Space–time interactions between free-ranging red deer (Cervus elaphus) and grazing livestock on a fenced pasture

Livestock farming, which produces milk and meat, is practised by nearly all cultures. But in addition to this, grazing livestock fulfils other fundamentally important roles, one of which is the maintenance of open landscapes, which in turn provides habitat for many rare plant and animal species. Whether wild large herbivores, such as red deer (Cervus elaphus), benefit from the management of the open landscapes is unknown, and there are conflicting views on the nature of the interaction between livestock and wildlife. To investigate this question, we studied interactions in space and time between red deer and livestock in an 787 ha fenced pasture in eastern Germany. First, we used faecal pellet counts to determine whether red deer were using the fenced pasture and whether livestock were using the entire available pasture. Second, we fitted ten red deer, as well as eight Konik horses (Equus ferus caballus) and eight Heck cattle (Bos taurus) with GPS-collars. These data allowed us to perform analyses of space–time interactions. The faecal pellet data showed that red deer did indeed use the fenced pasture and that livestock used the entire extent of the pasture. Further, we verified a neutral space–time interaction between red deer and livestock within the pasture. Thus, there was no indication of a negative effect of the presence of livestock on red deer inside the pasture. For these reasons, we recommend that low-intensity grazing by cattle and horses can maintain open landscapes without detrimental effects on the use of the area by red deer.


Introduction
Livestock grazing is a major use of about 30% of the ice-free terrestrial surface of the planet (Steinfeld et al. 2006), and so it can be an important influence on ecosystem function (Caro et al. 2014). Livestock are mainly used for milk and meat production, but grazing by large mammals can also be used as a tool in landscape management and thus for nature conservation. In Europe, open-and semi-open landscapes with dry soils, including grass-and heathlands, are critical habitats for highly adapted and often rare plants and animals. The use of grazing animals is a feasible and widely used method for maintaining open landscapes (van Wieren 1995). Provided that livestock is kept to moderate intensities, grazing can increase plant species diversity (Bakker et al. 1983;Dostalek and Frantik 2008;Henning et al. 2017;Benthien et al. 2018) as well as structural diversity (Bakker et al. 1983;Henning et al. 2017). Furthermore, grazing is effective in preventing shrub and tree development (Dostalek and Frantik 2008;Jauregui et al. 2009) and can have positive effects on vegetation and associated animal species, such as bird species dependent on semi-open landscapes (Holechek et al. 2006;Köhler et al. 2016). Although moderate grazing has become a widely used tool for restoring open landscapes (Newton et al. 2009), the use of domesticated species like cattle, horses, sheep and goats remains controversial. Keeping of livestock always requires appropriate fencing or shepherding. Depending on style, 1 3 61 Page 2 of 9 fences can be barriers or even trap and kill wildlife (Pocorny et al. 2017). They can also lead to habitat fragmentation (Bolger et al. 2008). In addition, the presence of livestock may have direct effects on wildlife, e.g. via competition for food. Not surprisingly, there is no simple consensus and evidence for competition between livestock and wildlife varies. McMahan (1964), Bowyer and Bleich (1984), Austin and Urness (1986), Ragotzkie andBailey (1991), Stewart et al. (2002) and Wagoner et al. (2013) found negative impacts of grazing livestock on food availability for deer, whereas Neal (1981), Urness (1981), Reiner and Urness (1982), Gordon (1988), Willms et al. (1979) and Yeo et al. (1993) observed positive effects like, e.g. greater biomass and less mature forage. Furthermore, both livestock and deer can use the same space but at different times (Cooper et al. 2008;Lindgren and Sullivan 2014). Thus, livestock and native deer can have a negative space-time interaction (Cohen et al. 1989;Loft et al. 1993;Coe et al. 2001). The aim of the management is to create semi-open landscapes of high values for biodiversity by using livestock as a management tool but also acknowledge habitat preferences of wildlife (e.g. red deer). In order to effectively manage semi-open landscapes, we need to understand the space-time interaction between red deer and the horses and cattle that graze in the same landscape.
Therefore, we specifically asked the following questions: 1. Do red deer enter the fenced pasture when livestock are present and is the pasture therefore a part of their habitat? 2. Do livestock and red deer use the whole area of the fenced pasture? 3. Is there overlap in space and time by livestock and red deer? 4. Do interactions differ between red deer and cattle or red deer and horses?
To answer these questions, we used faecal pellet counts (question 1 and 2) as well as GPS-data from collared red deer and two livestock species (horses and cattle; question 3 and 4).

Study area
We conducted the study in the 2116 ha area called "Oranienbaumer Heide" (N 51.774603, E 12.364772, 70 m a.s.l.) located in Saxony-Anhalt, Germany. The mean annual precipitation is about 560 mm, including rare snowfall, and a mean annual temperature of 9.2 °C. Between 1945 and 1989, the Oranienbaumer Heide was used as a military training area (John et al. 2010), which has resulted in a semi-open landscape with a mosaic of heath-and grasslands as well as various successional stages of forests. The cessation of military training in 1989 led to the rapid encroachment of woody species, such as silver birch (Betula pendula), Scots pine (Pinus sylvestris) and aspen (Populus tremula). Common ground cover species included common heather (Calluna vulgaris), bushgrass (Calamagrostis epigejos) and common broom (Cytisus scoparius) (Jentzsch and Reichhoff 2013;Henning et al. 2017).
In addition to red deer, roe deer (Capreolus capreolus) and wild boar (Sus scrofa) occur on the study site. Since 2013, a single female wolf (Canis lupus) has been regularly detected in the study area. From 2015 onwards, a wolf couple was confirmed, and the first offspring was born in 2017 (Landesamt für Umweltschutz Sachsen Anhalt and Wolfskompetenzzentrum Iden 2017).
The Oranienbaumer Heide is surrounded by forests and extensive farming areas. Hunting is permitted several times each year. In 2008, an 787 ha pasture was enclosed by a 1-m electric fence, and Heck cattle and Konik horses (stocking rate approximately 0.2 livestock units/ha) were introduced (e.g. Henning et al. 2017) in an attempt to reduce the encroachment of woody species. During the time of the study, about 50 Heck cattle and 40 Konik horses were kept in the Oranienbaumer Heide. Human disturbances result from daily inspections of the fence and from activities on several public pathways (cyclists and hikers) through the pasture and the surrounding area (Gillich et al. 2021).

Faecal pellet count
We studied the use of the fenced pasture by using faecal pellet counts. This method is widely accepted for assessing habitat use (Archibald et al. 2005;Schwarz et al. 2017;Bobrowski et al. 2020), especially if the number of sympatric ungulates species is low (Spitzer et al. 2019). Only red and roe deer are known to inhabit the study area in detectable numbers, and their faeces are better distinguishable in winter and spring (Spitzer et al. 2019). We counted faecal pellet groups of red deer, cattle and horses in May 2016 and May 2017, to identify the use of the fenced pasture (787 ha study site) and the surrounding area (1329 ha control site) by red deer and also to investigate whether livestock used the entire fenced pasture. Regarding red deer, we recorded only pellet groups consisting of more than seven individual pellets (Schwarz et al. 2017) to avoid overestimations, because smaller pellet groups might indicate moving animals. We used a systematic design for data collection (see also Bobrowski et al. 2015). We randomly selected the position of a sampling grid, where sampling plots were regularly distributed 250 m apart. We investigated 343 sampling plots (each sampling plot measuring 2 × 50 m), covering a total area of 34,300 m 2 . Of these plots, 134 were located within the fenced pasture, whereas 209 plots were located outside the fenced pasture as a control. The number was adjusted to the size of the fenced pasture and the control area. We examined in both years exactly the same plots for red deer, horses and cattle faeces.

GPS-telemetry
All animal capture and marking activities in this study were performed by trained professionals. Permissions to conduct animal research and to use anaesthetics were approved by the state administration of Saxony-Anhalt (203.6.1-42,502-9-016 HNEE/408.2.1-65,020/1). Special permission to catch and mark wildlife was granted by the state agency for environmental conservation of Saxony-Anhalt (43.15-22480-02/2014).
Between February 2015 and March 2016, we caught six female (two as subadults and four as adults) and four male (all of them as subadults) red deer (n = 10) by using three 500 m 2 , manually triggered drop-net capture systems, baited with sugar beets and placed outside the fenced pasture but in close vicinity (max. 200 m) to the fence. Immediately after capture, the animals were immobilised with xylazine (Rompun® TS, Bayer AG, Leverkusen, Germany). Whilst under anaesthesia, we equipped the animals with ear-tags (TypiFix™, Schlieren, Switzerland and Dalton GmbH, Newark Nottinghamshire, UK) and GPS-GSM-collars (Pro Light and GPS PLUS both from VECTRONIC Aerospace GmbH, Berlin, Germany). In collaboration with the owners of the livestock, we fitted eight female Heck cattle (n = 8) and eight female Konik horses (n = 8) with GPS-collars (Pro Light from VECTRONIC Aerospace GmbH, Berlin, Germany).
Every 2 h, the collars determined a GPS-position and transmitted it via GSM (Global System for Mobile) to our server over a period of 27 months (28 January 2016 to 23 April 2018). Four of the red deer died or the collar stopped working during the study. For the data analysis, we only used the GPS-positions determined from four or more satellites to guarantee a high precision. Final analysis was performed on 72,731 GPS-positions from red deer, 29,749 GPS-positions from horses and 37,258 GPSpositions from cattle.

Data analysis and statistics
To test for differences in the number of found faecal pellets of red deer faecal pellets and GPS-positions between the fenced and non-fenced area, we used Mann-Whitney U test, and to compare the interactions between Konik horses and Heck cattle to red deer, we used paired sample t-test, both in SPSS Version 17 (IBM, Chicago, USA). Furthermore, to investigate the interaction behaviour by red deer and livestock, we used the Jacob's Index (Dunn 1979;Kenward 2001) function in the software RANGES 9 Version 1.8 (Kenward et al. 2014) based on work by Kaulhala and Holmala (2006), Michler (2016) and Gillich et al. (2021). For this analysis, we used the data of only eight red deer, which were present within the fenced pasture.
We used simultaneous locations (recorded within a time buffer of 3 min) for each dyad of animals to characterise their temporal association (Minta 1992). Note, that not all red deer were entering the fenced pasture. There was a subset of animals that we could use for further analysis, because of the restricted time buffer. The spatial and temporal association between animals was calculated using the average distance of their simultaneous localisations (D O , observed distance) and the average distance of all recorded localisations of one animal (within the observation period), to all localisations of a second animal (D R , randomised distance). If there were n pairs of locations from animal 1 at the moment i (x 1i , y 1i ) and animal 2 at the moment i (x 2i , y 2i ) for each dyad, the observed mean distance between them was (n = number of simultaneous localisations of animal 1 and animal 2), which can be compared with the expected mean distance from animal 1 at the moment (x 1i , y 1i ) to all other localisations of animal 2 at the moment j (x 2j , y 2j ), that is obtained by randomising all possible pairs of locations at which the animals were detected (Kenward 2001). The Jacob's index (J X ) is obtained by equation: Jacob's index ranges between − 1 and + 1, where − 1 is a measure of maximum avoidance and + 1 is a measure of maximum attraction. Paired sample t-tests were conducted to determine differences in space-time interaction between red deer and horses as well as between red deer and cattle.

Faecal pellet counts
In 2016, we recorded 496 red deer faecal pellet groups, of which 148 were inside the fenced pasture (29.8%), mainly in the northeast, central and southern area (Fig. 1A, left). This distribution of faecal pellets differed significantly between the fenced pasture and the area outside the fence (n = 343; Z = − 2.175; p = 0.030). In 2017, we recorded a total of 196 groups of pellets from red deer (Fig. 1A, right). Within the fenced pasture, we found 65 pellet groups, (33.5%), mainly in southern and central area. The number of found faecal pellets inside and outside the fence in 2017 was not significantly different (n = 343; Z = − 0.667; p = 0.505). Despite a decrease of almost 40% of red deer faecal pellet groups from 2016 to 2017, the percentage distribution of red deer faecal pellet groups inside the fenced pasture increased by 3.7%. These results reflected a similar use of the fenced pasture by red deer in both years. In the fenced pasture, we documented 290 faecal patties from cattle on 85 plots in 2016 (Fig. 1B, left) and 327 faecal patties on 96 plots in 2017 (Fig. 1B, right).
In 2016, we found 927 faecal pellet groups from horses on 125 plots in the fenced pasture (Fig. 1C, left), whereas we found 609 faecal pellet groups on 122 plots in 2017 (Fig. 1C, right). The widely documented faeces from horses and cattle, with slight majorities in the north (Fig. 1), indicated an almost complete use of the fenced pasture by livestock.

GPS-telemetry
We analysed 72,731 GPS-positions from ten collared red deer of both sexes (Fig. 2). During the entire evaluation period, only two of collared red deer remained outside the fenced study area (Table 1). The centre of their home range was located south of the fenced pasture. Therefore, we did not include the GPS-data from these two red deer in further analyses. From the other eight collared red deer, we obtained 8,874 GPS-positions from inside the fenced pasture, mainly in the northern half, which corresponds 12.2% of the total observations. Thus, we detected a significant difference between the fenced pasture and the non-fenced area regarding the distribution of red deer (n = 8; Z = − 5.224; p ≤ 0.001). The eight red deer spent, on average, 14.5% of their time within the fenced pasture.

Space-time interaction between red deer and livestock
In order to analyse the space-time interaction using Jacob's index, we used the simultaneously recorded pairings between 8 red deer and 8 horses (Fig. 3), as well as between 8 red deer and 8 cattle (Fig. 4). Note, that not all the animals paired within the determined time buffer. We found n = 34 possible pairings between red deer and horses and n = 53 possible pairings between red deer and cattle to analyse using Jacob's index. Both the mean and median of Jacob's index between red deer, horses and cattle were 0.00. The Jacob's index for red deer and horses was between − 0.16 and 0.18 and for red deer and cattle between − 0.16 and 0.17.
The results of the paired sample t-test, based on the means of the Jacob's index values, showed no significant differences between the space-time interaction of red deer with cattle and red deer with horses, respectively (n = 87; T = −0.957; p = 0.370).

Discussion
A widely used method for maintaining species-rich, semiopen landscapes is year-round grazing by large herbivores (Henning et al. 2017). In addition to lower costs and risks, grazing is often preferred over cutting or mowing, because extensive grazing can result in patterns of vegetation at different scales (Bakker and de Jonge 1998), positively influence vegetation composition (Köhler et al. 2016;Benthien et al. 2018) and create mosaics of eutrophic/mesotrophic sites (Bakker and Van Diggelen 2006). But year-round grazing by large livestock and the associated fencing may influence movements of local wildlife.
In order to investigate the space use on the fenced pasture by livestock and red deer, we conducted two faecal pellet counts. Our faecal pellet data indicated a general use of the fenced pasture by free-ranging red deer and an entire use by grazing livestock with a slightly stronger tendency to the northern area. This holistic use of the pasture by livestock at our study site was also found by Kochs (2016). Despite the fact that we documented almost 40% fewer red deer faecal pellets in the second year, we still observed the same ratio of species' faecal pellets on the fenced pasture. The difference in the total number of faecal pellets found (from red deer and livestock) could be related to differences in weather and the associated activity of dung beetles (Groth et al. 2011).
The space utilisation of the collared red deer showed that individual red deer used the fenced pasture differently. The faecal pellet counts as well as the telemetry data showed that the fence, used in our study area, was not an absolute barrier for the red deer.
There are several studies focused on shared space and overlap in space use between livestock (mainly cattle) and deer, (white-tailed deer (Odocoileus virginianus), mule deer (Odocoileus hemionus) and elk (Cervus canadensis)), mostly from North America (Schieltz and Rubenstein 2016), but little is known about space-time interaction between free-ranging red deer and cattle and/or horses. Most research has observed a high overlap of habitat use between deer and cattle (Cohen et al. 1989;Yeo et al. 1993;Cooper et al. 2008), which is in accordance with our findings. However, Stewart et al. (2002) found that elk and mule deer avoided areas grazed by cattle.
In our space-time interaction analyses, we found no indication of avoidance or attraction behaviour of red deer to The whole study area (2116 ha, white: control site 1329 ha) is black framed; the fenced pasture (787 ha) is a grey polygon. The number of faecal pellet groups is numbered in the circles the presence of Heck cattle and Konik horses on the fenced pasture. Eight of ten collared red deer used the fenced area and were in contact with the grazing livestock. There was also no difference between cattle or horses in their effects on the red deer. This neutral interaction was also observed by Michael (1967) for white-tailed deer to cattle and horses. But equally, there are other studies which contradict our results and demonstrate a strong temporal separation between elk, mule deer, whited-tailed deer and cattle (Loft et al. 1993;Coe et al. 2001;Cooper et al. 2008). Thus, the individual deer species may react differently to the presence of large grazing livestock. Additionally, the habitat type may have an influence on the reactions of deer towards livestock, since visual cover can decrease vigilance and reactivity of red deer (Whittington and Chamove 1995). Furthermore, the livestock species, the stocking rate of livestock and the habituation  time of the deer to the presence of livestock could influence their behaviour. Weiss et al. (2022) found that red deer avoid areas with temporary grazing by sheep and goats. Finally, the presence of people and wolves in the area might have a general influence on the movement behaviour of red deer and grazers. Therefore, further studies are needed to evaluate more different potential influencing factors on interaction behaviour between livestock and local wildlife. Fig. 3 Temporal interaction between red deer and Konik horses in the study area Oranienbaumer Heide from 28 January 2016 to 23 April 2018. Jacob's index ranges from + 1, if animals used the same area at the same time (attraction), to − 1, if they use the same area only at different times (avoidance). The value zero represents a neutral behaviour between the animal species. Boxplot middle bar = median. Box = first and third quartiles Fig. 4 Temporal interaction between red deer and heck cattle in the study area Oranienbaumer Heide from 28 January 2016 to 23 April 2018. Jacob's index ranges from + 1, if animals used the same area at the same time (attraction), to − 1, if they use the same area only at different times (avoidance). The value zero represents a neutral behaviour between the animal species. Boxplot middle bar = median. Box = first and third quartiles In summary, we did not find any indication of avoidance or attraction behaviour of free-ranging red deer to the presence of low-intensity grazing Heck cattle and Konik horses within a large, fenced pasture. Consequently, we can support a low intensity to moderate use of these grazer species with the fence used here as a possible management tool to maintain a semi-open landscape.