed significantly higher abundance in average in the last half 258 of the time series, and hare, wild and domestic cat, long-eared owl and hen harrier 259 significantly lower (one-tailed permutation tests on mean, p < 0.001) ( Fig. 4 and 6).  variations and categorize them as sub-samples of 'low' or 'high' densities (see Fig. 4 and 280 6). Table 3 shows conversion coefficients from KAI to densities, presents the maximum 281 density values observed, and summarizes the estimations obtained using distance 282 sampling by density categories ('low' or 'high'). Considering the relative aggregation of 283 the domestic cat close to buildings, we provide one density estimate for the entire study 284 area, and another for a buffer of 300 m (night) or 250 m (day) around buildings.

300
The largest predator densities were reached during the high density peaks of grassland 301 Fig 8. Distance to buildings of domestic cats for the night and day roadside counts (n obs = 320 and n obs = 101, respectively).  Fig 9. Variations in densities for each species (n.km −2 ). Variations in biomass (kg.km −2 ) and theoretical daily food intake (kg.km −2 .day −1 ) are presented in the annexes 3 and 4.  Table 4. Density (ind.km −2 ) and theoretical daily food intake, TFI (kg.km −2 .day −1 ) in the low (LD) and high (HD) density phases of grassland vole populations. Numbers between parentheses are percentages.    populations [58]. This response was interpreted as being the result of predation switches 339 during the decline phase of the voles, with a supposed relaxation of the predation 340 pressure on the capercaillie during the high density peak, that is well documented e.g. 341 in Scandinavian ecosystems [23,59,60].

342
The variations in the populations of other species were independent of the grassland 343 vole populations over the study time span.

344
Long term changes in the predator community structure  (Fig. 3a). Massive use of anticoagulant rodenticide, here bromadiolone, is 352 known for its deleterious side effects on vole predators [9], with a canid sensitivity that is 353 more than 3 times higher than that of felids [61], and this effect has been proven to have 354 drastically decreased the fox population in the area at the end of the 1990s [62] until the 355 beginning of our study. This difference in sensitivity might explain simultaneously 356 relatively large cat populations due to extremely limited effects of poisoning.

357
Furthermore, Jacquot et al. [8] have shown how the fox population has recovered on 358 a regional scale after the change in rodent control practices. In our study, the predator 359 community shifted from a very low fox density of 0.1 ind.km −2 (CI95% 0.01-0.3) 360 foraging in grassland up to a much larger fox abundance of 2.6 ind.km −2 (CI95% 361 2.2-3.2), with a peak at 4.9 ind.km −2 in autumn 2012 (followed by a stabilization or a 362 slight decrease with an epidemic of sarcoptic mange, which is still ongoing). This value 363 is one of the highest population densities reported in rural landscapes of Europe [63,64]. 364 This increase was concomitant with a sudden and dramatic decrease in the hare 365 population during a low density phase of the vole populations, and with a decrease in 366 wild and domestic cats. This result strongly suggests that those declines might be the 367 consequences of the increase in the fox population, possibly by direct predation or by 368 creating a 'landscape of fear' [65,66], thus limiting the distribution of the prey species 369 to shelter-areas where they could not be detected by roadside counts (houses, forest, 370 etc), or both. In Australia, fox removal experiments showed in one study that cats 371 foraged more in open habitats where foxes were removed [67] and in two others that 372 they were more abundant [68,69]. Furthermore, in western Poland, the hare population 373 during the same year had 1.7 times higher density in response to fox removal [70], and 374 responded positively to sarcoptic mange epidemics that depressed the fox population in 375 Scandinavia [23]. We did not observe changes in the spatial distribution of species 376 between the first and second half of the study, making the 'landscape of fear' hypothesis 377 less likely herein, thus suggesting a major role for direct predation.

378
However the long-term increase in the European badger population since the rabies 379 vaccination in the early 1980s has been well documented in Europe [71][72][73]. In our 380 study, the sudden increase since summer 2013 remains unexplained. increase [74]. In our study, the lack of data regarding Mustela sp. and Martes sp. does 396 not permit us to determine whether those compensations observed in a community 397 subset extend to the whole community of vole predators. Earlier studies in the area and 398 a nearby valley of Switzerland [75,76]   commonly forcing other birds including raptors to drop prey [38]. Their behaviour has 427 not been systematically studied in our area, and the importance of small mammals in 428 the diet is not yet known; however, all the behaviours mentioned above, including 429 scavenging on dead animals, hunting voles and forcing raptors, have been occasionally 430 observed [81]. Thus, one can hardly infer conclusions about the impact of such an 431 opportunistic species in this ecosystem e.g., on vole regulation. Mechanically, however, 432 their number likely has a chronic impact on species that are vulnerable to predation 433 such as small game and bird nests.

434
The other species are more specialized towards small mammal prey. Cricetids) constitute the main prey of wildcats, and they can account for 97% of the 440 diet composition [83], while lagomorphs and birds generally appear as alternative prey. 441 However, when the availability of lagomorphs increases, wildcats can substantially shift 442 their diet towards them [84].

443
In the area, the dietary response of the red fox to variations of grassland vole 444 relative densities differed between M. arvalis (no response) and A. terrestris (Holling's 445 type III-like) [85]. M. arvalis could make up to 60% of prey items in faeces even at very 446 low densities (range from 0-80% of prey items over the whole range of vole densities), and barn owl showed that switching between prey depends on the proportion of the 454 prey available among other prey (frequency dependence), as commonly thought, but 455 also on the total amount of prey (density dependence), with a non-linear frequency and 456 density dependent interactions [25]. area, and the increase in predator populations was likely not enough alone to trigger the 488 decline in vole populations. However, our results suggest that predators during the low 489 density phase were sufficient to considerably slow down the growth phase or even cause 490 the extinction of vole populations locally.

491
Furthermore, our study documented that domestic cat populations could reach much 492 higher densities of 2.4-9.1 ind.km −2 up to more than 18 ind.km −2 around villages 493 within a 250-500 m radius, except during winter nights when they likely prefer to stay 494 warmly at home. In south-central Sweden, Hansson [92] observed that domestic cats, 495 supplied with continuous alternate food, were able to dampen the population 496 fluctuations of the field vole, compared to more or less cat-free areas. In villages some 497 kilometres from our study area, Delattre et al. [27,93] reported a systematic decrease in 498 the abundance of common vole colonies around villages near our study area during 499 similar fluctuations of vole abundance, within an area extending 300 to 400 m from the 500 village edge. This gradient persisted throughout a complete vole population fluctuation. 501 They subsequently hypothesized that this lower density of voles might be the result of 502 cat predation around villages. This figure and our estimates indicate that the 503 combination of domestic cat density and diet, added to the density and diet of other 504 predators, is sufficient to explain this effect.

505
The specific distribution of domestic cats, close to villages, can also cause spatial Overall, our results indicate that in such ecosystem with large variations of grassland 511 prey, the structure of the predator community can change over the long term without 512 changing its overall TFI variation pattern over a rodent cycle. Although the role of 513 small and medium mustelid populations remain unknown, the higher predator densities 514 observed during the grassland rodent peak were mostly due to mobile birds of prey that 515 followed the rodent population increase. However, our results suggest that resident 516 predators alone during the low density phase of grassland rodent populations were able 517 to slow-down the increase or even to cause the extinction of rodent populations locally, 518 but the whole predator community alone was unable to explain the population decrease 519 observed after a high density peak. In such a system, the carrion crow was numerically 520 the largest population with the largest TFI, but its impacts on the ecosystem could not 521 be clearly assessed due to its eclectic diet. After a shift in rodent control practices and a 522 much more moderate usage of anticoagulant rodenticides, the red fox population 523 recovered and then stabilized at much larger densities, which likely negatively impacted 524 hare, wildcat and domestic cat populations. The domestic cat population was 525 aggregated close to buildings, with a 400 m buffer where the vole population was 526 generally lower.

527
From an applied viewpoint, our results strongly suggest that, in such a highly 528 productive and connective grassland system favourable to grassland voles, any means 529 aimed at increasing the populations of predators during the low density phase (e.g. 530 hedgerow networks, roosts, cats around villages, etc.) should lead to better control of 531 grassland small mammal populations (slowing down the increase phase) [94]. However, 532 the impacts of a management with large densities of cats around human settlements on 533 other wildlife [95,96] and pathogen organism transmission (e.g. Toxoplasma 534 gondi ) [97, 98] should be considered. Moreover, in such systems and due to unavoidable 535 prey switches some populations such as the European hare can be caught in a predation 536 sink and can be sustained only at low density. Management options aimed at increasing 537 these vulnerable populations by culling predators (e.g. the red fox, etc.) would conflict 538 with the interests of other stakeholders interested in small mammal pest control. The 539 prohibitive costs and manpower for culling a large number of predators over the long 540 term and the ethical concerns associated with such management should prevent this 541 approach, which has most often been shown to be unsuccessful [99][100][101] and not 542 accepted socially [102]. Other tactics should be sought, including adaptive hunting plans 543 and demand, modification of habitats and landscapes favouring other equilibria in the 544 community, which implies evidence-based and constructive dialogue about management 545 targets and options between all stakeholders of such socio-ecosystems [103]. 546