Local-scale habitat configuration makes a niche for wildlife encroaching into an urban landscape: grubbing sites of wild boar Sus scrofa in a city matrix

Urban environments may offer certain species diverse and abundant food resources of natural and anthropogenic origin. However, the local-scale configuration of habitats and urban infrastructure may influence foraging decisions regardless of the availability of food. In recent years, the expansion of wild boar Sus scrofa into areas significantly transformed by humans has been observed in many parts of its range. Grubbing (rooting) is a major foraging mode of the species, during which disturbance of the upper soil layers enables these animals to find and consume food items. However, the factors that determine the selection of grubbing sites in the urban landscape, where the balance between food availability and the avoidance of humans may influence foraging decisions, are not known. Our aim was to identify local-scale factors that influence grubbing site selection and the size of grubbed patches in an urban landscape. The characteristics of 108 wild boar grubbing sites in the city of Kraków (Poland) were compared to randomly selected control sites. The probable presence of a grubbing site was positively correlated with the proportion of meadows and fallow land in the vicinity and with increases in both canopy cover and distance to pavements. The size of a grubbed patch was positively correlated with the percentage of meadows in the vicinity, increasing distance to buildings and decreasing distance to pavements. We found a non-random pattern of grubbing sites in the urban landscape and indicated that the local-scale configuration of vegetation and urban infrastructure contribute to foraging site selection by wild boar. Our study highlights that the encroachment of wildlife into the urban landscape is a complex process, driven by both resource availability and the avoidance of human-related disturbances.


Introduction
The wild boar Sus scrofa is a highly plastic species, capable of adapting its activities and diet to local conditions. It is thus able to colonize a broad spectrum of habitats, from natural to anthropogenic, the latter including urban areas (Sjarmidi and Gerard 1988;Podgórski et al. 2013;Ballari et al. 2015). Being endowed with an excellent spatial memory and sense of smell (Maselli et al. 2014), wild boar have developed complex strategies for moving around, which take into account food availability, weather conditions, terrain and human pressure . In urban areas, wild boar find a wide variety of foods, both natural and anthropogenic (Stillfried et al. 2017b). Another aspect favouring the presence of these animals in such areas is the minimal pressure from hunting, which is severely restricted for safety reasons; and in any case, animal culling meets with widespread public disapproval (Tack 2018). Moreover, predation pressure from large carnivores, normally a significant factor reducing wild boar populations, is usually much reduced in urban areas (Tack 2018). In consequence, numbers of wild boar have increased very considerably during the last 20 years and the animals have successfully colonized urban areas throughout their range of distribution (Cahill et al. 2012;Tack 2018).
The wild boar is a food generalist, whose diet depends on the season and the resources available in a given place (Schley and Roper 2003). By far the greatest part of the wild boar's diet consists of plant-based food, mostly garnered in crop fields; animal food -mainly invertebrates, small mammals and birds' eggs -is eaten in much smaller amounts (Herrero et al. 2006;Ballari and Barrios-García 2014). Wild boar normally consume the above-ground parts of plants, but if these are less abundant or unavailable, they grub within the soil in search of the underground parts of plants or animals (Sandom et al. 2013;Petrelli et al 2022). Grubbing (or rooting) is a particularly important foraging technique that disturbs the litter or turf and the upper soil layers. Wild boar also employ grubbing when there is a local and/or periodic surplus of organisms that are an important source of energy in their diet, such as insects (Genov 1981;Laznik and Trdan 2014;Scandurra et al. 2016) or rodents, which may be eaten along with food they have stored (Focardi et al. 2000), and also parts of plants like geophytes (Palacio et al. 2013).
In the presence of a habitat mosaic and varied degrees of human pressure, animals are compelled to modify their spatial and temporal utilization of the land (Lowry et al. 2013). The movements of wild boar are governed mostly by the need to find food (Toger et al. 2018), but in contrast to areas devoid of human pressure, where wild boar are active 24 h a day (Podgórski et al. 2013), in urban areas they feed mainly at night (Cahill et al. 2003), which leads to temporal segregation between them and humans (Johann et al. 2020). Also, in order to avoid predator attacks or encounters with humans while hunting, the activity of wild boar is spatially differentiated, giving rise to spatial segregation between them and potential sources of mortality (Brown et al. 1999;Keuling et al. 2008;Theuerkauf and Rouys 2008). Human hunting pressure and farming activities such as crop harvesting force wild boar to move between habitats (Dardaillon 1986;Tolon et al. 2009) and even to temporarily abandon previously occupied land (Sodeikat and Pohlmeyer 2003;Thurfjell et al. 2009). Finally, wild boar usually avoid those urban landscapes with high levels of human pressure, preferring sparsely populated areas with accessible grassland and ruderal habitats (Honda 2009;Baś et al. 2017).
On the one hand, the foraging ecology of wild boar has been well studied in woodland and in open, farming areas (Massei et al. 1996;Geisser and Reyer 2004;Ballari and Barrios-García 2014). On the other, little is known about this aspect of the wild boar's biology in highly urbanized habitats, where plentiful sources of natural and anthropogenic food encourage these animals to encroach into them (Stillfried et al. 2017b;Castillo-Contreras et al. 2018). Against this, the presence of people and their dogs, as well as linear barriers like roads and buildings, can hamper their free movement when searching for food (Ohashi et al. 2013;Thurfjell et al. 2015). In doing so, wild boar may damage urban greenery, both private and municipal, situations that may provoke conflict (Barrios-Garcia and Ballari 2012). Such damage to urban greenery, not to mention collisions with road vehicles (Zuberogoitia et al. 2014) and occasional attacks on people (Chauhan et al. 2009), combine to engender negative attitudes among the public towards these animals (Kotulski and König 2008); at best, the wild boar are scared off, but if persistently causing problems, may be shot (Geisser and Reyer 2004;Conejero et al. 2019).
The foraging of wild boar in the urban landscape is a complex process, driven by the availability of suitable, good quality habitat and food resources, the spatial arrangement of urban infrastructure and socio-economic aspects on the one hand, and the avoidance of human-related disturbances on the other. The aim of this work was to evaluate the habitat and anthropogenic factors that mediate grubbing site selection by wild boar and the size of grubbed patches in the urban landscape. Our working hypothesis therefore predicted that the choice faced by wild boar with regard to their foraging sites would be a trade-off between the availability of food-rich habitats and the avoidance of anthropogenic disturbances. Consequently, wild boar should prefer habitats with potentially large food resources, including meadows and fallow land with nut-and fleshy fruit-bearing trees. However, the species should avoid areas with high levels of human pressure in the vicinity of buildings and transportation routes and select sites that offer potential side and top cover, which may influence the animals' feeling of security during foraging.

Study area
This study was carried out in Kraków (50°05' N, 19°55' E), the second largest city in Poland (Fig. 1). It has a population of 774 839, a population density of 2371 persons/km 2 and covers an area of 326.85 km 2 (GUS 2019). The area within the city's boundaries consists of a densely built-up city centre with limited green spaces, suburbs with a moderate number of buildings and different types of greenery, and scattered buildings typical of a rural landscape that are surrounded by areas of natural and semi-natural vegetation. 47% of the city's overall area consists of densely built-up areas with their accompanying urban greenery (gardens, squares, road verges, playgrounds, allotments and orchards, parks and cemeteries); 37% of this area comprises open spaces like arable land (14%), spontaneous vegetation on fallow land (13%), meadows and pastures (8%) and wetland vegetation (2%); 11% of the area is wooded, and includes natural and semi-natural scrub (5%), deciduous and mixed forest (4%), and damp, riparian forest and transformed tree stands (2%).
In the period of 2010-2019, the mean annual temperature in the Kraków was 9.5 °C. July was the warmest month (mean air temperature 19.8 °C) and January was the coldest one (-1.9 °C). Average temperatures are higher in the city centre than on the outskirts (Matuszko and Piotrowicz 2015). The average annual rainfall was ca 653 mm, and precipitation occurred on average on 160 days a year. Mean annual precipitation was the highest in July (98 mm) and the lowest in February (26 mm). Snow cover lasted about 37 days on average; in 2019, there were 18 days with snow cover (UMK 2019). During the winter preceding the study period (2018/2019), the mean snow cover was 1.8 cm thick and melted in mid-February. During the period of field works (April-May 2019) mean temperature was 11.3 °C and precipitation was 200 mm.

Field methods
The area within the administrative borders of Kraków was first divided into a grid of 1 × 1 km squares based on the point with coordinates 50°N 20°E. Then, one hundred sample plots (1 km 2 ) on which grubbed patches of ground were to be surveyed ( Fig. 1) were selected at random. Between 3 April and 26 May 2019, grubbing sites were searched for on all the sample plots, where all suitable habitats, i.e. urban greenery, woodland, parks, meadows, fallow land and farmland, were present. Not included were sites such as fenced gardens and military areas, to which access was impossible. The route taken by the observer on a sample plot was recorded on a GPS device. The mean distance he/ she covered per sample plot was 4266.8 m (± 1295.7 SD; range 569.9-7347.7), his/her mean speed of movement over the sample plot was ca 4 km/h, and the mean duration of a sample plot survey was ca 1 h. The distance covered and the amount of time allocated to the survey of a given sample plot resulted from the diversification and location of potentially suitable habitats within the urban landscape (single, large and easily accessible patches vs. numerous, small and scattered over the plot area). The exact positions of the grubbing sites were recorded on the GPS device. A grubbing site was regarded as one where the foraging of wild boar had led to the soil being turned over and/or exposed, and such sites were considered separate if they were at least 100 m apart, or 50 m apart if there was an intervening obstacle such as a building, fence, road or river. The adopted distances were intended to assure independence of measurements on grubbing sites and were approximately 200% larger than the radius of the largest grubbed patch found during the fieldwork. To avoid overrepresentation of data from intensively grubbed plots, a maximum of 5 grubbing sites were recorded on a given sample plot.
When a grubbing site was identified, its area was measured and an inventory of its habitat features compiled in a 10 m radius buffer zone around the site (Table 1). For each grubbing site found on a sample plot, an adjacent, random reference site was selected with habitat parameters identical to those on the grubbing site. The positions of the reference sites were established at random within the borders of a sample plot in places accessible to wild boar (excluding fenced areas) and in habitats where grubbing was possible (i.e. not in large areas covered with an impervious surface, e.g. pavements, roads, car parks). A minimum distance of 100 m was set between a grubbing site and its associated reference site.

Habitat variables
The actual habitat characteristics of the grubbing sites and paired reference sites were established during the fieldwork ( Table 1). The set of habitat variables potentially influencing grubbing site selection was determined on the basis of the  Lombardini et al. 2017). The descriptions of the grubbing sites include the features characterizing the side and top cover, which may affect the animals' feeling of security during foraging. Coverage by tall grasses, herbaceous plants and shrubs/trees, which can all act as a side cover for the feeding site was assessed (Table 1). Wild boars forage on the above-ground parts of plants, their fruits and seeds, and also on their roots (Cuevas et al. 2010(Cuevas et al. , 2012Ballari and Barrios-García 2014). The seeds of trees such as beech Fagus spp. and oak Quercus spp. are an important component of the wild boar's diet (Herrero et al. 2005), and the intensity of grubbing increases in seed-rich areas (Groot Bruinderink and Hazebroek 1996). Surface waters and wetlands are likely to encourage the presence of wild boar (Borowik et al. 2013), as the animals have opportunities to feed and bathe there (Giménez-Anaya et al. 2008). Hence, the distances to potential food/water sources and the scale of their availability (percentage) at each grubbing site were determined (Table 1).
Arable land provides food for wild boars and can have a significant impact on their movements between patches of suitable habitat . In the urban landscape, wild boars show a strong preference for areas of a natural or semi-natural character, such as forests or meadows (Stillfried et al. 2017a), and they usually avoid human-related structures (Honda 2009;Baś et al. 2017). Even so, the areas used by wild boar in cities are often located in close proximity to households or roads (Stillfried et al. 2017a). Moreover, anthropogenic waste and rubbish are potential sources of food for them (Stillfried et al. 2017b). With respect to the above, the distances of grubbing sites from structures related to human activity, possibly having an impact on habitat selection and foraging intensity, were measured (Table 1). Potential food sources like game baiting sites and infrastructure components that might have a deterrent effect, such as roads and buildings, were also surveyed (Table 1).

Data handling and analyses
The differences between the habitat variables at the grubbing and random sites were analysed using Student's t test. The relationships between the habitat variables and the probability of occurrence of grubbing sites were analysed using generalized linear models with binominal distribution and log link function. The relationships between the habitat variables and the size of grubbing sites were analysed using generalized linear models with a Gaussian distribution and log link function. Multicollinearity between variables was controlled with Pearson's rank correlation matrix prior to the modelling procedure. The correlation between pairs of variables did not exceed 0.5. The set of all habitat variables (Table 1) was used as continuous variables in the starting model. Backward selection was applied to obtain the final model with all variables significant at p < 0.05.

Results
Grubbed patches of ground resulting from the foraging of wild boar were recorded on 45% of the sample plots (N = 100), randomly selected within the city borders. In total, 108 such grubbed patches were found with a mean of 2.4 patches per plot (± 1.2 SD, range 1 -5). The mean surface area of the grubbing sites was 177.3 m 2 (± 477.2 SD, range 1 -3700). The grubbing sites were situated significantly farther away from buildings and pavements than the random sites (Table 2). In addition, there was a smaller percentage of arable land in areas with grubbing sites than with random sites (the difference approached significance, Table 2). The presence of wild boar grubbing sites was positively correlated with the percentages of meadows and fallow land  (Table 3). Moreover, increases in both canopy cover and distance to pavements were positively correlated with the presence of grubbing sites (Table 3). The size of wild boar grubbing sites was positively correlated with the percentage of meadows in the vicinity (Table 4) and also with distance to buildings, but negatively with distance to pavements (Table 4). The presence of trees and shrubs that produce seeds or fruits that are potentially an attractive source of food for wild boar was not correlated with either the probability of grubbing sites being present (Table 3) or their size (Table 4).

Discussion
The results of the study indicate that wild boars commonly search for food of natural origin in an urban landscape, but their selection of foraging (grubbing) sites is mediated by the local-scale composition of habitat types and anthropogenic structures. Although the distribution of wild boars is determined primarily by the availability of sites offering food resources , the presence of natural shelters for hiding and resting in is an important driver of changes in the spatio-temporal locations of wild boar foraging sites (Thurfjell et al. 2009). Woodland habitats and their margins, especially deciduous and mixed forests, are regarded as the wild boar's preferred biotope, because fruit-and nut-bearing trees, invertebrates and plant tubers are readily available in them, and also because its complex spatial structure offers sufficient shelter (Meriggi and Sacchi 2001;Honda 2009). The grubbing sites in the study area were situated mainly in meadows and on fallow land, i.e. habitats that appear to offer an abundant underground biomass of the plants and/or invertebrates that make up the wild boar's diet (Laznik and Trdan 2014;Tack 2018). However, as these habitats are subject to the natural succession of vegetation, they are often locally covered with a mosaic of shrub vegetation, grasses and other herbaceous plants that can provide cover for wild boar searching for shelter within the city. The grubbing sites we surveyed were situated along a broad urbanization gradient, from the moderately built-up outskirts of the city to areas near the city centre (Fig. 1). But in the mosaic of urban habitats there is potential danger from a number of sources, such as road traffic, barriers to migration, as well as noise and light pollution (Ciach and Fröhlich 2019). In addition, the presence and activities of people compel the animals to greater vigilance, causing them to move around more often (Padié et al. 2015). Even though wild boar inhabiting urban areas are more tolerant towards the presence of humans than their rural counterparts, they avoid highly urbanized areas and manmade structures (Marino and Colvin 2015;Stillfried et al. 2017a). The patches of ground grubbed up by wild boar were situated at a significantly greater distance from pavements than the random sites, and a denser tree-canopy closure also increased the probability of grubbing sites being present. The network of pavements crossing natural and semi-natural habitats that have managed to persist in the city matrix are associated with the greater activity of people (walkers, cyclists). Wild boar thus perceive the man-made infrastructure as a risk-laden component of the "urban landscape of fear" (Gaynor et al. 2018). As already mentioned, opportunities for concealment can be provided by patches of meadows and fallow land that are overgrown with dense vegetation. However, as such dense vegetation is unattractive to most humans, who are thus unlikely to enter it, it becomes a fairly safe haven for wildlife in the urban landscape. Moreover, watercourses and railway or road verges with their associated dense vegetation are often the corridors along which animals can penetrate highly urbanized environments (Ignatieva et al. 2011). Such habitats provide shelter for wild boar, that might influence these animals to move between natural habitats and urban areas and look for food resources (Castillo-Contreras et al. 2018).
We found considerable variation in the size of the grubbed patches, which increased along with the increasing percentage of meadows in the vicinity and increasing distance to buildings but with decreasing distance to pavements. The magnitude of wild boar grubbing areas varies widely between habitat and soil types, and is annually and seasonally variable (Welander 2000). However, the size of grubbed patches also depends on the abundance of food available and the time spent foraging (Laznik and Trdan 2014). In areas with greater predation or human hunting pressure, wild boar are more vigilant and minimize their foraging intensity (Brown et al. 1999;Theuerkauf and Rouys 2008). In urban areas, a factor significantly reducing the feeling of safety, and thus the intensity of foraging, is the presence of humans (Gaynor et al. 2018). In our study, the size of grubbed patches increased with increasing distance to buildings, near which wild boar could be scared off by people. Surprisingly, grubbing sites near pavements were larger, even though the very presence of the pavements reduced the probability of grubbing sites being made there. Having a high level of ecological and behavioural plasticity (Fulgione et al. 2017), urban wild boar can cope with the presence of humans and work out strategies for avoiding contacts with people, e.g. by being active at night (Podgórski et al. 2013;. Some pavements by busy roads on the city's outskirts are rarely used by people, particularly at night, which is when wild boar usually forage. Moreover, some pavements are protected by acoustic screens, which may provide side cover for foraging wild boar. Therefore, habitats located in close proximity to pavements appear to provide attractive food resources and, when people are not close by, are commonly used by wild boars. The availability of high-energy anthropogenic food is regarded as a factor that encourages wild boar to enter urban landscapes (Stillfried et al. 2017b;Castillo-Contreras et al. 2018). In our study area, the presence of wild boar was not correlated with the proportion of arable land, even though food derived from crops of maize, potatoes and cereals makes up a key fraction of the wild boar's diet (Schley and Roper 2003;Thurfjell et al. 2009). Outside the growing season, the biomass of crops on cultivated fields is low, as is that of underground sources of food, so wild boar do not seek food there (Schley et al. 2008). Since the variable availability of food is a driver of habitat use by wild boar (Sütő et al. 2020), one might expect the selection of foraging sites in the urban landscape to differ between phenological seasons. Stillfried et al. (2017b) found that the diet of urban wild boar is dominated by naturally-occurring food, and that anthropogenic food is supplementary. Wild boar may search for new foraging areas, where human-related food resources are provided both intentionally as supplementary feed for wildlife or unintentionally as refuse deposited in rubbish bins or fruits produced by urban trees. Although some of the grubbing sites were located in urban greenery of highly urbanized areas near the city centre, we did not discover any grubbed patches near building structures. Therefore, our study suggests that urban wild boar do not search there for supplementary feed or refuse, or that such resources are inaccessible to them. Potential food in the form of seeds, nuts or fleshy fruits produced, e.g. by oak, walnut or cherry trees planted purely for their aesthetic effect are easily available during autumn when they are lying on the ground beneath the trees. However, our study indicates that the presence of trees and shrubs that produce nuts, seeds or fleshy fruits has no influence on either the probable presence of grubbing sites or their size. It is possible, however, that wild boar pick fruits off the ground surface without disturbing it, which reduces the chances of our discovering their foraging areas. It should be noted, that the consumption of human-related food resources by wild boar has been observed in many parts of its range (Cahill et al. 2012) and human activities might support synanthropic generalists (Shochat et al. 2006).
Grubbing disrupts the structure of the soil, altering its physico-chemical properties, so that where wild boar densities are high, their intensive grubbing exacerbates soil erosion (Wirthner et al. 2012). Grubbing also affects the quantities of seeds in the soil (Bueno et al. 2011) and can alter the environmental conditions essential for the growth of many plants, including non-native species (Brunet et al. 2016;Horčičková et al. 2019). Moreover, foraging by wild boar can significantly reduce populations of locally occurring animals (Amori et al. 2016;Casula et al. 2017). In consequence, such foraging may significantly affect the ecosystem (Barrios-Garcia and Ballari 2012) by altering the species composition, the structures of plant communities and the assemblages of animals inhabiting patches of vegetation and contributing to the spread of alien species. At the same time, it should be noted that in cities where wild boars are found, the media regularly reports on individuals or herds that have close contact with humans, e.g. rummage through rubbish, overturn bins, wallow in sandpits or cool off under fountains or in pools. Although such observations are still more anecdotal than strictly scientific, it should be assumed that wild boars, as cognitively complex animals, will progressively lose their fear of humans over time and increasingly use highly urbanized areas.

Conclusions
Populations and ranges of wild boar have increased significantly in recent decades (Massei et al. 2015). This is correlated with higher average temperatures in winter and the greater availability of food, especially as a result of the increasing area of maize cultivation (Geisser and Reyer 2005). Urbanization leads to the emergence of a habitat mosaic composed of variously utilized open areas adjoining more or less wooded areas. These offer suitable foraging sites and shelter, so they may be an additional factor encouraging the expansion of wild boar. The results of the present study indicate, however, that urban wild boar utilize sites that are influenced as little as possible by humans. In the urban landscape, therefore, the preference for foraging habitats is mediated by the presence of anthropogenic infrastructure such as pavements, which being a source of human-related disturbance are avoided. This strategy ensures the relatively safe acquisition of food by preventing direct contact with humans. Our study highlights that the encroachment of wildlife into the urban landscape is a complex process, driven by both resource availability and the avoidance of human-related disturbances.