Received: 15 April 2020 | Revised: 16 August 2020 | Accepted: 17 August 2020 DOI: 10.1002/ece3.6778 ORIGINAL RESEARCH Coexistence of predators in time: Effects of season and prey availability on species activity within a Mediterranean carnivore guild Marc Vilella1 | Mariona Ferrandiz-Rovira2,3 | Ferran Sayol4,5,6 1 Delegació d’Osona (GNO-ICHN), Institució Catalana d’Història Natural, Vic, Spain Abstract 2 CREAF, Cerdanyola del Vallès, Catalonia, 1. The degree of coexistence among predators can determine the structure of eco- Spain logical communities. Niche partitioning is a common strategy applied by species 3 Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Catalonia, Spain to enhance their coexistence. Diet, habitat, or time use can be responsible for 4 Centre for Biodiversity and Environment segregation among carnivore species, the latter factor being the least studied in Research, Department of Genetics, Mediterranean ecosystems. Terrestrial medium-sized carnivores (i.e., mesocarni- Evolution and Environment, University College London, London, UK vores) carry out important functions in ecosystems, and identifying their interac- 5 Department of Biological and tions is essential for their conservation. Environmental Sciences, University of 2. In this study, we explore the activity of a terrestrial mesocarnivore guild in order Gothenburg, Gothenburg, Sweden 6 Gothenburg Global Biodiversity Centre, to determine seasonal differences in daily activity patterns of competitors and Gothenburg, Sweden prey. We also investigate how the abundance of a common mesocarnivore prey in Correspondence the region, small mammals, influences the activity of predators. Ferran Sayol, Centre for Biodiversity and 3. During a year, camera trap devices (n = 18) were installed in Montseny Natural Environment Research, Department of Genetics, Evolution and Environment, Park (Catalan Pre-Coastal Range, North-East Iberian Peninsula), a region that University College London, London, UK. hosts five mesocarnivore species. Camera trapping detections were used to esti- Email: f.sayol@ucl.ac.uk mate their daily activity patterns and corresponding overlaps. We also surveyed Funding information small mammal plots (n = 5) in order to calculate prey abundance and test its effect Institució Catalana d'Història Natural; H2020 Marie Skłodowska-Curie Actions, on the relative activity of each carnivore species. Grant/Award Number: 838998; Diputació 4. Despite all target mesocarnivores are mainly nocturnal, the activity overlap among de Barcelona; Barcelona Zoo Foundation; Generalitat de Catalunya, Grant/Award them varies according to species particularities and season. Red fox (Vulpes vulpes) Number: 2017 SGR 1006 appears as a generalist species in terms of time use, whereas stone marten (Martes foina) and genet (Genetta genetta) show the most similar activity patterns and both of them seem to be positively influenced by small mammal abundance. Overall, the diversity found in the way mesocarnivore species use time could facilitate their coexistence. 5. Despite activity pattern similarities among carnivore species should not be di- rectly translated to negative interactions, they can have a strong influence in habi- tat and resource-limited ecosystems. Therefore, activity overlaps should be taken into account when discussing wildlife management actions. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd 11408 |  www.ecolevol.org Ecology and Evolution. 2020;10:11408–11422. VILELLA et al. | 11409 KEYWORDS activity overlap, activity pattern, camera trapping, mesocarnivore, relative activity index, small mammals, temporal niche 1 | I NTRO D U C TI O N applied to tropical landscapes, there is still a lack of information for many temperate species, especially within the Mediterranean area Mammal carnivores play an important role in terrestrial landscapes, (but see Barrull et al., 2013; Curveira-Santos, Marques, Björklund, as their top-down effects can regulate prey populations with conse- & Santos-Reis, 2017; Monterroso, Alves, & Ferreras, 2014; Torretta, quences spreading across the entire food web (Roemer, Gompper, Serafini, Puopolo, & Schenone, 2016). & Van Valkenburgh, 2009; Schmitz, Hambäck, & Beckerman, 2000). One way to evaluate the temporal niche of a population is the Consequently, they are commonly identified as umbrella spe- daily activity pattern (i.e., the relative activity of the population cies that promote the protection of entire ecosystems (Roberge at each moment of the day). Daily activity patterns of many large, & Angelstam, 2004), thus being targeted for conservation efforts medium, and even small mammal species (Torre, 2004) can be es- (Karanth & Chellam, 2009). In addition, some studies suggest the timated by means of camera trapping, as it enables obtaining pre- protection of entire carnivore guilds in order to preserve their eco- cise dates and times at which individuals are active. Subsequently, system function (Dalerum, 2013; Dalerum, Cameron, Kunkel, & temporal or activity overlaps—defined as the similarity between Somers, 2009). However, protecting multiple carnivore species can two activity patterns—can be also calculated. These similarities are be challenging if species exhibit strong competitive interactions usually described by using three variables: activity levels (i.e., the that compromise their coexistence (Fedriani, Fuller, Sauvajot, & number of hours in a day that a population is active) (Rowcliffe, York, 2000; Linnell & Strand, 2000). Kays, Kranstauber, Carbone, & Jansen, 2014), activity period (i.e., Interspecific competition for resources affects all trophic lev- the exact hours at which the population is active), and activity peaks els (Chesson & Kuang, 2008; Menge & Sutherland, 1976), but it is (i.e., specific hour of maximum activity). Camera trapping also en- especially intense among predators due to their large requirements ables to obtain a Relative Activity Index (RAI), an indirect measure of (Fedriani et al., 2000). Terrestrial carnivores have usually adapted species activity estimated by the number of independent detections their behavior in order to reduce negative interactions among them within a time interval (Bu et al., 2016). Species RAI complements the (Linnell & Strand, 2000), a process known as niche partitioning. information on daily patterns as it explores activity at a larger tem- Several niche axes that could play a role in carnivore coexistence poral scale. Finally, some other advantages of using camera trapping have been analyzed in a variety of ecosystems. Among them, land are a reduction in human sampling error and an increase in monitor- cover selection and diet preferences are some of the most stud- ing scales (Ahumada et al., 2011; Barea-Azcón, Virgós, Ballesteros- ied factors (Barrientos & Virgós, 2006; Carvalho & Gomes, 2004; Duperón, Moleón, & Chirosa, 2007; Burton et al., 2015; Peris, Tena, Curveira-Santos, Pedroso, Barros, & Santos-Reis, 2019; Kelly & & Villena, 2011). Holub, 2008; Linkie, Dinata, Nugroho, & Haidir, 2007; Rosalino, The way carnivores use time can vary according to a number Macdonald, & Santos-Reis, 2004; Ruiz-Olmo & López-Martín, 2001; of factors, such as meteorology, the presence of large apex preda- Torre, Arrizabalaga, & Ribas, 2009; Virgós, Romero, & Mangas, 2001). tors, human frequentation, or even ancient selective pressures that However, despite similar diet or habitat requirements, time parti- are not currently present (Kronfeld-Schor & Dayan, 2003; Parsons tioning could favor the coexistence of species that occupy the same et al., 2016; Rosalino, Macdonald, & Santos-Reis, 2005; Suraci, guild. Due to the popularization of camera trapping as a technique to Clinchy, Dill, Roberts, & Zanette, 2016). However, apart from daylight study wildlife, the number of studies focusing on the temporal niche variations across the year and other aspects derived from seasons of sympatric carnivores has recently increased (Azevedo, Lemos, (Zhao et al., 2020), species diet is one of the factors that could have Freitas-Junior, Rocha, & Azevedo, 2018; Bu et al., 2016; Karanth the strongest influence on carnivore activity (Penido et al., 2017). et al., 2017; Massara, Paschoal, Bailey, Doherty, & Chiarello, 2018; For instance, species with generalist diets can keep a relatively Mukherjee et al., 2019; see also Meredith & Ridout, 2018 and ref- high daily activity overlap among them because there is less risk erences therein). Some of them show that temporal avoidance can of developing exploitation competition (Bu et al., 2016). Moreover, reduce competition and thus facilitate species coexistence (Hearn carnivore activity might be influenced by prey fluctuations, as the et al., 2018; Herrera et al., 2018; Karanth et al., 2017; Lucherini most specialized predators tend to adapt their patterns according to et al., 2009; Massara et al., 2018; Mukherjee et al., 2019), espe- them (Azevedo et al., 2018; Brown, Kotler, & Bouskila, 2001; Foster cially in cases where species segregate between diurnal, crepuscu- et al., 2013; O’Connell et al., 2006). Finally, this overlap between lar, and nocturnal domains (Marinho, Fonseca, Sarmento, Fonseca, predator and prey daily activity patterns can vary across time and & Venticinque, 2020; de Satgé, Teichman, & Cristescu, 2017; Zhao space according to ecosystem dynamics (Barrull et al., 2013; Brown et al., 2020). These studies highlight the need to account for activity et al., 2001; Bu et al., 2016). For instance, a more intense relation- overlaps in order to gain knowledge on the interactions of compet- ship between predator and prey is expected when communities are ing carnivores. However, as most of this work on time use has been regulated in a bottom-up manner, a typical situation that occurs in 11410 | VILELLA et al. locations with resource-limited conditions during some periods of much spread. Instead, forest areas cover almost 80% of the surface, the year, such as the Mediterranean region (Meserve, Kelt, Milstead, being sclerophyll tree species the most abundant—Quercus ilex and & Gutiérrez, 2003). Q. suber (Torre et al., 2009). The main objective of this study was to investigate the tempo- As in many other areas within the Eastern part of the Iberian ral niche partitioning process in a terrestrial mesocarnivore guild Peninsula, large predator species—Iberian lynx (Lynx pardinus), that inhabits a Mediterranean ecosystem. To fulfill this aim, we de- Eurasian lynx (Lynx lynx), brown bear (Ursus arctos) and gray wolf ployed 18 camera trap stations in Montseny Natural Park (Catalan (Canis lupus)—are absent from the massif at least since the last cen- Pre-Coastal Range, North-East Iberian Peninsula) during 1 year and tury (Ruiz-Olmo & Aguilar, 1995). Consequently, the apex of the estimated two different activity variables of mesocarnivore species: regional trophic chain is occupied by five terrestrial mesocarnivore one at a daily level (activity pattern) and the other at a yearly scale species, all of them included in the study: red fox (Vulpes vulpes), (Relative Activity Index—RAI). From the former, we analyzed the ac- stone marten (Martes foina), European badger (Meles meles), common tivity overlap within the mesocarnivore guild and between predators genet (Genetta genetta), and European wildcat (Felis silvestris). Small and a common mesocarnivore prey in the region: small mammals. In mammals are one of the most common prey consumed by these addition, we collected data on small mammal abundance in five dif- predators in Mediterranean habitats (Carvalho & Gomes, 2004). ferent plots to estimate how it might influence the relative activity of However, regarding small mammal percentage in relation to over- predators. The results of this work highlight the importance of iden- all prey biomass in locations that are similar to the study area, the tifying the way carnivores use time and the variables that influence five mesocarnivores could be divided in three groups: species with it in order to understand potential intraguild coexistence issues and occasional consumption of small mammals (badger—<10% accord- define appropriate conservation measures to face them. ing to Rosalino, Loureiro, Macdonald, & Santon-Reis, 2005), spe- cies with regular consumption (red fox and stone marten—around 30% according to Padial, Ávila, & Gil-Sánchez, 2002), and species 2 | M ATE R I A L S A N D M E TH O DS with specialized consumption (wildcat and genet—from 80% to 90% according to Ruiz-Olmo & Aguilar, 1995 and Torre, Ballesteros, & 2.1 | Study area and target species Degollada, 2003, respectively). Other resources exploited by these species are birds (especially the stone marten), invertebrates (typ- Data were collected in Montseny Natural Park (Catalan Pre-Coastal ically the badger), or carrion (mainly the red fox). Finally, fruits are Range, North-East Iberian Peninsula) during a whole year, from July often consumed by red fox, stone marten and badger (Carvalho & 2018 to July 2019. Montseny is the largest and highest massif of Gomes, 2004; Rosalino, Loureiro, et al., 2005). In relation to the the Catalan Pre-Coastal Range, halfway between the Mediterranean habitat, all five mesocarnivores are typical of Mediterranean eco- Sea and the Pyrenees (Figure 1). Montseny stands out by its large systems, especially the forested ones (Barrull et al., 2013; Ferreras, altitude range, from 150 to more than 1,700 m. Consequently, it of- Díaz-Ruiz, Célio, & Monterroso, 2016; Rosalino et al., 2004; Santos- fers an extended gradient of environmental conditions and habitats Reis et al., 2004). Besides, no important avoidance behavior has been within a continuous protected area of 55,090 ha. Despite the fact reported among these five species at a broad scale, thus coexisting −2 that 51,760 people live in this area (~100 people km ), traditional in many large areas (Barrull et al., 2013; Ferreras et al., 2016). Finally, human activities such as agriculture or ranching are not currently home range size is also similar among them: stone marten, genet, F I G U R E 1 Montseny Natural Park position within the Iberian Peninsula, indicated as a green polygon at a regional scale. Locations of camera trapping sites, represented by the mean coordinates obtained from the different device positions within each of them, and small mammal sampling plots are symbolized according to the legend VILELLA et al. | 11411 and wildcat mean home ranges are estimated to measure between 2 minimum trigger delay possible. Camera time was set according to 2 and 3 km (Monterroso, Brito, Ferreras, & Alves, 2009; Santos-Reis Coordinated Universal Time (UTC), as the time standard approaches et al., 2004); while badger and red fox tend to have slightly larger true or apparent solar time the most in the region. No lures or baits territories: mean size between 4 and 5 km2 (Ferreras et al., 2016; were used as the aim was to capture species natural behavior: Some Rosalino et al., 2004). authors reported that attractors can have different effects depend- ing on individuals and species (Barea-Azcón et al., 2007; Rovero & Zimmermann, 2016; Torre et al., 2009). Accounting for local land- 2.2 | Camera trapping scape irregularities (slope, edges, etc.), devices were placed at around 30 cm (exceptionally up to 100 cm) above the ground in order Camera trap devices were set in 18 different sampling sites (these to focus on the images at a height of 20–50 cm, according to the units will be referred to as “camera trapping sites” hereafter), trying size of the target species. Finally, they were secured and vegetation to maximize the geographical and habitat range of Montseny massif around them was cleared to avoid false triggering when necessary. covered. As a result, we sampled a considerable part (533–1,465 m) Most of the cameras operated continuously for nearly a year of the altitude gradient of the protected area. Cameras were always (328.28 ± 50.78; mean nights camera−1 ± SD), resulting in a total of installed in evident fauna trails located at suitable forested habi- 5,909 trap nights, expected to be enough to get a considerable num- tats in order to favor mesocarnivore detectability (Curveira-Santos ber of target species detections (Burton et al., 2015). Devices were et al., 2019; Karanth et al., 2017; Linkie et al., 2007; Mukherjee revised once every 30–40 days to obtain the images and check bat- et al., 2019; Zielinski & Kucera, 1995). Therefore, note that we were teries. In order to minimize the bias that might produce the camera mainly recording activity patterns derived from animal movements position itself—for instance, if by chance a camera was placed near a across forested sites, as opposed to activity in other contexts, such badger sett, badger activity would be overestimated—devices were as hunting in open fields. To facilitate camera revisions, site distribu- moved within each site to 50–150 m from the initial position every tion drew six groups (Figure 1). However, the 18 camera trapping 12–20 weeks, assuring that these changes did not coincide with sea- sites were always located at a minimum distance of 800–1,200 m son solstices or equinoxes. from each other, a sufficient interval to theoretically assume in- Mesocarnivore pictures taken by camera traps were identified dependence among them (Kelly & Holub, 2008; Moruzzi, Fuller, at species level but not for small mammal ones, in which species DeGraaf, Brooks, & Li, 2002; O’Connell et al., 2006). was not evident in most cases (Figure 2). Different individuals of the Devices belonged to two different brands: Cuddeback (C1 and same species were not identified due to the absence of individual C2, models that only differ in the flash type) and Browning (Strike fur marks in four of the five taxa. The rare cases in which more than Force HD Pro). Cameras were set to operate 24 hr a day, to take three one individual of the same species was present in a picture were photo-bursts every time the sensor was triggered and to apply the considered as single detections in order to facilitate the analysis. F I G U R E 2 Example of some camera trapping pictures obtained during the study. From top left to bottom right: red fox Vulpes vulpes, stone marten Martes foina, European badger Meles meles, common genet Genetta genetta, European wildcat Felis silvestris, and wood mouse Apodemus sylvaticus 11412 | VILELLA et al. Picture metadata was extracted with ExifTool application and used groups according to the astronomical season in which it had been to build a database including camera trapping site, species, and date obtained: spring–summer (before 23 September 2018 and after 20 and time of each picture by using the R package camtrapR (Niedballa, March 2019) or autumn–winter (between 24 September 2018 and Sollmann, Courtiol, & Wilting, 2016). Two pictures of the same spe- 19 March 2019). cies at the same camera were only considered as independent detec- tions (contacts) if they were separated by a minimum time interval of 30 min. This interval—together with the 1-hr interval—is commonly 2.4.1 | Daily activity patterns used in camera trapping studies and enabled us to maximize the number of theoretically independent contacts, allowing a more pre- Data were pooled from independent detections (contacts) of all cise estimation of daily activity patterns (Azevedo et al., 2018; Bu camera trapping sites to calculate daily activity patterns of meso- et al., 2016; Jiménez et al., 2017; Kelly & Holub, 2008). carnivore species and small mammals. Firstly, detection times were extracted and transformed into radians (2π radians = 24 hr) before fitting a Kernel density function to them for each species and season 2.3 | Small mammal sampling with R package overlap (Ridout & Linkie, 2009). By comparing Kernel density functions in pairs, activity overlap indices were calculated Five small mammal sampling plots were set within the area (see in three different contexts: (a) overlap between seasons within a Figure 1), and data were collected in two campaigns lasting three species, (b) overlap of mesocarnivore species within a season, and nights each, one during November–December 2018 and the other (c) overlap of a mesocarnivore species and small mammals within a during June–July 2019. These plots are part of a larger network season. Activity overlap indices correspond to the area shared by that aim to monitor small mammal biodiversity (SEMICE Project) in the two functions compared, thus normally taking values from zero the Mediterranean region (Torre, Arrizabalaga, Freixas, Pertierra, to one: the lower the value, the more different activity patterns & Raspall, 2011)—three of them have been running for some years, are. Here, indices were calculated by means of Dhat 4 and Dhat 1 whereas two additional ones were set for the purpose of this study, equations according to the sample sizes obtained: for pairwise following the guidelines of the project (Torre et al., 2011). Each plot comparisons in which the smallest sample size was lower than 50 consisted of 36 live traps (18 Sherman and 18 Longworth model (a total of 16 comparisons out of 36), Dhat 1 was used as recom- units), which were distributed either in a 6 × 6 or in a 9 × 4 grid de- mended by Meredith and Ridout (2018), whereas Dhat 4 was used pending on landscape features and always with a distance of ~15 m in the other comparisons. When using Dhat 4 index, we took 1.0 as between traps, covering a total area of 0.56 ha. All traps were filled bandwidth, and for Dhat 1, 0.8 was applied (Ridout & Linkie, 2009). with a mixture of flour and tuna in oil as bait and a piece of apple Bandwidth is the parameter that regulates the adjustment of the (ca. 10 g) to ensure enough individual hydration during the capture density functions to the times observed; resulting in spikier or softer period, as the piece is never totally consumed. Besides, a handful curves (Rovero & Zimmermann, 2016). Finally, for Dhat 1 calcula- of hydrophobic cotton was added so that the captured individuals tions, density was estimated in 144 equally spaced points (one every could make a nest and feel safe. Traps were checked at dawn so that 10 min), a sufficient amount to obtain accurate estimates (Ridout & small mammals were inside the traps the minimum time possible. Linkie, 2009). Confidence intervals (95%) for overlap indices were All individuals were identified at species level, marked with an ear estimated by bootstrapping 1,000 samples from the Kernel func- tag and released. Capture and handle small mammals authorization tions and calculating the overlap index for each iteration within each was issued by Generalitat de Catalunya (Departament de Territori i pairwise comparison by using the same R package overlap (Ridout Sostenibilitat) with identification code SF/0554/2019. All the han- & Linkie, 2009). Smooth bootstraps (i.e., values are taken from the dling and sampling adhered to the ASAB/ABS Guidelines for the density function instead of from the observed times) were used as Use of Animals in Research. Assuming that during the three-night they take into account the probability of animals being active and campaign all individuals inhabiting the area sampled are captured not detected (Rovero & Zimmermann, 2016). Due to excessively low at least once, small mammal relative abundance was estimated as seasonal mesocarnivore sample sizes obtained in several camera the number of different individuals captured throughout the whole trapping sites (spring–summer: n = 30.94 ± 19.61, autumn–winter: campaign, an index that has been proved to give accurate population n = 21.89 ± 18.08; mean detections site−1 ± SD), it was not possible estimates (Torre et al., 2003). to calculate and compare patterns among different sites, hence ac- tivity overlap was only assessed at a large scale. To test whether two activity patterns could be considered as 2.4 | Data analysis significantly different, we used R package activity (Rowcliffe, 2019). With this purpose, a null distribution of 1,000 random overlap indi- Data analysis is divided in two main sections: (a) species daily ac- ces was created by using bootstrap samples that contained values tivity patterns (mesocarnivores and small mammals) and (b) preda- taken indiscriminately from any of the two density distributions to tor RAI as a function of prey abundance. In both sections, we also be contrasted. Then, the observed or “true” overlap index was com- included the effect of season by dividing the information in two pared to this null distribution to check the probability that it had VILELLA et al. | 11413 F I G U R E 3 Proportion of independent camera trapping detections (contacts) for each mesocarnivore species in each camera trapping site arisen by chance (p). We considered that activity patterns were sig- as continuous variables had been measured in different units, they nificantly different when p < .05. were scaled by calculating the number of standard deviations at which each value was located from the variable mean. For each species, we constructed models from all possible combinations of 2.4.2 | RAI and prey availability explanatory variables by using the dredge and lmer functions from R packages MuMIn (Bartoń, 2015) and lme4 (Bates, Mächler, Bolker, RAI per camera trapping site was calculated for each mesocarnivore & Walker, 2015), respectively. Then, we calculated cumulative AICc species as the number of independent detections per 100 days of weights of the variables (importance function from MuMIn R pack- camera trapping operation (Davis, Kelly, & Stauffer, 2011), assuming age) in order to determine which of them were the most likely to species are equally detectable across time. RAIs were calculated for influence mesocarnivore RAI: the ones with a cumulative AICc the whole year and for every season, as well as for each mesocarni- weight ≥ 0.50 (Barbieri & Berger, 2004). We used the correction of vore species and for all of them together. Before fitting the models, Akaike's Information Criterion (AICc), as it is more suitable for small species year-round RAIs per site were tested for spatial autocorrela- sample sizes. Finally, species most parameterized models (saturated) tion by means of Moran's I index—using mean coordinates obtained were checked for normality and homoscedasticity of the residuals. from the different camera positions within each site to calculate In this analysis, only camera trapping sites located at less than 3 km distance between camera trapping sites—and resulted in no signifi- from a small mammal sampling plot were used (n = 13, Appendix cant effects (see Appendix S1). Besides, to test for differences in de- S5), a distance easily covered by mesocarnivores according to their tection rates between both camera trap models (i.e., Browning and home ranges (Ferreras et al., 2016; Monterroso et al., 2009; Rosalino Cuddeback), we did a permutation test for the mean differences in et al., 2004; Santos-Reis et al., 2004). Besides, if a site had more mesocarnivore year-round RAI per site. Specifically, we generated a than one small mammal sampling plot within this 3 km buffer (n = 4), null distribution of random differences obtained by permuting cam- small mammal abundance was calculated as the mean of these plots. era models among sites for 999 times. We could then compare the Generally, habitat features of the location where the small mammal observed mean difference between models to the ones expected plot was set were similar to the ones characterizing the camera trap- by chance and calculate p as the number of samples from the null ping sites associated with it. model that were greater in absolute value than the observed differ- ence. Overall, the effect of camera model on detection rate was not significant (Appendix S2). Finally, we checked that human frequenta- 3 | R E S U LT S tion was relatively low in most of the sites (Appendix S3) and linear mixed models (LMM) indicated that human pressure did not affect During the whole camera trapping period, 951 independent meso- mesocarnivore seasonal activity (Appendix S4). carnivore detections were obtained (n = 52.83 ± 29.47; mean detec- Small mammal availability effect on mesocarnivore activity was tions site−1 ± SD). Each species appeared in almost all the 18 camera tested by means of linear mixed models (LMM). For each mesocar- trapping sites, except for the wildcat (Figure 3): red fox (18 sites: nivore species, its seasonal RAI per site was taken as the response n = 456, 47.95% of mesocarnivore detections), stone marten (17 variable, while season and small mammal relative abundance—gath- sites: n = 231, 24.29%), European badger (15 sites: n = 134, 14.09%), ering all the species in a single variable—were the main fixed explan- common genet (18 sites: n = 103, 10.83%), and European wildcat (6 atory variables. Besides, the RAI of the other four mesocarnivore sites: n = 27, 2.84%). In spring–summer, cameras captured 58.57% species was also included as a fixed covariate to control for differ- (n = 557) of the total mesocarnivore detections, for a 41.43% ences in detection rates caused by camera position. Finally, cam- (n = 394) in autumn–winter. Camera trapping detection of small era trapping site was added as a random factor. Before modeling, mammals followed the opposite pattern, with 34.65% (n = 220) of 11414 | VILELLA et al. independent detections associated to spring–summer and 65.35% more hours—than in spring–summer. During the latter season, their (n = 415) to autumn–winter. However, their detections were concen- activity showed an only and important peak at 02:00 hr solar time trated in only 9 of the 18 sites overall. Beyond the mentioned taxa, (Figure 4f). six other wild mammal species were captured during the sampling When comparing mesocarnivore species to one another, red fox (Appendix S6). was the species that temporally overlapped the least with the other four mesocarnivores in both seasons (Figure 5). Overlap indices also showed that stone marten, badger, and genet temporally over- 3.1 | Daily activity patterns lapped more among them in spring–summer than in autumn–winter, whereas red fox and wildcat comparison followed the opposite pat- As expected, the five mesocarnivore species and the small mammals tern. All comparisons between red fox and other mesocarnivore spe- exhibited a clear nocturnal activity pattern (Figure 4). During the cies, except for the wildcat, resulted in daily activity patterns being whole year, red fox was the species with the largest activity level— significantly different (Figure 5). On the other side, stone marten and that is, its activity pattern spanned more hours than the rest. Stone genet were the species with the most similar daily routines on aver- marten and genet had an irregular activity pattern, with two clear age (Figure 5). peaks in spring–summer. Finally, badger and wildcat showed a uni- Activity overlap indices resulting from the comparisons between modal pattern with the activity peak centered between 21:00 hr and mesocarnivore and small mammal patterns showed relevant differ- midnight solar time (Figure 4). ences (Figure 6). Stone marten and genet were the species that, in Regarding daily activity pattern variations, red fox and stone spring–summer, overlapped the most with this prey. On the other marten were the only mesocarnivore species that showed significant hand, overlap values became more homogenous among the differ- differences between seasons. However, these shifts did not occur ent predator species in autumn–winter. in the same direction: while red fox began and reduced its activity earlier during autumn–winter, stone marten increased its activity around dawn the same season (Figure 4a,b). Despite not showing 3.2 | RAI and prey availability significant differences, the daily activity of the badger seemed to increase toward dusk in autumn–winter (Figure 4c), a similar process Gathering the five small mammal sampling plots, 291 individuals of 7 as the red fox, whereas genet daily activity shift tended to imitate species were captured (see Data Accessibility Statement). The num- the stone marten one (Figure 4d). Finally, the wildcat was the me- ber of captures varied more between seasons than among different socarnivore with least activity pattern differences between seasons sampling plots (Figure 7). However, two distinct population dynam- (Figure 4e). On the other hand, the small mammal population showed ics were seen: The two plots located above 1,400 m (“Passavets” a considerably larger activity level in autumn–winter—around three and “Turó de l’Home”) showed larger small mammal abundance in F I G U R E 4 Daily activity patterns of the five mesocarnivore species (a–e) and small mammals (f). Line shape and colour according to the season (legend at the bottom). At the top left corner of each panel, Dhat 4 (a, b, c, f) or Dhat 1 (d, e) activity overlap indices obtained by comparing the pattern of both periods, with its 95% confidence interval. At the bottom, p associated with the overlap index, testing for significant differences between seasonal activity patterns. Hours are expressed in UTC and midnight is the central hour VILELLA et al. | 11415 F I G U R E 5 Comparisons between the daily activity patterns of target mesocarnivore species in spring–summer and autumn–winter, resulting in 10 comparisons per season. Graphs on red background (left side) belong to spring–summer, whereas graphs on blue (right side) belong to autumn–winter; function colors in accordance with species (line colour on each draw). At the top of each graph: Dhat 4 (or Dhat 1 for wildcat: both seasons; and genet: autumn–winter) activity overlap indices obtained with their 95% confidence interval. At the bottom: p associated with the overlap index, testing for significant differences. Hours are expressed in UTC and midnight is the central hour autumn–winter, whereas the other plots (550–1,200 m) were more Park. This broad distribution was expected according to the little productive in spring–summer. preference shown by these species when selecting habitat features In mesocarnivore seasonal RAI models, stone marten and genet within Mediterranean forests (Barrull et al., 2013; Ferreras et al., activity was positively influenced by prey availability (Figures 8 and 2016) and to the nonrandom locations of camera trapping sites. This 9), whereas the rest of predator species were not affected by it spatial overlap observed at a large scale highlights the need to ex- (Figure 9). Other mesocarnivore RAI was a relevant explanatory vari- plore potential mechanisms that mesocarnivores use in order to re- able for the red fox, indicating that the canine showed more activity duce intraguild negative interactions. Although different degrees of at sites where its competitors were frequently detected (Table 1, diet specialization have been suggested as an important way to avoid Figure 9). To a lesser degree, this variable also influenced positively competition of sympatric carnivores (Barrientos & Virgós, 2006), our badger and wildcat RAI (Table 1, Figure 9). Finally, season did not results provide more information about the way some species use appear to have any meaningful role according to its cumulative AICc time and support the idea that temporal segregation could be an weights (<0.40) (Table 1). additional mechanism that favors the coexistence within complex Mediterranean carnivore guilds (see Monterroso et al., 2014). In this study, we have identified different activity pattern shapes, with the 4 | D I S CU S S I O N resulting variability in the levels of overlap among species. In addi- tion, we found that some activity features at both daily and yearly Our study found that four out of five target mesocarnivore species — levels are related to season and prey. Overall, time partitioning all except the wildcat — are widely spread across Montseny Natural within a carnivore community should be considered when applying 11416 | VILELLA et al. F I G U R E 6 Comparisons between the daily activity patterns of target mesocarnivore species and small mammals in spring–summer and in autumn–winter, resulting in five comparisons per season. Graphs on red background (left) belong to spring-summer, whereas graphs on blue (right) belong to autumn–winter; function colors in accordance with species (line colour on each draw). At the top of each graph: Dhat 4 (or Dhat 1 for wildcat: both seasons; and genet: autumn-winter) activity overlap indices obtained with their 95% confidence interval. At the bottom: p associated with the overlap index, testing for significant differences. Hours are expressed in UTC and midnight is the central hour measures that aim to protect the guild as a whole (Davis et al., 2011; other hand, activity of small mammals appears to be even more af- Lucherini et al., 2009). fected by season than mesocarnivore one. Being mostly nocturnal, The daily activity pattern of all target mesocarnivores is markedly they would follow the yearly variation of darkness hours. However, nocturnal. However, some activity pattern shifts occur throughout seasonal differences observed in this work should be regarded with seasons, as found in other studies (Barrull et al., 2013; Camps, 2008; caution, as more years of data would be required to test for the con- Torretta et al., 2016), leading to seasonal variations in the degree of sistency of the patterns we report (see Marinho et al., 2020 as an overlap among species. These shifts fall mainly on activity peaks, example). such as the case of the red fox—its activity peak moves toward dusk In general, activity overlap indices among mesocarnivore species in autumn–winter, or on activity levels, such as the stone marten—it obtained here are higher than in other carnivore community studies increases the amount of hours of activity during this same season. which performed a similar analysis (Bu et al., 2016; Curveira-Santos One of the most apparent results is that stone marten, badger, and et al., 2017; Karanth et al., 2017; Mukherjee et al., 2019): previous genet temporal overlaps decrease in autumn–winter, a period when works where a few target species were essentially diurnal, allow- some carnivore feeding opportunities become more challenging ing a broader temporal niche segregation within the guild. Here, de- (Carvalho & Gomes, 2004; Padial et al., 2002) and the increase of spite carnivore activity is gathered at night hours, we still detected darkness hours enables more distinct daily activity patterns. On the that each species differs in the way it temporally overlaps with VILELLA et al. | 11417 F I G U R E 7 Small mammal relative abundance per survey plot. Bars coloured according to the season (legend at the top) F I G U R E 8 Effect caused by small mammal relative abundance on two mesocarnivore RAI models—stone marten (a) and genet (b). On the y-axis: stone marten or genet RAI per site, calculated as the number of independent detections obtained in 100 operational days. On the x-axis: small mammal relative abundance per plot, calculated as the number of different individuals captured the others. For instance, the red fox is active during a much larger intraspecific competition might be occurring and this could result amount of hours than other species do. This generalist behavior of in an apparent wider temporal niche at population level (Bolnick the canine toward time use coincides with previous studies (Barrull et al., 2003). To distinguish between these hypotheses, future work et al., 2013; Ferreras et al., 2016) and could be associated with its should try to identify individual specimens and study their activity largely flexible requirements in both spatial (O’Connell et al., 2006) patterns. and resource dimensions (Carvalho & Gomes, 2004; Ruiz-Olmo & It is expected that species with similar sizes, habitats, and diets Aguilar, 1995). Nevertheless, as it is the most detected carnivore, should have a low activity overlap to reduce negative interactions, 11418 | VILELLA et al. F I G U R E 9 Cumulative AICc weight for the explanatory variables that were considered as candidates to influence seasonal RAI per site of mesocarnivores and were thus used to construct models with all combinations of variables for each species. *RAI other: RAI obtained by gathering the other four carnivore species TA B L E 1 Model β coefficients (SE in brackets) of the most parsimonious model (lowest AICc) that included each variable and their respective cumulative AICc weights (in bold, AICc w+ > 0.50) Species Model parameters Intercept Small mammals Season RAI other Vulpes vulpes Coefficients (β) 1.36 (0.40) −0.28 (0.26) 0.13 (0.44) 1.35 (0.48) AICc w+ – 0.28 0.21 0.87 Martes foina Coefficients (β) −0.09 (0.14) 0.21 (0.07) −0.15 (0.17) 0.06 (0.09) AICc w+ – 0.83 0.24 0.21 Meles meles Coefficients (β) 0.03 (0.16) 0.12 (0.14) 0.01 (0.27) 0.27 (0.15) AICc w+ – 0.23 0.19 0.48 Genetta genetta Coefficients (β) −0.24 (0.07) 0.21 (0.07) 0.06 (0.17) 0.05 (0.07) AICc w+ – 0.84 0.24 0.22 Felis silvestris Coefficients (β) −0.52 (0.06) −0.01 (0.06) −0.18 (0.11) 0.09 (0.05) AICc w+ – 0.20 0.39 0.51 Note: Linear mixed models (LMM) were used, with mesocarnivore species seasonal RAI per site as response variable. Autumn–winter season is taken as the reference level for the factor Season. *RAI other: RAI obtained by gathering the other four carnivore species. SE: standard error. AICc: corrected Akaike's information criterion. w+: cumulative weight. such as the case of yellow-throated marten (Martes flavigula) and more generalist than the one shown by the genet and they can se- masked palm civet (Paguma larvata) (Bu et al., 2016). In our case, quentially use their shared resources. On the other hand, accord- despite these similarities exist between stone marten and genet ing to overlap indices observed, the wildcat seems to slightly avoid (Barrientos & Virgós, 2006; Carvalho & Gomes, 2004), they appear the activity peak of its closer competitor in terms of diet, the genet. to have the highest activity overlap, on average, of all mesocarnivore However, this activity difference should be taken cautiously due to pairs analyzed. Consequently, results would suggest the presence the low number of wildcat detections obtained, obeying its modest of potential negative interactions, such as interference competition. densities in the region (Sayol, Vilella, Bagaria, & Puig, 2018). Further In fact, Santos-Reis et al. (2004) reported such situation, showing analyses in the area should aim to check spatial and temporal over- local territory exclusion between stone marten and genet individ- lap at once as we could not estimate local temporal segregation, uals. Nevertheless, Barrientos and Virgós (2006) observed the ex- which might be different from the one observed at a large scale istence of some discreet feeding mechanisms that could facilitate (Monterroso et al., 2014). In addition, it could also be interesting to their coexistence, as the feeding behavior of the stone marten is study other factors that can potentially have an influence on these VILELLA et al. | 11419 carnivore activity patterns, such as land management or the pres- We can conclude that there is a relatively high interspecific activ- ence of domestic animals (Ahumada et al., 2011; Curveira-Santos ity overlap among the Mediterranean mesocarnivores studied here, et al., 2019; Rosalino, Macdonald, et al., 2005). mainly due to their shared nocturnality. Nevertheless, some sea- Activity pattern shapes in carnivore species not only might re- sonal differences appear on species activity patterns, which can lead flect interactions among them, but could also respond to prey be- to either an increase or reduction of temporal segregation among havior (Azevedo et al., 2018; Mukherjee et al., 2019). Therefore, as them, potentially affecting intraguild coexistence. In this work, de- small mammals are a common prey for our target mesocarnivores spite small mammal features seem to have an effect on stone mar- (Carvalho & Gomes, 2004), we expected an association between ten and genet behavior, activity particularities shown by predators mesocarnivore and small mammal daily activity patterns. In general, cannot be directly associated to their respective feeding habits. In the activity overlap indices between predators and prey obtained general, in terms of time use, red fox appears as the most relaxed here are relatively high in comparison with previous studies (Foster competitor of the guild studied, whereas stone marten and genet et al., 2013). This is especially the case of stone marten and genet are the species with the highest activity overlap, eluding their similar in spring–summer and the wildcat in autumn–winter. Genet and size and habitat preferences. Cases like the latter should be carefully wildcat predilection for small mammals (Ruiz-Olmo & Aguilar, 1995; regarded and conservation efforts should guarantee enough optimal Torre et al., 2003) could partially explain their temporal associa- habitat for both predators (Gompper, Lesmeister, Ray, Malcolm, & tion with this prey. In addition, predator–prey activity overlaps are Kays, 2016). Preserving entire carnivore guilds is essential to main- more homogenous in autumn–winter than in spring–summer. In tain the ecological stability of ecosystems, and here is where the es- winter, consumption of small mammals tends to increase probably tablishment of large and diverse protected areas—such as Montseny because other prey, namely birds and insects, as well as fruits, are Natural Park—together with the characterization of species require- usually less accessible (Carvalho & Gomes, 2004; Padial et al., 2002). ments along different niche dimensions can be useful. Therefore, according to our results in the area sampled, autumn– winter could be the period when small mammals are targeted the AC K N OW L E D G M E N T S most by mesocarnivores. We want to express our most sincere appreciation to all people who According to ecological theory of species competition, tempo- provided help during the fieldwork of this study. For revision of cam- ral and dietary niches should trade-off with each other (Carothers era trap devices and picture identification, we thank Jordi Boixader, & Jaksić, 1984). In our study area, for instance, one would expect Marcos Fernández, Alberto Muñoz, Marçal Pou, and Arnau Tolrà. that the activity overlap between badger and genet—the species For help in small mammal sampling, we thank Ignasi Torre, Cristina with the most distinct diets (Rosalino, Loureiro, et al., 2005; Torre Terraza, Marta Jutglar, Anna Vilà, and Jordi Grajera. We also want to et al., 2003)—should be the highest, whereas the pairs genet/wild- thank Museu de Ciències Naturals de Granollers and Dani Guinart— cat or red fox/stone marten should have a low overlap, according biologist of Montseny Natural Park—for the logistic help during the to their similar feeding requirements (Padial et al., 2002; Torre fieldwork, and Albert Peris for his advice during the analysis. Finally, et al., 2003). However, these predictions were not met in our study we really appreciate the helpful and constructive comments given at any of the seasons. On the other hand, theory suggests that the by the associate editor and two anonymous reviewers. This study most specialized species at capturing small mammals (i.e., genet and was funded by Diputació de Barcelona, Institució Catalana d'Història wildcat) would temporally overlap significantly more with this prey Natural (ICHN), and Barcelona Zoo Foundation. Ferran Sayol re- than the other mesocarnivores. In this case, although prey overlap ceived funding from the European Union's Horizon 2020 research values of both species are relatively high, the genet was not the and innovation program under the Marie Skłodowska-Curie grant predator that overlapped the most with small mammals in any of the agreement No 838998. seasons and the wildcat only in autumn–winter, by a tight difference with respect to the other predators. Therefore, temporal and trophic C O N FL I C T O F I N T E R E S T niche dimensions seem to be contributing rather independently to The authors declare no competing interests. the niche partitioning process of the studied guild. Prey abundance is a factor that could explain variations in me- AU T H O R C O N T R I B U T I O N S socarnivore RAIs. Here, we found that stone marten and genet RAIs Marc Vilella: Conceptualization (equal); data curation (lead); for- were positively linked with small mammal abundance: They were mal analysis (lead); funding acquisition (supporting); methodology also the predators with the highest temporal overlap with this prey (lead); project administration (supporting); software (lead); writ- in spring–summer. This positive effect was more expected for the ing–original draft (lead); writing–review and editing (equal). Mariona genet than for the stone marten due to the important dependence Ferrandiz-Rovira: Conceptualization (equal); supervision (equal); on small mammals shown by the former (Torre et al., 2003). However, validation (equal); writing–review and editing (equal). Ferran Sayol: prey activity pattern coincidences with predators, as well as the ef- Conceptualization (lead); data curation (supporting); formal analysis fects of its availability, should be carefully regarded because other (supporting); funding acquisition (lead); methodology (lead); project factors not analyzed here could be implicated in these predator–prey administration (lead); supervision (equal); validation (equal); writing– relationships, such as habitat features (Karanth et al., 2017). review and editing (equal). 11420 | VILELLA et al. DATA AVA I L A B I L I T Y S TAT E M E N T Carvalho, J. C., & Gomes, P. (2004). Feeding resource partitioning among four sympatric carnivores in the Peneda-Gerês National Park The data generated for this study are available on Dryad (https://doi. (Portugal). Journal of Zoology, 263, 275–283. https://doi.org/10.1017/ org/10.5061/dryad.w6m90​5qn8). S0952​83690​4 005266 Chesson, P., & Kuang, J. J. (2008). The interaction between predation ORCID and competition. Nature, 456, 235–238. https://doi.org/10.1038/ Marc Vilella https://orcid.org/0000-0002-0732-0985 natur​e 07248 Curveira-Santos, G., Marques, T. A., Björklund, M., & Santos-Reis, Mariona Ferrandiz-Rovira https://orcid. M. (2017). Mediterranean mesocarnivores in spatially structured org/0000-0001-8548-2851 managed landscapes: Community organisation in time and space. Ferran Sayol https://orcid.org/0000-0003-3540-7487 Agriculture, Ecosystems & Environment, 237, 280–289. https://doi. org/10.1016/j.agee.2016.12.037 Curveira-Santos, G., Pedroso, N. M., Barros, A. L., & Santos-Reis, M. REFERENCES (2019). Mesocarnivore community structure under predator control: Ahumada, J. A., Silva, C. E. F., Gajapersad, K., Hallam, C., Hurtado, J., Unintended patterns in a conservation context. PLoS One, 14, 18. Martin, E., … Andelman, S. J. (2011). Community structure and https://doi.org/10.1371/journ​al.pone.0210661 diversity of tropical forest mammals: Data from a global cam- Dalerum, F. (2013). Phylogenetic and functional diversity in large carni- era trap network. Philosophical Transactions of the Royal Society vore assemblages. Proceedings of the Royal Society B, 280, 9. https:// B: Biological Sciences, 366, 2703–2711. https://doi.org/10.1098/ doi.org/10.1098/rspb.2013.0049 rstb.2011.0115 Dalerum, F., Cameron, E., Kunkel, K., & Somers, M. (2009). Diversity and Azevedo, F. C., Lemos, F. G., Freitas-Junior, M. C., Rocha, D. G., & depletions in continental carnivore guilds: Implications for prioritiz- Azevedo, F. C. C. (2018). Puma activity patterns and temporal over- ing global carnivore conservation. Biology Letters, 5, 35–38. https:// lap with prey in a human-modified landscape at Southeastern Brazil. doi.org/10.1098/rsbl.2008.0520 Journal of Zoology, 305, 246–255. https://doi.org/10.1111/jzo.12558 Davis, M. L., Kelly, M. J., & Stauffer, D. F. (2011). Carnivore co-existence Barbieri, M. M., & Berger, J. O. (2004). Optimal predictive model selec- and habitat use in the Mountain Pine Ridge Forest Reserve, Belize: tion. Annals of Statistics, 32, 870–897. https://doi.org/10.1214/00905​ Carnivore co-existence in a neotropical pine forest. Animal Conservation, 36040​0 0000238 14, 56–65. https://doi.org/10.1111/j.1469-1795.2010.00389.x Barea-Azcón, J. M., Virgós, E., Ballesteros-Duperón, E., Moleón, M., & de Satgé, J., Teichman, K., & Cristescu, B. (2017). Competition and co- Chirosa, M. (2007). Surveying carnivores at large spatial scales: A com- existence in a small carnivore guild. Oecologia, 184(4), 873–884. parison of four broad-applied methods. Biodiversity and Conservation, https://doi.org/10.1007/s0044​2-017-3916-2 16, 1213–1230. https://doi.org/10.1007/s1053​1-006-9114-x Fedriani, J. M., Fuller, T. K., Sauvajot, R. M., & York, E. C. (2000). Competition Barrientos, R., & Virgós, E. (2006). Reduction of potential food inter- and intraguild predation among three sympatric carnivores. Oecologia, ference in two sympatric carnivores by sequential use of shared 125, 258–270. https://doi.org/10.1007/s0044​20000448 resources. Acta Oecologica, 30, 107–116. https://doi.org/10.1016/j. Ferreras, P., Díaz-Ruiz, F., Alves, P. C., & Monterroso, P. (2016). Factores actao.2006.02.006 de la coexistencia de mesocarnívoros en parques nacionales de ambiente Barrull, J., Mate, I., Ruiz-Olmo, J., Casanovas, J. G., Gosàlbez, J., & Salicrú, mediterráneo, Proyectos de investigación en parques nacionales, M. (2013). Factors and mechanisms that explain coexistence in a Vol. 2011-2014, (pp. 321–339). Madrid, Spain:Organismo Autónomo Mediterranean carnivore assemblage: An integrated study based on Parques Nacionales. camera trapping and diet. Mammalian Biology, 79, 123–131. https:// Foster, V. C., Sarmento, P., Sollmann, R., Tôrres, N., Jácomo, A. T. A., doi.org/10.1016/j.mambio.2013.11.004 Negrões, N., … Silveira, L. (2013). Jaguar and puma activity patterns Bartoń, K. (2015). MuMIn: Multi-Model Inference R package version 1.15.6. and predator-prey interactions in four Brazilian biomes. Biotropica, Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear 45, 373–379. https://doi.org/10.1111/btp.12021 mixed-effects models using lme4. Journal of Statistical Software, 67, Gompper, M. E., Lesmeister, D. B., Ray, J. C., Malcolm, J. R., & Kays, R. 1–48. https://doi.org/10.18637/​jss.v067.i01 (2016). Differential habitat use or intraguild interactions: What Bolnick, D. I., Svanback, R., Fordyce, J. A., Yang, L. H., Davis, J. M., Hulsey, structures a carnivore community? PLoS One, 11, 18. https://doi. C. D., & Forister, M. L. (2003). The ecology of individuals: Incidence org/10.1371/journ​al.pone.0146055 and implications of individual specialization. The American Naturalist, Hearn, A. J., Cushman, S. A., Ross, J., Goossens, B., Hunter, L. T. B., & 161(1), 1–28. https://doi.org/10.1086/343878 Macdonald, D. W. (2018). Spatio-temporal ecology of sympatric fe- Brown, J. S., Kotler, B. P., & Bouskila, A. (2001). Ecology of fear: Foraging lids on Borneo. Evidence for resource partitioning? PLoS One, 13, 25. games between predators and prey with pulsed resources. Annales https://doi.org/10.1371/journ​al.pone.0200828 Zoologici Fennici, 38, 17. Herrera, H., Chávez, E. J., Alfaro, L. D., Fuller, T. K., Montalvo, V., Bu, H., Wang, F., McShea, W. J., Lu, Z., Wang, D., & Li, S. (2016). Spatial Rodrigues, F., & Carrillo, E. (2018). Time partitioning among jaguar co-occurrence and activity patterns of mesocarnivores in the tem- Panthera onca, puma Puma concolor and ocelot Leopardus pardalis perate forests of Southwest China. PLoS One, 11, 15. https://doi. (Carnivora: Felidae) in Costa Rica's dry and rainforests. Revista De org/10.1371/journ​al.pone.0164271 Biología Tropical, 66(4), 1559–1568. Burton, A. C., Neilson, E., Moreira, D., Ladle, A., Steenweg, R., Fisher, Jiménez, J., Nuñez-Arjona, J. C., Rueda, C., González, L. M., García- J. T., … Boutin, S. (2015). REVIEW: Wildlife camera trapping: A Domínguez, F., Muñoz-Igualada, J., & López-Bao, J. V. (2017). review and recommendations for linking surveys to ecological Estimating carnivore community structures. Scientific Reports, 7, 10. processes. Journal of Applied Ecology, 52, 675–685. https://doi. https://doi.org/10.1038/srep4​1036 org/10.1111/1365-2664.12432 Karanth, K. U., & Chellam, R. (2009). Carnivore conservation at the cross- Camps, D. (2008). Activity patterns of adult common genets Genetta roads. Oryx, 43, 2. https://doi.org/10.1017/S0030​60530​8 43106X genetta (Linnaeus, 1758) in Northeastern Spain. Galemys, 20, 47–60. Karanth, K. U., Srivathsa, A., Vasudev, D., Puri, M., Parameshwaran, R., & Carothers, J. H., Jaksić, F. M., & Jaksic, F. M. (1984). Time as a niche dif- Kumar, N. S. (2017). Spatio-temporal interactions facilitate large car- ference: The role of interference competition. Oikos, 42, 403–406. nivore sympatry across a resource gradient. Proceedings of the Royal https://doi.org/10.2307/3544413 Society B, 284, 10. https://doi.org/10.1098/rspb.2016.1860 VILELLA et al. | 11421 Kelly, M. J., & Holub, E. L. (2008). Camera trapping of carnivores: ecosystem. Journal of Wildlife Management, 70, 1625–1633. https:// Trap success among camera types and across species, and habi- doi.org/10.2193/0022-541X tat selection by species, on Salt Pond Mountain, Giles County, Padial, J. M., Ávila, E., & Gil-Sánchez, J. M. (2002). Feeding habits and Virginia. Northeastern Naturalist, 15, 249–262. https://doi. overlap among red fox (Vulpes vulpes) and stone marten (Martes foina) org/10.1656/1092-6194 in two Mediterranean mountain habitats. Mammalian Biology, 67, Kronfeld-Schor, N., & Dayan, T. (2003). Partitioning of time as an 137–146. https://doi.org/10.1078/1616-5047-00021 ecological resource. Annual Review of Ecology Evolution and Parsons, A. W., Bland, C., Forrester, T., Baker-Whatton, M. C., Schuttler, S. Systematics, 34, 153–181. https://doi.org/10.1146/annur​ev.ecols​ G., McShea, W. J., … Kays, R. (2016). The ecological impact of humans ys.34.011802.132435 and dogs on wildlife in protected areas in eastern North America. Linkie, M., Dinata, Y., Nugroho, A., & Haidir, I. A. (2007). Estimating Biological Conservation, 203, 75–88. https://doi.org/10.1016/j. occupancy of a data deficient mammalian species living in trop- biocon.2016.09.001 ical rainforests: Sun bears in the Kerinci Seblat region, Sumatra. Penido, G., Astete, S., Jácomo, A. T. A., Sollmann, R., Tôrres, N., Silveira, Biological Conservation, 137, 20–27. https://doi.org/10.1016/j. L., & Filho, J. M. (2017). Mesocarnivore activity patterns in the semi- biocon.2007.01.016 arid Caatinga: Limited by the harsh environment or affected by inter- Linnell, J. D. C., & Strand, O. (2000). Interference interactions, co-existence specific interactions? Journal of Mammalogy, 98, 1732–1740. https:// and conservation of mammalian carnivores. Diversity and Distributions, doi.org/10.1093/jmamm​al/gyx119 6, 169–176. https://doi.org/10.1046/j.1472-4642.2000.00069.x Peris, A., Tena, L., & Villena, A. (2011). Abundancia de ginetas (Genetta Lucherini, M., Reppucci, J. I., Walker, R. S., Villalba, M. L., Wurstten, A., genetta) en un encinar mediterráneo. Estimación mediante trampeo Gallardo, G., … Perovic, P. (2009). Activity pattern segregation of car- fotográfico. Galemys, 23, 73–79. nivores in the High Andes. Journal of Mammalogy, 90, 1404–1409. Ridout, M. S., & Linkie, M. (2009). Estimating overlap of daily activity https://doi.org/10.1644/09-MAMM-A-002R.1 patterns from camera trap data. Journal of Agricultural, Biological, Marinho, P. H., Fonseca, C. R., Sarmento, P., Fonseca, C., & Venticinque, and Environmental Statistics, 14, 322–337. https://doi.org/10.1198/ E. M. (2020). Temporal niche overlap among mesocarnivores in a jabes.2009.08038 Caatinga dry forest. European Journal of Wildlife Research, 66(2), 1–13. Roberge, J. M., & Angelstam, P. (2004). Usefulness of the umbrella spe- https://doi.org/10.1007/s1034​4-020-1371-6 cies concept as a conservation tool. Conservation Biology, 18, 76–85. Massara, R. L., Paschoal, A. M. O., Bailey, L. L., Doherty, P. F. Jr, & https://doi.org/10.1111/j.1523-1739.2004.00450.x Chiarello, A. G. (2018). Ecological interactions between ocelots and Roemer, G. W., Gompper, M. E., & Van Valkenburgh, B. (2009). The eco- sympatric mesocarnivores in protected areas of the Atlantic Forest, logical role of the mammalian mesocarnivore. BioScience, 59, 165– southeastern Brazil. Journal of Mammalogy, 97, 1634–1644. https:// 173. https://doi.org/10.1525/bio.2009.59.2.9 doi.org/10.1093/jmamm​al/gyw129 Rosalino, L. M., Loureiro, F., Macdonald, D. W., & Santon-Reis, M. (2005). Menge, B. A., & Sutherland, J. P. (1976). Species diversity gradients: Dietary shifts of the badger (Meles meles) in Mediterranean wood- Synthesis of the roles of predation, competition, and temporal lands: An opportunistic forager with seasonal specialisms. Mammalian heterogeneity. The American Naturalist, 110, 351–369. https://doi. Biology, 70, 12–23. https://doi.org/10.1078/1616-5047-00172 org/10.1086/283073 Rosalino, L. M., Macdonald, D. W., & Santos-Reis, M. (2004). Spatial Meredith, M., & Ridout, M. (2018). Overview of the overlap package, R structure and land-cover use in a low-density Mediterranean popula- Project (pp. 1–9). https://cran.r-proje​c t.org/web/packa​ges/overl​ap/ tion of Eurasian badgers. Canadian Journal of Zoology, 82, 1493–1502. vigne​t tes/overl​ap.pdf https://doi.org/10.1139/z04-130 Meserve, P. L., Kelt, D. A., Milstead, W. B., & Gutiérrez, J. R. (2003). Rosalino, L. M., Macdonald, D. W., & Santos-Reis, M. (2005). Activity Thirteen years of shifting top-down and bottom-up control. rhythms, movements and patterns of sett use by badgers, Meles BioScience, 53, 633. https://doi.org/10.1641/0006-3568 meles, in a Mediterranean woodland. Mammalia, 69, 395–408. Monterroso, P., Alves, P. C., & Ferreras, P. (2014). Plasticity in circa- https://doi.org/10.1515/mamm.2005.031 dian activity patterns of mesocarnivores in Southwestern Europe: Rovero, F., & Zimmermann, F. (2016). Camera trapping for wildlife research. Implications for species coexistence. Behavioral Ecology and Exeter, UK: Pelagic Publishing Ltd. Sociobiology, 68(9), 1403–1417. https://doi.org/10.1007/s0026​ Rowcliffe, J. M., Kays, R., Kranstauber, B., Carbone, C., & Jansen, P. 5-014-1748-1 A. (2014). Quantifying levels of animal activity using camera trap Monterroso, P., Brito, J. C., Ferreras, P., & Alves, P. C. (2009). data. Methods in Ecology and Evolution, 5, 1170–1179. https://doi. Spatial ecology of the European wildcat in a Mediterranean org/10.1111/2041-210X.12278 ecosystem: Dealing with small radio-tracking datasets in spe- Rowcliffe, M. (2019). Animal Activity Statistics. R Package, 17. cies conservation. Journal of Zoology, 279, 27–35. https://doi. Ruiz-Olmo, J., & Aguilar, À. (1995). Els grans mamífers de Catalunya i org/10.1111/j.1469-7998.2009.00585.x Andorra. Barcelona, Spain: Lynx Edicions S.L. Moruzzi, T. L., Fuller, T. K., DeGraaf, R. M., Brooks, R. T., & Li, W. (2002). Ruiz-Olmo, J., & López-Martín, J. M. (2001). Relaciones y estrategias Assessing remotely triggered cameras for surveying carnivore distri- ecológicas de los pequeños y medianos carnívoros forestales. In J. bution. Wildlife Society Bulletin, 30, 380–386. Camprodon & E. Plana (Eds.), Conservación de la biodiversidad y gestión Mukherjee, S., Singh, P., Silva, A. P., Ri, C., Kakati, K., Borah, B., … forestal (pp. 397–414). Bracelona, Spain: Edicions de la Universitat de Ramakrishnan, U. (2019). Activity patterns of the small and me- Barcelona. dium felid (Mammalia: Carnivora: Felidae) guild in Northeastern Santos-Reis, M., Santos, M. J., Lourenço, S., Marques, J. T., Pereira, I., India. Journal of Threatened Taxa, 11, 13432–13447. https://doi. & Pinto, B. (2004). Relationships between stone martens, genets org/10.11609/​jott.4662.11.4.13432​-13447 and cork oak woodlands in Portugal. In D. J. Harrison, A. K. Fuller, Niedballa, J., Sollmann, R., Courtiol, A., & Wilting, A. (2016). & G. Proulx (Eds.), Martens and fishers (Martes) in human-altered envi- CamtrapR: An R package for efficient camera trap data manage- ronments (pp. 147–172). New York, NY: Springer-Verlag. https://doi. ment. Methods in Ecology and Evolution, 7, 1457–1462. https://doi. org/10.1007/0-387-22691​-5_7 org/10.1111/2041-210X.12600 Sayol, F., Vilella, M., Bagaria, G., & Puig, J. (2018). El gat salvatge, Felis O’Connell, A. F., Talancy, N. W., Bailey, L. L., Sauer, J. R., Cook, R., & silvestris (Schreber, 1777), al Prepirineu oriental: Densitat de les po- Gilbert, A. T. (2006). Estimating site occupancy and detection blacions del Lluçanès i el Bisaura. Butlletí De La Institució Catalana probability parameters for meso- and large mammals in a coastal D’història Natural, 82, 185–191. 11422 | VILELLA et al. Schmitz, O. J., Hambäck, P. A., & Beckerman, A. P. (2000). Trophic cas- genetta), in central Spain. Canadian Journal of Zoology, 79, 1544–1551. cades in terrestrial systems: A review of the effects of carnivore re- https://doi.org/10.1139/cjz-79-9-1544 movals on plants. The American Naturalist, 155, 141–153. https://doi. Zhao, G., Yang, H., Xie, B., Gong, Y., Ge, J., & Feng, L. (2020). Spatio- org/10.1086/303311 temporal coexistence of sympatric mesocarnivores with a single apex Suraci, J. P., Clinchy, M., Dill, L. M., Roberts, D., & Zanette, L. Y. (2016). Fear carnivore in a fine-scale landscape. Global Ecology and Conservation, of large carnivores causes a trophic cascade. Nature Communications, 21, 11. https://doi.org/10.1016/j.gecco.2019.e00897 7, 7. https://doi.org/10.1038/ncomm​s10698 Zielinski, W. J., & Kucera, T. E. (1995). American marten, fisher, lynx, and Torre, I. (2004). Distribution, population dynamics and habitat selection of wolverine: Survey methods for their detection (No. PSW-GTR-157). small mammals in Mediterranean environments: The role of climate, veg- Albany, CA: U.S. Department of Agriculture, Forest Service, etation structure, and predation risk (Ph.D Thesis). Barcelona, Spain: Pacific Southwest Research Station. https://doi.org/10.2737/ Universitat de Barcelona. PSW-GTR-157. Torre, I., Arrizabalaga, A., Freixas, L., Pertierra, D., & Raspall, A. (2011). Primeros resultados del programa de seguimiento de micromamíferos S U P P O R T I N G I N FO R M AT I O N comunes de España (SEMICE). Galemys, 23, 81–89. Torre, I., Arrizabalaga, A., & Ribas, A. (2009). Estudio de la comunidad Additional supporting information may be found online in the de carnívoros del P.N. del Montseny (Catalunya) mediante trampeo Supporting Information section. fotográfico. Galemys, 21, 165–180. Torre, I., Ballesteros, T., & Degollada, Y. A. (2003). Cambios en la dieta de la gineta (Genetta genetta Linnaeus, 1758) con relación a la dis- How to cite this article: Vilella M, Ferrandiz-Rovira M, Sayol F. ponibilidad de micromamíferos: ¿Posible preferencia por el topillo rojo? Galemys, 15, 12. Coexistence of predators in time: Effects of season and prey Torretta, E., Serafini, M., Puopolo, F., & Schenone, L. (2016). Spatial and availability on species activity within a Mediterranean temporal adjustments allowing the coexistence among carnivores carnivore guild. Ecol Evol. 2020;10:11408–11422. https://doi. in Liguria (NW Italy). Acta Ethologica, 19(2), 123–132. https://doi. org/10.1002/ece3.6778 org/10.1007/s1021​1-015-0231-y Virgós, E., Romero, T., & Mangas, J. G. (2001). Factors determining “gaps” in the distribution of a small carnivore, the common genet (Genetta