Sett Use, Density and Breeding Phenology of Badgers in Mediterranean Agro-Sylvo-Pastoral Systems

animals Article Sett Use, Density and Breeding Phenology of Badgers in Mediterranean Agro-Sylvo-Pastoral Systems Marcelo Silva †, Luís Miguel Rosalino †, Sandra Alcobia and Margarida Santos-Reis *   Citation: Silva, M.; Rosalino, L.M.; Alcobia, S.; Santos-Reis, M. Sett Use, Density and Breeding Phenology of Badgers in Mediterranean Agro-Sylvo-Pastoral Systems. Animals 2021,11, 2663. https:// doi.org/10.3390/ani11092663 Academic Editors: Nicola Marples, Aoibheann Gaughran and Enda Mullen Received: 1 August 2021 Accepted: 9 September 2021 Published: 10 September 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). cE3c—Centre for Ecology, Evolution and Environmental Changes, Faculdade de Ciências, Universidade de Lisboa, 1749-016 Lisboa, Portugal; [email protected] (M.S.); lmr[email protected] (L.M.R.); [email protected] (S.A.) *Correspondence: mmr[email protected] † These authors contributed equally to this work. Simple Summary: Understanding carnivores social structure variation is pivotal for properly addressing conservation challenges and solutions. The European badgers is a social carnivore for which most of the available information regarding how this species is socially organized derives from central west populations. This article describes the group composition, den use patterns and breeding phenology of a Mediterranean population of badgers. We showed that badger live in low density, in relatively small groups, composed by 2–4 adult animals and ca. 2 cubs, born in winter. These patterns, representing a variation of what was described for other populations, show that badgers take advantage of the landscape context, where human-related resources and mild environmental conditions allow badger to reach higher densities than in many southern populations, and to reproduce earlier than their northern counterparts. Abstract: Carnivores social organization varies widely, from strongly social to solitary predators. European badgers are facultative social carnivores that also shows a geographical variation in social structure. These patterns derive mainly from central/west European regions, with an underrepresentation of Mediterranean populations that face different conservation challenges, especially regarding group composition, sett use patterns and breeding phenology. We addressed these traits topics for a population inhabiting a Portuguese agro-silvo-pastoral system. Based on monthly monitoring of 34 setts and continuous camera-trapping surveys of 12, we showed that setts surrounded by diversified vegetation and located in sandy sites are more used, a pattern probably linked to food availability and ease of sett excavation and maintenance, respectively. Badgers followed a general pattern regarding group size (2–4 adults), but showed an intermediate population density ( 0.49–0.73 badgers/km2 ), with values higher than those estimated for other Mediterranean environments, but lower than for central-western populations. This, together with the breeding (November/January) and cub emergence (1.8 cubs/sett; March/April) periods, indicates an ecological adaptation to the landscape context, where human-related resources and mild environmental conditions allow badger to reach higher densities than in many southern populations, and to reproduce earlier than their northern counterparts. Keywords: camera-trapping; Meles meles; density; social organization; reproduction 1. Introduction Carnivores show a wide array of social organizations, ranging from solitary (e.g., wildcats, Felis silvestris) to highly social species, such as the meerkat (Suricata suricatta) [ 1 ]. However, less than 20% are considered social predators [ 2 ], and even those show distinct degrees of sociality. While wolves (Canis lupus) can form packs with more than 40 animals sharing hunting, territory patrolling and defense, cub rearing, and other activities [ 3 ], others, such as European badgers (Meles meles), form clans usually with less than seven Animals 2021,11, 2663. https://doi.org/10.3390/ani11092663 https://www.mdpi.com/journal/animals Animals 2021,11, 2663 2 of 15 individuals that share the same refuge and territory and show some intra-clan interactions, but lack the social sophistication showed by highly social species [ 4 ]. The way badger social groups are organized and use the communal refuges they build (known as setts) is however highly variable across its distribution range and this variability still poses questions about how the landscape context may shape this species’ social patterns. Throughout Europe, the size of badger social groups ranges from singe pairs in southern Spain [ 5 ] to seven adult individuals in Wytham Woods, UK (although reaching 30 adults/yearling badgers in unusual situations, such as those found in Woodchester Park, UK, in 1989) [ 6 ]. Such social structure determines the species density, which peaks in the UK (38 ind./km 2 ) and reaches its lowest value in Eastern Europe (e.g., Czech Republic, 0.12 ind./km 2 [ 7 ]. This variation in density is frequently linked to climate variation (e.g., wetter climate favors earthworm’s abundance and, indirectly, badgers), abundance and availability of potential sett sites and/or food, and level of disturbance, such as human population density, road density or hunting pressure [8]. Iberia, corresponding to the species south-western range limit, seems to be a challenging environment for badgers, which occur at low densities (0.13–0.67 ind./km 2 , in Spain, and 0.36–0.48 ind./km 2 in Portugal) [ 9 – 12 ]. This population structure seems to be determined mostly by the low availability of food resources [ 11 ] or sett sites [ 13 ] in Mediterranean environments. However, estimations of badger density are limited to very few regions in Iberia—Doñana National Park (S) and Park of Collserola (NE), in Spain, and Serra de Grândola in Portugal [ 10 – 12 ]—not representing the full range of environmental conditions badgers explore in Iberia. Thus, more data are needed to allow a better understanding of how Iberian populations respond to and are influenced by the Iberian landscape context, because this region is expected to suffer drastic environmental changes in the next future due to climate warming [14,15]. Setts play a central role in badger socio-ecology, acting not only as refuge sites (where individuals spend as much as 70% of their time), but also as focal points for reproduction, a nursery, and a social exhibition and interaction arena [ 4 ]. For all these reasons, badgers density is often linked to sett density, especially when considering the main setts, i.e., larger setts (>5 entrances) with signs of a regular use throughout the year [ 16 ]. However, in low density areas, as Iberia, this relation is weaker, and keeps valid only including also smaller and less frequently used setts (often called, secondary, outliers or subsidiary) [ 16 ]. Whatever the type of setts considered, their use pattern is determined by their internal structure, landscape context and disturbance risk [17–19]. In the Mediterranean region, human presence has shaped the landscape for millennia, converting natural environments to agriculture land, pasture for livestock or human settlements/infrastructures [ 20 ], affecting wildlife ecological patterns. In Iberia, badgers show low sett use rates, with animals frequently changing between setts in consecutive days (ca.50% of the occasions), and the re-use rate greatly declining after five nights [ 21 ]. This pattern seems to be linked to anthropic disturbances, such as cattle breeding and removal of forest understory around the setts [22]. Setts are also a pivotal structure in badger reproduction, as birth occurs in underground sett chambers, and cubs stay inside until they can explore the outside environment [ 4 ]. While mating can occur throughout the year, most available birth records are concentrated in February, due to delayed implantation, with a litter size of 2–3 cubs (range 1–5) [ 4 ]. However, this information originates mainly from UK populations, with scarce data available for Mediterranean populations, where the highly different climate may produce variations to the British pattern. Available data for the Iberian Peninsula indicate that in captive animals mating occurred between February and May, with the birth of 1–3 cubs in February [ 23 ], while 3–4 cubs per sett have been reported for a wild population in Portugal [12]. Considering the specific characteristics of Mediterranean landscapes, and the lack of information on the ecology of badgers in this region, our study aimed to assess badger: Animals 2021,11, 2663 3 of 15 (1) density; (2) sett use patterns and underlying determinants; and (3) breeding phenology and litter size. To fulfill these goals, we tested eight hypotheses (Table 1). Table 1. Hypotheses tested targeting three life-history patterns: sett use, density and breeding phenology. For each, we detail the underlying reasoning and the supporting references. Pattern Hypothesis Reasoning Supporting References Density H1—Badger’s density will be lower when compared to the species core range area, but within the limits reported for Iberian populations Studies implemented in Iberia estimate a population density of 0.13–0.67 ind./km2, significantly lower than that estimated for central/west European populations (mean = 9.2 ind./km2, SD = 10.5) [4,9–12] Sett use pattern H2—Sett use is promoted by the occurrence of dense vegetation surrounding the sett In human-shaped landscapes, the occurrence of dense vegetation provides a more protective context, where animals may socially interact more safely (including cubs) [24,25] H3—Setts located in areas with a easily diggable and well-drained soil will be more frequently used Badgers prefer well drained and cohesive soils as sett sites [26,27] H4—Sett use is promoted by the proximity to feeding patches (e.g., olive yards, wildlife feeding stations) Sett location near patches providing food resources allows badgers to save energy and time in their foraging bouts [22,28] H5—Setts subject to high disturbance (e.g., cattle, roads proximity) will be less used by badgers Anthropogenic disturbance is known to affect sett use by badgers [22,29–31] H6—Setts located in areas where perceived competition can be high will be less used Evidence exist that the presence of wild boars constrains badgers presence, due to resource competition. Inter-group competition is also considered a passive range exclusion mechanism. [32,33] Reproduction phenology H7—Badger mating peaks in January/February and cubs will start emerging from the setts in April Badger mating can occur all year round, with two peaks in Winter/Spring (main) and Summer/Autumn; cubs emerge eight weeks after. [11,34] H8 —Litters will be composed of 2–3 cubs Although available data for Portugal indicate 3-4 cubs/litter, the average values for most badger populations is lower [12,34] 2. Materials and Methods 2.1. Study Area The study was conducted in the “Charneca do Infantado” (38 ◦ 48 0 N 8 ◦ 49 0 W; Figure 1 ), a 100 km 2 farmstead managed by Companhia das Lezírias S.A., mainly devoted to the production of cork (cork oak, Quercus suber, woodlands—67.50 km 2 ), wood (maritime pine, Pinus pinaster, plantations—9.71 km 2 ; Eucalyptus, Eucalyptus globulus, plantations; 4.76 km 2 ), pine nuts (stone pine, Pinus pinea, plantations—5.08 km 2 ), and agricultural goods (e.g., corn, Zea mays—2.50 km 2 ; rice, Oryza spp.—2.40 km 2 ; olives, Olea europaea—0.70 km 2 ). This multiuse agroforestry system also includes other activities, namely seasonal cattle raising, with cattle grazing in the farmstead in Autumn/Winter, and hunting mostly of wild boars (Sus scrofa) and partridges (Alectorys rufa), for which several artificial feeding stations with cereals and water are available throughout the farmstead [35]. The region has a typical Mediterranean climate, with hot and dry summers and mild but rainy winters. During the study period the mean temperature and precipitation were of 16.8 ◦ C and 361 mm, respectively. “Charneca do Infantado” is located in a plain lowland area, with few relatively deep valleys formed by temporary watercourses. Only Animals 2021,11, 2663 4 of 15 one permanent stream crosses the farmstead (“Ribeira de Vale Cobrão”), although several artificial water points and temporary ponds are scattered throughout the areas. Animals2021,11,26634of15  artificialfeedingstationswithcerealsandwaterareavailablethroughoutthefarmstead [35]. TheregionhasatypicalMediterraneanclimate,withhotanddrysummersandmild butrainywinters.Duringthestudyperiodthemeantemperatureandprecipitationwere of16.8°Cand361mm,respectively.“CharnecadoInfantado”islocatedinaplainlowland area,withfewrelativelydeepvalleysformedbytemporarywatercourses.Onlyone permanentstreamcrossesthefarmstead(“RibeiradeValeCobrão”),althoughseveral artificialwaterpointsandtemporarypondsarescatteredthroughouttheareas.  Figure1.Studyarealocation(A),andspatialdistributionofdetectedbadgersetts(B),highlightingthoseselectedfor intensivemonitoringusingcamera‐traps(C).Differentcolorsrepresentdistinctbadgersocialgroups,andmainsettsare representedbybiggercircles. 2.2.SettSurvey Wefirstreviewedandcompiledallsettrecordsregisteredinpreviousstudies conductedinthestudyarea(e.g.,[22,36]).Allthesesetts(n=33)wererevisitedtoconfirm theirusebybadgerandnewsettswereintensivelysearchedinthestudyarea.Theareais dividedincattlegrazingplotsand,betweenSeptember2016andApril2017,wesurveyed 20thatfulfilledthecharacteristicsconsideredpromotersofsettemplacement(vegetation andsoilfeatures;see[4];andavoidingintensiveagricultureareas,suchasriceandcorn fields,waterchannelandreservoirs,etc.).Theseplots,totalizing15km2(meanarea1.45 km2,sd=1.05,range:0.05–6.66km2)weresurveyedbasedonseverallinearandparallel transects,coveringtheentirecell.Thedistancebetweenthetransectswasabout50meters, Figure 1. Study area location ( A ), and spatial distribution of detected badger setts ( B ), highlighting those selected for intensive monitoring using camera-traps ( C ). Different colors represent distinct badger social groups, and main setts are represented by bigger circles. 2.2. Sett Survey We first reviewed and compiled all sett records registered in previous studies conducted in the study area (e.g., [ 22 , 36 ]). All these setts (n= 33) were revisited to confirm their use by badger and new setts were intensively searched in the study area. The area is divided in cattle grazing plots and, between September 2016 and April 2017, we surveyed 20 that fulfilled the characteristics considered promoters of sett emplacement (vegetation and soil features; see [ 4 ]; and avoiding intensive agriculture areas, such as rice and corn fields, water channel and reservoirs, etc.). These plots, totalizing 15 km 2 (mean area 1.45 km 2 , sd = 1.05 , range: 0.05–6.66 km 2 ) were surveyed based on several linear and parallel transects, covering the entire cell. The distance between the transects was about 50 meters, since we surveyed a buffer of 25 m in each perpendicular direction. In more closed areas (woodlands, for example), the distance between transects was shortened to ensure that the whole area was surveyed, as the visual detection of setts decreases in dense vegetation areas. Assuming that main setts are located in the core area of each group territory [ 24 ], and considering that badgers home range in a similar Portuguese landscape was 4.46 km 2 [ 12 ], Animals 2021,11, 2663 5 of 15 we assumed that all secondary setts located within a radius of 1191 m from a main sett (see Section 2.3) will belong to the same social group. The only exceptions were secondary setts either separated by an insurmountable geographical barrier (e.g., permanent stream) or located within the buffer boundaries of two neighbor main setts (in such situation they were grouped with the closest main sett). 2.3. Sett Monitoring To determine the sett use pattern, all detected setts were visited monthly to determine badger signs of presence, and estimate a sett use index, ranging from 0 (never active) to 7 always active. A sett was considered active if nearby its entrances we found: (1) freshly excavated soil, (2) recently used latrines; (3) fresh nest material (Figure S1; Supplementary Material); and/or (4) high abundance of footprints. Setts that showed frequent and abundant activity signs (sett index = 6–7), or where reproduction was detected, were considered main setts [ 12 ], all the remaining being classified as secondary [ 24 ]. The main setts were the target of an intense monitoring program, using a camera trapping approach [ 37 ]. We installed 14 Moultrie ® m-990i Gen2 (Calera, AL, USA) camera-traps (Figure S2; Supplementary Material), equipped with a 16 Gb memory card, set to take 30s low resolution videos (854–480 pixels) when the heat/movement sensor was activated, with a minimum time interval between consecutive videos of 5s. Cameras were installed 30 cm above the ground, attached to a tree or a wooden stick, facing one or more active sett entrances. In two setts with more than one entrance active, and that could not be simultaneously monitored with a single camera, an additional one was set to ensure the monitoring of all active entrances. Cameras were checked weekly to replace the batteries and memory cards. Sett monitoring occurred between November 2016 and April 2017, and each camera was active for 180 days, for an overall monitoring period of 2520 camera-trapping nights. 2.4. Sett Characterization In order to assess which environmental drivers, were explaining the sett use pattern, we used 21 variables to characterize each sett (Table 2). Variables associated to vegetation composition and structure were assessed, in situ, within the minimum convex polygon that encompassed all the sett entrances, plus a buffer of 2 m. The Simpson diversity index [ 38 ] was estimated based on the herbaceous, shrub and bare soil cover in each polygon. Later, we built a Geographic Information System (GIS; QGIS ® version 2.8.3—Wien, Beaverton, OR, USA [ 39 ]), that encompassed remote sensing information regarding the: type of soil (data provided by the Companhia das Lezíria, S.A, Samora Correia, Portugal); land use [ 38 ]; location of setts, game feeding stations and watering pivots, paved and unpaved roads and water points (to allow the estimation of the distance of each sett from these landmarks; data provided by the Companhia das Lezíria, S.A.); grazing pressure per plot (data provided by the Companhia das Lezíria, S.A.); and wild boar occurrence determined in a previous study [40]. 2.5. Group Size and Breeding Phenology Each video was carefully viewed and we registered the number of individuals identified, the age class (i.e., adult, cub; based on animal’s size and survey month [ 4 ], the interaction between the animal and the sett (i.e., entering or leaving the sett, excavating or handling nest material), as well other behaviors or social interactions (e.g., antagonistic or breeding behaviors). To estimate the number of individuals per social group we used the highest number of badgers recorded simultaneously. However, this estimation was only considered valid if that number of individuals was observed in, at least 1% of the total recorded videos of each sett. This criterion was defined to reduce the risk of counting, as members of the monitored group, individuals from other social groups that may be visiting the main sett of another group. This conservative approach considers that all badgers detected simultaneously belong to the same social group. Animals 2021,11, 2663 6 of 15 Table 2. Variables used to characterize each sett during the monthly monitoring visit, with reference to the hypothesis to which they were linked, the type of variable and its range/units. Variable Description Type Range/Units H2 Tree Tree cover Ordinal 1–5 (1—low; 5—high) 1,2 Shrubs Shrub cover Continuous Percentage 1,2 Herbs Herbaceous cover Continuous Percentage 1,2 Shrub_H Shrub mean high Continuous Cm 1,2 Bare_soil Bare soil cover Continuous Percentage 1,2 Veg_div Vegetation diversity Continuous 0–1 (Simpson diversity index) Understory Dominant understory Categorical 4 Categories (Cattle pastures, Shrubland, Natural pastures, Natural pastures with shrubs) Land use Type of land use Categorical 4 categories (Cork oak woodland, Mixes wood, Natural pasture, Pine forest) H3 Soil_mat Main soil material Categorical 4 categories (Sand, Rock, Silt/Clay, Roads 3)1 Soil Type of soil 2 Categories (Podzols, Regosols) H4 Dist_ol Distance to the nearest olive yard Continuous m Dist_piv Distance to the nearest watering pivot Continuous m Dist_feed Distance to the nearest game artificial feeders Continuous m Dist_wat Distance to the nearest water source Continuous m H5 Grazz_16–17 Grazing pressure between 2016–2017 Continuous Grazing intensity 4 Grazz_07–17 Cumulative grazing pressure between 2007–2017 Continuous Grazing intensity 4 N_grazz Number of consecutive years, since 2007, without cattle grazing Continuous Number of years Dist_road Distance to the nearest paved road Continuous m Dist_road2 Distance to the nearest dirt road Continuous m H6 Dist_sett Distance to the nearest badger sett Continuous m Wildboar Wild boar abundance Continuous Number of wild boar signs [40] 1 Variables collected in situ, with the minimum convex polygon encompassing all the sett entrances. 2 Variables used in the Principal Components Analysis (PCA). 3 Setts located under dirt roads. 4 Expressed as the number of cattle per day and per hectare, in the pot where each sett was located [35]. Animals 2021,11, 2663 7 of 15 Regarding reproduction, we considered that a mating event occurred when we detected in the video a sexual interaction between two adult badgers. When in subsequent videos the mating event still occurred, we registered the entire sequence as a single mating event. Mating events less than 2 minutes long were considered short-duration mating or copulation events, and those that occurred for more than 2 minutes (typically 12 minutes or more; see results) were registered as long-duration mating events [ 34 ]. Long mating events are expected to increase breeding success [41]. 2.6. Data Analysis 2.6.1. Drivers of Sett Use In the modelling procedure applied for assessing the environmental drivers of sett use by badgers (hypotheses H2–H6), the number of monthly visits positive for signs of activity (ranging from 0 to 7) was included as dependent variable (Tables 1and 2). Due to the high number of variables associate to H2 we conducted a Principal Component Analysis (PCA) [ 42 ], using the variables tree, shrub, herbaceous and bare soil cover and shrub height (Table 2); principal components that cumulatively explained >80% of the information of the original variables, were used as candidate drivers in the subsequent modelling procedure (see 3.1). Spatial autocorrelation of data was assessed using the Moran I Index [ 43 ], and we evaluated multicollinearity between all candidate drivers (except those used in the PCA) using the Spearman correlation coefficient, ρ [ 42 ]. When two variables were highly correlated (ρ> 0.70), we excluded the one less correlated with the dependent variable [44]. For each hypothesis, we run Generalized Linear Models (GLM) [ 42 ], using a Poisson distribution and a “log” link function, corresponding to all combination of the variables associated to the hypothesis (Table 2). Resulting models were ranked according to the Akaike Information Criterion, with correction for small sample size (AICc) [ 45 ]; those that showed a ∆ AICc < 2 (i.e., the difference between the AICc of a model and the lowest AICc value in the model set [ 46 ]) were considered the best models in explaining sett use variation. If more than one model fulfilled this criterion, we used a model averaging procedure [47] to assess variables average coefficient and the 95% confident interval (95% CI). Variables with a 95% CI that did not include the 0 (i.e., we can determine more precisely if its effect is positive or negative), and that showed a higher relative importance, were considered more influential [ 48 ]. Relative importance was estimated as the sum of the Akaike weight (w; probability of a model being the best model) [ 45 ], of all the models that included the variable of interest. The variables that, for each hypothesis, showed a significant (p< 0.05) or almost significant (p< 0.1) influence on sett use pattern, were included in overall models (combined hypothesis), that postulated that the pattern of sett use is determined by a combination of drivers linked to distinct ecological/environmental processes (vegetation, soil, food, disturbance or competition). Finally, we compared the AICc of the best model of each hypothesis and considered the hypothesis with more support as that showing the lowest AICc value. The goodness of fit of the best overall model was assessed by estimating the R 2 , which identifies the proportion of variability of the original data explained by the model [ 47 ]. Finally, we also tested the autocorrelation of models residuals [46]. 2.6.2. Badger Density Estimation We used the maximum number of different animals detected in each main sett (see above) to estimate the mean number of animals per social group for the entire study area. We then estimated badger density as: Mean number of adult badgers per main sett estimated for the study area ×Estimated number of social groups Surveyed area 43 km2 Animals 2021,11, 2663 8 of 15 We assumed that the studied badger population followed the behavioral pattern described for other areas of its distribution range, with each social group having one main sett, used all year round and where reproduction occurs [49]. 3. Results Of a total of 33 setts identified in the area in the frame of previous studies, only 22 still occurred (i.e., were not destroyed). Twenty-three new setts were added as a result of the field work, totalizing 45 setts for monitoring purposes (Figure 1B), resulting in a sett density of 1.05 setts/km2. 3.1. Drivers of Sett Use Pattern From the 45 setts we detected in the study area, we only used 34 in our modeling procedure, since 11 were only detected close to the end of the study not providing robust data regarding their use by badgers. No significant spatial autocorrelation was detected (Moran I = − 0.053; p= 0.822). From all the candidate co-variates we excluded “Grazz_07–17” and “Grazz”, since they were significantly correlated with “Grazz_16–17 (rGrass_07–17–Grazz = 0.77, p< 0.001; rGrass_07–17–Grazz_16–17 = 0.77, p< 0.001; rGrazz–Grazz_16–17 = 0.79 , p< 0.001), and the later was more correlated with the dependent variable (rs = 0.258). Furthermore, we also excluded “Dist_piv” as it was significantly correlated to “Dist_ol” (r = 0.97, p< 0.001), and the “Dist_ol” was more correlated to the frequency of sett use (rs = −0.237). The first two compoments of PCA analysis produced for the vegetation associated variables reached a cumulative explained variance of 80.1% and thus were used as candidate variables in models procuced to test Hypothesis 2 (Vegetation drivers). The association between both components and the original variables is presented in Table 3. Table 3. Correlation between the original variables and the PCA components used in the modeling procedure. Variable description is detailed in Table 2. Original Variable Component 1 Component 2 Tree −0.431 - Shrubs −0.501 −0.398 Herbs −0.507 −0.132 Shrub_H −0.547 −0.271 Bare_soil −0.114 0.865 We produced a total of 91 models, but only a few were considered the best models (i.e., ∆ AICc < 2) for each hypothesis: four for H2, one for H3, two for H4, three for H5 and two for H6 (Table 4). From the variables included in these best models, only “Vegetation diversity”, “Distance to the nearest olive yards”, ”Main soil material”, “Distance to the nearest paved road”, “Distance to the nearest dirt road” and “Grazing pressure in 2016–2017 ” had a significant influence in the dependent variable (p< 0.001) and thus were used as candidate variables for the combined model (Table 4). Models for the combined hypothesis showed the highest fit (i.e., lower AICc and Overall ∆ AICc < 2). Best models included the variables “Main soil material”, “Vegetation diversity”, “Distance to the nearest dirt road” and “Grazing pressure in 2016–2017”. Based on the model averaging procedure we can state that setts that were more used by badgers during the study period were located in sites with a higher diversity of vegetation (i.e., higher Simpson diversity index), higher levels of grazing in the 2016–2017, and softer soils (e.g., silt/clay soils were avoided) respect to poorly used setts (Table 5). The best model was able to explain ca. 40% of the variance in the frequency of sett use (R 2 = 0.397), indicating an acceptable goodness of fit [ 42 ], and no spatial autocorrelation of residuals was detected (Moran I = −0.09; p= 0.55). Animals 2021,11, 2663 9 of 15 Table 4. Best models for each tested hypothesis ( ∆ AICc < 2). For each model we include the degree of freedom (df), the Akaike Information Criterion, corrected for small samples (AICc), the AICc difference for the lowest AICc in each hypothesis ( ∆ AICc), the Akaike weight (Weight), and the AICc difference for the lowest AICc of all produced models (Overall ∆ AICc). Variables’ acronyms are described in Table 2. Model df LogLik AICc ∆AICc Weight Overall ∆AICc Null 1 −87.721 169.6 0 - 11.0 H2—Vegetation Understory + Veg_div + PCA1 6 −73.950 163.0 0 0.209 4.4 Understory + Veg_div 5 −75.574 163.3 0.3 0.182 4.7 Veg_div 2 −70.981 164.3 1.3 0.107 5.7 Veg_div + PCA1 3 −78.977 164.8 1.8 0.088 6.2 H3—Soil Soil_mat 4 −75.027 159.4 0 0.781 0.8 H4—Food Dist_ol 2 −82.114 168.6 0 0.346 10.0 Dist_ol + Dist_wat 3 −81.902 170.6 2.0 0.128 12.0 H5—Disturbance Dist_road2 + Grazz_16–17 3 −78.930 164.7 0 0.255 6.1 Dist_road2 + Grazz_16–17 + N_grazz 4−78.370 166.1 1.4 0.123 7.5 Grazz_16–17 2 −81.036 166.5 1.8 0.104 7.9 H6—Competition Dist_sett 2 −83.464 171.3 0 0.223 12.7 Wildboar 2 −83.715 171.8 0.5 0.174 13.2 Combined hypothesis Soil_mat + Veg_div 5 −73.204 158.6 0 0.151 0 Dist_road2 + Grazz_16-17 + Veg_div 4 −74.792 159.0 0.4 0.123 0.4 Soil_mat + Grazz_16-17 + Veg_div 6 −72.095 159.3 0.7 0.104 0.7 Soil_mat 4 −75.027 159.4 0.8 0.097 0.8 Veg_div + Grazz_16-17 3 −76.340 159.5 0.9 0.095 0.9 Table 5. Variables included in the average model of the best models produced to explain the variation in sett use frequency (Combined hypothesis—Table 4). For each variable the coefficient ( β ), standard error (SE), z-value and significance (p), 95% confidence intervals (95% CI) and relative importance are presented. (See Table 2for variables’ description). Variable βSE z-Value p95% CI Relative Importance Intercept 0.719 0.601 1.175 0.240 [−0.480, 1.919] - Soil_mat (Silt/clay) −0.740 0.244 2.919 0.004 [−1.237, −0.243] 0.62 Soil_mat (Rock) −0.115 0.246 0.450 0.653 [−0.618, 0.388] 0.62 Soil_mat (Roads) 0.264 0.403 0.629 0.530 [−0.559, 1.087] 0.62 Veg_div 1.989 0.862 2.122 0.034 [0.145, 3.651] 0.83 Dist_road2 −0.157 0.091 1.662 0.096 [−0.342, 0.028] 0.22 Grazz_16-17 0.194 0.092 2.041 0.041 [0.007, 0.381] 0.56 3.2. Social Groups Size and Density Estimates We identified 12 setts that showed a higher and more continuous activity patterns (Figure 1C) and were therefore considered main setts. On average main setts included 3.85 ±3.32 (ranging from 1–7) entrances, 3 ± 2.2 of which were active simultaneously (i.e., with footprints, latrines, revolved soil, bedding material; range 1–6). Secondary setts had only an average of 1.41 ± 1.66 active entrances, ranging from 0 to 4 (often none showed signs of activity), from an overall 2.93 ± 2.80 total entrances (range: 1–9). The index of sett use was higher for main setts (6.14 ±1.22) than for secondary setts (3.37 ±2.23). Overall, we manage to register badgers activities in main setts on average 150 ±26 night per set, representing 18,966 videos of 30 seconds (i.e., 158 hours of video). In 46% of the videos (n= 8654) we detected at least one badger, and in 7% (n= 1281) other mammal species: cattle, wild boars, wild rabbit (Oryctolagus cuniculus), common genet (Genetta genetta), stone marten (Martes foina), red fox (Vulpes vulpes) and Egyptian mongoose (Her-