Introduction

Humans provide forage to ungulates both intentionally, by delivering it to feeding sites, and unintentionally, by leaving food leftovers at compost piles and rubbish dumps (Smith 2001; Sorensen et al. 2014) or not protecting agricultural crops (Oro et al. 2013; Milner et al. 2014). Winter supplementary feeding, which involves providing forage to feeding sites during winter, is a widely practiced management approach for ungulates in temperate ecosystems of Europe and North America (Milner et al. 2014). In temperate latitudes, during winter, ungulates are exposed to higher thermoregulatory and mobility costs due to harsh weather conditions expressed in low ambient temperature and snow cover (Jensen et al. 1999; Arnold et al. 2006; Signer et al. 2011; Ossi et al. 2015). However, since supplementary fodder is an alternative and easily accessible food resource in winter, it allows ungulates to maintain high body condition, thereby enhancing their reproduction, survival (Krasińska and Krasiński 2013; Peterson and Messmer 2007) and trophy quality (Putman and Staines 2004; Cooper et al. 2006). Nevertheless, winter supplementary feeding of ungulates is also implemented for various other management goals. For instance, diversionary feeding aims at modification of ungulate movement and spatial behaviour to decrease their impact on forest and agricultural crops (van Beest et al. 2010a; Smith et al. 2001; Náhlik et al. 2005; Katona et al. 2014; Krasińska and Krasiński 2013; Jaroszewicz et al. 2017), limit ungulate-vehicle collisions (Wood and Wolfe 1988; Andreassen et al. 2005) or prevent disease transmission between livestock and wildlife (Brook 2008; Sorensen et al. 2014).

Undeniably, supplementary and diversionary feeding practices demontrated various benefits (e.g. reduced number of mule deer Odocoileus hemionus collisions along highways in Utah, USA (Wood and Wolfe 1988); increased winter survival (Krasińska and Krasiński 2013)), however, these practices have often failed to achieve their intended management goals (Cooper et al. 2006; van Beest et al. 2010b; Kowalczyk et al. 2011; Felton et al. 2017). Moreover, supplementary feeding can lead to unintended outcomes, such as aggregation around feeding sites, which may increase probability of intra- and interspecific disease and parasite transmission (Sorensen et al. 2014; Kołodziej-Sobocińska et al. 2016b), elevate impact of ungulates on vegetation composition and structure near feeding sites (Dunkley and Cattet 2003; Côté et al. 2004; Putman and Staines 2004; Mathisen and Skarpe 2011) and increase stress level caused by competitive feeding behaviour (Ceacero et al. 2012). Furthermore, by consuming supplementary forage prepared from sources from outside of the targeted area (i.e. hay), wild herbivores can contribute to seed dispersal of non-native vegetation seeds (Jaroszewicz et al. 2017).

Although in many places supplementary feeding was continuously delivered to feeding sites during periods of natural forage shortage, their utilization exhibited high spatio-temporal variation. Ungulates were more prone to visit feeding sites that provided high-quality fodder (Rajský et al. 2008) or those that have been present in the environment for longer duration and they were familiar with (Gundersen et al. 2004; Ranc et al. 2021). In addition, the visitation frequency to feeding sites by ungulates in winter was also positively associated with the severity of weather conditions. However, the effect of weather conditions on feeding sites visitation was mostly pronounced in small-bodied species or in young which body condition and survival are strongly dependent on weather conditions in winter (Gaillard et al. 1993; Loison and Langvatn 1998). For example, roe deer (Capreolus capreolus) used the supplemental feeding sites during harsh environmental conditions (Ossi et al. 2017; Ranc et al. 2021) and decreased utilization of artificial fodder in the time of spring green-up (Ossi et al. 2020), whereas red deer (Cervus elaphus) intensified the intake of supplementary fodder as soon as snow appeared (Náhlik et al. 2005).

Even though the factors affecting winter supplementary feeding in various ungulate species have been broadly studied (Milner et al. 2014), there has been no studies on this problem for European bison (Bison bonasus) which future existence depends on adequate management actions. With our study, we tried to bridge this gap, and additionally, we wanted to highlight the issue of wildlife supplementary feeding in the context of ongoing climate change by studying the winter space use of European bison that are supplementally fed in Białowieża Forest (NE Poland). This is especially important because weather conditions have undergone substantial changes in recent decade due to global warming. Increasing temperature in winter has caused elongation of the growing season and prolonged availability of natural forage for ungulates from northern latitudes (Oechel and Billings 1992).

European bison were reintroduced to Białowieża Forest in 1953 after their extinction in the wild in 1919 (Kowalczyk et al. 2011, 2013). The conservation plan recommended winter supplementary feeding as one of the practices that aimed at maintaining body condition (Kowalczyk et al. 2013) and mitigating conflicts with humans by decreasing the damage made by bison to tree stands and farm crops (Krasińska and Krasiński 2013; Jaroszewicz et al. 2017). Hence, European bison in Białowieża Forest have been supplementally fed at feeding sites throughout the winter period (December-March). We expect this management practice to have a significant effect on the winter space use of bison since the majority of the Białowieża Forest population utilizes these feeding sites during winter (Krasińska and Krasiński 2013). However, the utilization of feeding sites by bison should be associated with the variation in winter conditions. We predicted that (1) bison would concentrate their space use at feeding sites in the middle of the winter season (January and February) when natural forage is remarkably limited and weather conditions are the most severe (low temperatures, presence of a snow cover); (2) independently of the day of winter, bison will be closer to feeding sites on colder days with snow cover due to higher energy requirements and decreased natural food accessibility. However, we expected that (3) the effect of harsh winter conditions should be more pronounced in males than in females. Since females aggregate in large herds in winter, they require abundant food resources, therefore, they should be strongly dependent on utilization of feeding sites regardless of weather conditions. In turn, male utilization of feeding sites should be more weather-dependent because bulls roam solitary or in small groups and, during mild weather, they can utilize less abundant and dispersed food resources (small hay stacks, meadows, anthropogenic food).

Materials and methods

Study area

The study was conducted in the Polish part of the Białowieża Primeval Forest (BPF, 625 km2) located in north-eastern Poland (Fig. 1). BPF is one of the best-preserved temperate forests in Europe (Melis et al. 2007) and one of the very few forests on the European Plain where fragments of close to primeval forest remained (Sabatini et al. 2018; Jaroszewicz et al. 2019). BPF is covered by a mosaic of forest communities (coniferous, fresh and wet deciduous forest, and early successional stands) (Jędrzejewska and Jędrzejewski, 1998) dominated by Norwegian spruce Picea abies (25%), black alder Alnus glutinosa (21%), Scots pine Pinus sylvestris (19%), European hornbeam Carpinus betulus (11%), birch Betula sp. (7%), oaks Quercus sp. (6%) and small-leaved lime Tilia cordata (5%) (Modzelewska et al. 2020). Open habitats (6% of BPF area) include river valleys, forest meadows and glades with human settlements (Fig. 1; Sokołowski 2004; Kowalczyk et al. 2013, 2021). BPF is inhabited by five ungulate species, roe deer, red deer (Cervus elaphus), moose (Alces alces), European bison and wild boar (Sus scrofa), as well as two large predators: grey wolf (Canis lupus) and Eurasian lynx (Lynx lynx). The most abundant species is red deer, while other ungulates occur at lower densities (Borowik et al. 2013; Bubnicki et al. 2019), and the estimated number of European bison in the BPF in 2020 was 715 individuals (Raczyński 2021). European bison in the BPF are supplementally fed during winter (from December to March; Krasińska and Krasiński 2013), which leads to their aggregation around feeding sites (up to 100 individuals; Hayward et al. 2011; Kołodziej-Sobocińska et al. 2016a). Supplementary fodder primarily consists of hay, hay silage and occasionally beetroots (Kowalczyk et al. 2011). The hay for European bison is brought to feeding sites in summer and stored in ruffed haystacks (hay stacked around a tall, central pole) with free access to bison. The hay is usually consumed by bison between November and January, and supplemented by managers until March with different frequency, depending on the location. In the centre of the forest, where the largest herds of bison occur, hay is provided 3–5 times a week while in more distant locations on the edges of the forest usually once or twice a week (Radwan et al. 2010).

Fig. 1
figure 1

The distribution of winter feeding sites and GPS/VHF locations of male and female European bison in winter (December–March) from 2005 to 2012 in Białowieża Primeval Forest

Full size image

The climate in the BPF is transitional between continental and Atlantic types, with clearly marked warm and cold seasons. During the study period (2005–2012), the temperature ranged from –24.4 to 10.6 °C, and for 174 (47%) days, the temperature was below 0 °C. The snow persisted per 139 days a year on average, and its depth varied between 0.01 and 0.55 m. (Czarnecki et al. 2020; IMGW 2021, accessed via “climate” package).

Data collection and processing

In the study, we used tracking data of 43 European bison (24 males and 19 females) equipped with GPS or VHF collars during winter season (from 1 December to 31 March) between 2005 and 2012. Bison were immobilized from a car or hides with a dart gun (Dan-Inject) and etorphine (Arnemo 2003).

For the tracking data of European bison in BPF, we had to consider that 1) tracking duration of each bison was different; 2) during the same time period, both GPS and VHF collars were used to track different bison; and 3) for single individuals, the usage of collars during tracking period was mixed (i.e. only VHF or GPS collar used, transition either from VHF to GPS or from GPS to VHF). Therefore, bison locations (even for the same individuals) were collected with different intensity (VHF: from 1 to 12 locations per day; GPS: 24 locations per day). Moreover, VHF telemetry was conducted only during daylight hours, while GPS tracking was done 24 hours a day. To correct for possible circadian difference in the use of feeding sites by bison, we extracted locations fixed exclusively during daylight hours (between 08:00 am and 04:00 pm) — the period for which both VHF and GPS locations were available (19,906 locations: December to March, 2005–2012). Next, to unify the number of locations collected daily by GPS and VHF collars, for each bison and day we extracted one random location from all locations available for considered daylight hours (5386 bison locations in total).

Then, we georeferenced all bison winter feeding sites (n = 38) in Białowieża Forest and calculated distances between each bison location and the closest feeding site in QGIS software ver. 3.4.15 (QGIS Development Team 2021) (Fig. 1); thus, the number of distances was equal to the number of locations (n = 5386). Since part of the studied individuals could have been members of the same winter herds, their locations could be spatially autocorrelated. To deal with this problem, we decided to perform our analyses on population rather than individual level. For each unique day in the data, we calculated the mean and 95% quantile of available distances, separately for sexes. The calculations of mean and 95% quantile were restricted to days when at least 5 distances of bison to the nearest feeding site were available (range: from 1 to 13 distances per day) for a given sex (males: n = 107 days; females: n = 342 days; in total: n = 449).

Statistical analyses

To verify the association between considered explanatory variables and the mean distance of bison to feeding sites in winter, we fitted a generalized additive model (GAM1) in the R package “mgcv” (Wood 2011). We chose the GAM framework because for different predictors, we expected either linear or non-linear relationship with a dependent variable and GAM allows for flexible testing of these two types of relationships (Wood 2011). We used gamma family with a logarithmic link to conform to the distribution of continuous response variable: mean bison distance to winter feeding sites (n = 449). In GAM1, we included the main effect of the following explanatory variables: day of winter (continuous variable; from 1 December to 31 March for each considered winter period), presence of snow cover (categorical variable; 1, snow present; 0, snow absent), residuals of mean daily temperature (continuous variable) and sex (categorical variable). We also considered interactive effects between day of winter and sex and between residuals of mean daily temperature and sex. As the correlation between considered explanatory variables was moderate (R < 0.49), all variables were retained in the analyses.

Because we expected the day of winter to be nonlinearly related to mean bison distance to feeding sites, we fitted this variable with a spline (smoothing function), while other explanatory variables were fitted as linear (parametric) predictors. We expected limited amount of curvilinearity in the relationship between the day of winter and mean distance of bison to feeding sites; thus, we considered models using splines with a low number of knots (from 3 to 5) and compared them with AIC. The model with knot number equalling 4 had the lowest AIC values and this model was chosen as the best one. As the majority of the variation in mean daily temperature is expressed by seasonal dynamics (higher temperatures in the beginning and the end of winter and lowest in the middle of winter), we removed the seasonal component from the mean daily temperature by extracting residuals from the second-order polynomial regression model testing the relationship between the day of winter and mean daily temperature (LeBreton et al. 2013; García et al. 2020). Extracted residuals were then used as a variable in GAM1 as a residual mean daily temperature. This way we mitigated non-linear collinearity between the day of winter and mean daily temperature and obtained new interpretation of the mean daily temperature. Therefore, we could explain whether higher or lower temperature than expected on any day, based on predictions from the second-order polynomial model, related to the mean distance of bison to the feeding sites. Furthermore, we added the number of bison distances to closest feeding site in each unique day (range: from 5 to 13) as a weight to the GAM1, to give higher importance for mean bison distances which were calculated based on a larger sample size. Due to limited number of location in the majority of the studied winter seasons, we were not able to test the inter-seasonal variation in the mean bison distance to the feeding sites; thus, we added winter season (2004/2005–2011/2012) as a categorical random factor (penalized regression term) to GAM1 (Wood 2011).

Next, we fitted a generalized additive model with gamma error structure (GAM2) with 95% quantile of bison distance to feeding sites as a response variable to predict maximum distance of male and female bison to feeding sites in each considered month (December to March) under warm and cold scenarios. To fit GAM2, we used the same variables and model settings as in GAM1. To predict 95% quantile of bison distance to feeding sites, we used estimates of the GAM2 parameters. Then, we made predictions of 95% quantile for the 15th day of each month (from December till March; 15th, 46th, 76th and 105th day of winter respectively). In (1) warm scenario, the mean daily temperature was set to the maximum residual of mean daily temperature that was registered in a given month and for snow absent, whereas in (2) cold scenario, we used the minimum residual of mean daily temperature in respective month and presence of snow. Predictions for both scenarios were made separately for males and females and averaged over winter seasons. Predicted distances were used as radii for drawing buffers around feeding sites in the study area. Finally, we calculated the area covered by merged buffers to estimate maximum area utilized by male and female European bison in BPF under warm and cold conditions.

Finally, we investigated the relationship between mean bison distance to feeding sites and snow depth. From the dataset used in GAM1-2, we extracted days for which snow depth was at least 1 cm (n = 243), and for this dataset, we fitted a generalized additive model with gamma error structure (GAM3). We set the mean distance of bison to feeding sites as response variables and main and interactive effect of snow depth (continuous variable) and sex (categorical variable) as explanatory (parametric) predictors. Winter season was fitted as a random factor.

To evaluate the assumptions of GAM1-3, we inspected plots presenting distribution of models (deviance) residuals against the fitted values. We did not find any pattern in residual distribution; thus, we considered our models to be fitted correctly. All statistical analyses were done in R ver 4.0.5 (R Core Team 2021).

Results

The mean distance of bison to the feeding sites during daytime hours (between 08:00 am and 04:00 pm) (± SD) ranged from 0.03 ± 0.34 to 2.33 ± 0.34 km for females and from 0.18 ± 0.45 to 2.17 ± 0.45 km for males (Fig. 1S, Appendix 1). For both sexes, mean distance to feeding sites varied throughout the season (GAM1; Fig. 2; Table 1, females: P < 0.001; males: P < 0.05) in a non-linear pattern (GAM1; Table 1). Females were the closest to the feeding sites in the middle of winter, while at the beginning and the end of winter they roamed further from the feeding sites (Fig. 2A). For example, on the 76th day of winter (February 15), females were on average 90 m closer to the feeding sites than on the 15th day of winter (December 15) and 210 m closer than on the 105th day of winter (March 15). Males were the farthest from the feeding sites at the beginning of winter and approached feeding sites in the middle of winter (Fig. 2B). For example, on the 15th day of winter (December 15), males were on average 290 m farther from the feeding sites than on the 76th day of winter (February 15) and 360 m farther away than on the 105th day of winter (March 15). Furthermore, independently of the day of winter, distance of male and female bison from the feeding sites increased with increasing corrected mean daily temperature (GAM1; Table 1; Fig. 3; P = 0.04). With corrected mean daily temperature decreasing by 5ºC, the distance to the feeding sites of females decreased on average by 100 m, while for males, the distance to the feeding sites decreased by 140 m. Bison distances to feeding sites were negatively related to snow presence (GAM1, Fig. 3, Table 1, P < 0.001). Bison were by 120 m closer to the feeding sites when snow was present than in snowless conditions (Fig. 3).

Fig. 2
figure 2

The relationship between the day of winter (1 December–31 March), snow presence and the mean distances of female (A) and male (B) European bison (Bison bonasus) to feeding sites in Białowieża Primeval Forest in winter (December–March) in 2005–2012

Full size image
Table 1 Results of the general additive models testing the association between (1) the mean distance of European bison to winter feeding sites (GAM1), (2) 95% quantile of distance of European bison to winter feeding sites (GAM2) in Białowieża Primeval Forest between 2005 and 2012 and the set of explanatory variables: residuals of mean daily temperature, day of winter, snow presence and sex
Full size table
Fig. 3
figure 3

The relationship between the residuals of mean daily temperature, snow presence and the mean distances of female (A) and male (B) European bison (Bison bonasus) to the feeding sites in Białowieża Primeval Forest in winter (December–March) in 2005–2012

Full size image

The predictions of GAM2 indicated that 95% quantile of bison distance to feeding sites and thus, the area of Białowieża Primeval Forest occupied by bison, differed with weather conditions. The area used by European bison decreased on days when residual mean daily temperature was low and snow present (i.e. cold conditions) (Table 2). In cold conditions, males utilized larger area than females only at the beginning of winter (Table 2), while in mild weather conditions (snow absent and the highest residual mean daily temperature for this month), males utilized larger area than females throughout the whole winter (Table 2). In January, for mild weather conditions, female and male bison were predicted to utilize, respectively, more than 4 and 28 times larger area than under cold conditions (snow present and the lowest residual mean daily temperature) (Table 2; Fig. 4).

Table 2 The predicted 95% quantile distances (± SE) of European bison (Bison bonasus) to the feeding sites in Białowieża Primeval Forest and maximum area utilized by European bison under the warm and cold weather scenarios in winter
Full size table
Fig. 4
figure 4

The distribution of feeding sites in Białowieża Primeval Forest with marked buffers representing the maximum area utilized by male and female European bison (Bison bonasus) under the (A) cold and (B) warm weather conditions in a given month. For more details, see the “Statistical analyses” section

Full size image

Mean distance of bison to the feeding sites was significantly negatively associated with snow depth (GAM3, Table 1, P < 0.001). Males responded to increasing snow depth more strongly than females (GAM3, Fig. 5, Table 1, P < 0.05). With snow depth increasing by 5 cm, the distance to feeding sites of males decreased on average by 65 m while for females by 20 m.

Fig. 5
figure 5

The relationship between snow depth and mean distances of male and female European bison (Bison bonasus) to the feeding sites in Białowieża Primeval Forest in winter (December–March) in 2005–2012

Full size image

Discussion

In our work, we examined the space use of European bison around winter feeding sites in Białowieża Primeval Forest under contrasting weather conditions. According to our analyses, bison were the closest to feeding sites in the middle of winter (January and February) when weather conditions were most severe and natural food resources remarkably limited. Females roamed farther from the feeding sites, mainly at the end of winter (March) when some herbaceous vegetation starts to appear in open habitats. Males approached feeding sites later (late January/early February) and remained in the vicinity of feeding sites throughout March. For females, such behaviour contradicts the previous findings which showed that ungulates kept utilization of the feeding sites even when natural vegetation is already present in spring (e.g. red deer (Bradesco 2012)). Moreover, we found that temperature affected bison space use independently of the time of season, with bison being closer to the feeding sites on days when daily temperature was below the expected for a given day of winter. Additionally, bison were also closer to the feeding sites on days when snow cover was present, especially when snow was deep. Our results showed that utilization of feeding sites by males was more related to harsh winter conditions than in females.

The severity of weather conditions affects the space use and thus, home range size of many ungulate species (Grignolio et al. 2004; Morellet et al. 2013; Borowik et al. 2020). Animals exposed to low temperature in winter reduce their activity to limit loss of energy needed for thermoregulation (white-tailed deer Odocoileus virginianus: Taillon et al. 2006; wild boar: Brivio et al. 2017). In addition, the presence of snow cover impedes animal mobility and makes movements costly in terms of energetic expenditures (Fox et al. 1992). These costs increase with increasing snow depth, as observed, for example, in mule deer (Odocoileus hemionus), elk (Cervus elaphus nelson; Parker et al. 1984), barren ground caribou (Rangifer tarandus granti, Fancy and White 1987) and roe deer (Hewison et al. 1998). Snow accumulation during winter may also limit access to vegetation, thereby increasing the amount of time required to encounter a food item (Robinson and Merrill 2012). To counteract the costs associated with severe weather conditions, ungulates utilize supplementary fodder when available, and concentrate their activity around feeding sites, which is reflected in reduced mobility and home range sizes (Guillet et al. 1996; van Beest et al. 2010a). This is especially true for European bison, which are grazers, unlike other ungulate species occurring in the same ecosystem, and their natural forage in winter consists in majority of dry grasses that are often unavailable due to thick snow cover (Krasińska and Krasiński 2013).

We also indicated the sex-specific differences in the effect of weather condition and day of winter on the distance of bison to feeding sites. Females were closer to feeding sites for the most of winter, which may suggest that females utilized feeding sites more often than males, while males approached feedings sites mainly during adverse weather conditions. These differences can be caused by sex-segregation observed in bison during this time of the year, which is associated with a high dimorphism in body size of European bison (Krasińska and Krasiński 2002). According to activity budget hypothesis, larger males in dimorphic ungulates should spend less time foraging and spent more time walking than females (Ruckstuhl and Neuhaus 2002). Therefore, European bison males in winter are usually solitary or move in small male groups while females with calves create large aggregations (Krasińska and Krasiński 1995). Moreover, males can find sufficient food supplies also outside the feeding sites (e.g. natural vegetation in forest glades or river valleys), while female herds, due to high forage requirements, are forced to utilize feeding sites or migrate outside the forest to feed on winter crops (Hofman-Kamińska and Kowalczyk 2012; Kowalczyk et al. 2013).

Although we did not use direct measure of the utilization of feeding sites by bison, and the efficiency of applied index (i.e. distance of bison to feeding sites) in describing supplementary forage use by bison remained unknown, we claim that the different distances of bison to the feeding sites in winter are expressing the variation in feeding site utilization. We did not know any other factor associated with studied feeding sites, or surrounding area, that could additionally affect differences in bison proximity to the feeding sites. However, it would be beneficial to confirm the observed relationships by the results of the studies made on direct measures of supplementary fodder use, as well as supplementing those analyses with the variables describing the quantity and quality of consumed supplementary forage and spatio-temporal variation in natural food supply.

In the presence of fast-occurring climate change and global warming, we expect a further decrease in winter severity; thus, its limiting effect on the mobility of ungulates will no longer occur at such scale and intensity as in the past. Increasing winter temperature and shorter persistence of snow cover will cause earlier vegetation green-up and ultimately, significant changes in the time of natural forage availability to herbivores (Barichivich et al. 2013; Keenan et al. 2014). Milder winters may also reduce changes in body condition and lower energetic requirements of animals, thus increasing independence from fodder supplementation. In the case of European bison, we can expect a limited intake of supplementary fodder and a larger area of the BPF utilized by bison during milder weather conditions. Such changes can have a diverse impact on the ecosystem functioning through their influence on nutrient cycling (Melis et al. 2007; Jaroszewicz et al. 2013), vegetation growth, forest succession (Kowalczyk et al. 2021) and the availability and distribution of carrion (Selva et al. 2003). Furthermore, increased bison mobility seems to be also favourable because smaller bison aggregations may decrease parasite and disease transmission (Kołodziej-Sobocińska 2016a, b) and diffuse browsing from feeding sites areas (Jaroszewicz et al. 2017). On the other hand, greater mobility and utilization of larger areas by bison during milder winter may also have negative impacts. For instance, bison may disperse non-forest plant species consumed with hay that is delivered to feeding sites which in consequence, may lead to changes in the composition of natural vegetation and plant diversity (Jaroszewicz et al. 2008, 2013). Greater mobility also increases probability of bison moving more frequently outside of BPF into farmland areas surrounding the forest, causing conflicts with farmers through damaging crops.

To summarize, our results showed that the utilization of feeding sites by bison was associated with winter weather conditions. Therefore, there is a high probability that ongoing global warming will cause further changes in supplementary forage use by European bison, which may have far-reaching ecological consequences for the ecosystem of the BPF. However, to address appropriately the problem of supplementary feeding under changing weather conditions, we urgently need further studies which will more precisely evaluate optimal supplementary feeding strategies under different winter severity scenarios, allowing to meet management or conservation goals while optimizing the costs. This is especially important in case of European bison which was restored after extinction in the wild (Plumb et al. 2020), as its future depends on appropriate conservation practices aimed on both, preservation and counteracting the actual and potential threats, and ensuring that population functioning is shaped by natural processes as much as possible.