How do agricultural practices affect the movement behaviour of European brown hares (Lepus europaeus)?

Agriculture, Ecosystems and Environment 292 (2020) 106819

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment journal homepage: www.elsevier.com/locate/agee

How do agricultural practices affect the movement behaviour of European brown hares (Lepus europaeus)?

T

W. Ullmanna,b,*, C. Fischerc, S. Kramer-Schadtd,e, K. Pirhofer-Walzlb,f, M. Glemnitzb, N. Blauma a

Department of Plant Ecology and Nature Conservation, University of Potsdam, Am Mühlenberg 3, 14476 Potsdam, Germany Leibniz-Centre for Agricultural Landscape Research (ZALF), Eberswalderstr. 84, 15374 Müncheberg, Germany c Restoration Ecology, Department of Ecology and Ecosystem Management, Technische Universität München, Germany d Department of Ecological Dynamics, Leibniz-Institute for Zoo and Wildlife Research (IZW), Alfred-Kowalke-Str. 17, 10315 Berlin, Germany e Department of Ecology, Technische Universität Berlin, Rothenburgstrasse 12, 12165 Berlin, Germany f Institute of Biology, Freie Universität Berlin, Altensteinstraße 6, 14195, Berlin, Germany b

ARTICLE INFO

ABSTRACT

Keywords: Resource change Range shift Energy expenditure Telemetry GPS Landscape complexity Germany Radio tracking

Agricultural landscapes are spatially and temporally dynamic habitats that force wildlife to interact with different management practices, such as harvests and mowing events which cause sudden changes in resource availability. Animals may avoid agricultural management events and the changed habitat, to search for undisturbed areas or they might use and explore such areas due to beneficial changes in vegetation structure. Further, landscape structure might influence the movement processes that are underlying reactions to agricultural management. Here we study how agricultural management events affect animal movement behaviour in two contrastingly structured agricultural landscapes. In 2014 and 2015 we collared 36 European brown hares (Lepus europaeus) with GPS-tags and accelerometers in Northeast Germany (large agricultural fields) and South Germany (small fields and comparatively more nonarable vegetation). We recorded hares’ movement behaviour for 4 days before and after agricultural management events without (e.g. fertilizer application) and with (harvest and mowing) immediate changes in resource availability, on the most common production cover types (wheat, grasslands, maize and rapeseed). We measured the number of GPS points in the focal fields, the range size (area covered within 4 days), the shift in range centre, and the hares’ energy expenditure (overall dynamic body acceleration). More GPS locations were found on fields following management events that affected resource availability, and less GPS fixes were recorded on the wheat fields after management events without resource changes. Compared to an equivalent period without management events, hares showed increased range shifts after harvesting maize and rapeseed fields, mowing grasslands and after management events without resource changes for most of the production cover types. Range sizes were only affected in wheat fields in Northeast Germany, where they increased after harvest and decreased after management events without resource changes. Energy expenditure was unaffected by agricultural management. Hares may profit from harvested fields, likely because they find food in form of fallen grains, improve their predator detection probability and generally prefer areas with low vegetation. The reaction to management events without the change of resources might depend on the specific type of management practices (e.g. organic vs. inorganic fertilizer). Landscape structure may play an important role as range sizes increase due to the necessity to reach distant alternative habitats. Hence, the provision of smaller fields with high crop diversity and sufficient alternative habitat patches throughout the year has the potential to maximise accessible resources and predator detection ability for hares and other farmland wildlife.

1. Introduction Agricultural landscapes cover roughly 38 % of the Earth’s terrestrial surface and 42 % of Europe (European Environment Agency, 2018a), of

⁎

which almost 60 % is intensively managed (Eurostat 2019). Agricultural intensification confronts farmland wildlife with a variety of challenges, such as longer travelling distances (Saïd and Servanty, 2005; Ullmann et al., 2018), frequent disturbances by agricultural

Corresponding author at: Department of Plant Ecology and Nature Conservation, University of Potsdam, Am Mühlenberg 3, 14476 Potsdam, Germany. E-mail address: [email protected] (W. Ullmann).

https://doi.org/10.1016/j.agee.2020.106819 Received 28 January 2019; Received in revised form 7 January 2020; Accepted 9 January 2020 0167-8809/ © 2020 Elsevier B.V. All rights reserved.

Agriculture, Ecosystems and Environment 292 (2020) 106819

W. Ullmann, et al.

machinery (Báldi and Faragó, 2007; Padié et al., 2015), and abrupt spatiotemporal changes due to the synchronous harvest of multiple crops fields (Schai-Braun et al., 2014; Vasseur et al., 2013). However, we lack a comprehensive understanding of how animals may respond or adapt to these often interlinked challenges in human-dominated dynamic systems. The dynamics of agricultural landscapes, including frequent disturbances and abrupt spatiotemporal resource changes, are produced by the local agricultural management regime. In particular, animals can experience: 1) management events with resource changes, such as harvesting of crop fields and mowing of grasslands, and 2) management events without the change of resources, e.g. the application of fertilizer and plant protection products. Both of these management types disturb farmland species, if only through the presence of heavy machinery in their habitat. Movement is one key process through which animals can adjust to these disturbances. However, the effects of management events on animal movement behaviour can be complex. Harvesting and mowing for example cause multi-layer habitat changes, important to wildlife movements, because they simultaneously remove cover and physical barriers (Rühe, 1999; Vercauteren and Hygnstrom, 1998). During the maturation phase of crops and grasses the fields are difficult to pass through, as it involves higher energy expenditure to move through high standing vegetation (Rühe, 1999). After the harvest of crops and the mowing of grasslands farmland wildlife can easily cross the newly accessible space. This could provide direct connections to important (semi-)natural habitat patches that could otherwise be reached only by circumventing the field with the high standing vegetation. Management events with resource changes also eliminate cover and forage by suddenly removing almost the entire standing biomass. However, they can concurrently provide forage in the form of fallen grain and space for freshly sprouting plants (Cimino and Lovari, 2003; Späth, 1989). Roe deer (Capreolus capreolus), for example, shift their home ranges away from maize fields after harvesting, supposedly to find new foraging grounds and/or cover possibilities (Cimino and Lovari, 2003; Vercauteren and Hygnstrom, 1998). In contrast, while European brown hares (Lepus europaeus) may display no shift in their home range, they may increase their home range size after cereal harvest to incorporate alternative foraging habitats (Schai-Braun et al., 2014). Agricultural management events with a change in resources (i.e. a massive removal of standing biomass), seem to shape animal movements, however there are few studies comparing these with direct effects of agricultural management events without resource changes. Roe deer were also observed to flee from agricultural machinery (Cimino and Lovari, 2003; Mrlik, 1990; Padié et al., 2015; Stankowich, 2008). These studies show that animals change their movement behaviour in response to disturbances linked to agricultural practices. How exactly animals react to agricultural management events can vary considerably, potentially applying a variety of movement processes (Drygala and Zoller, 2013; Schai-Braun et al., 2014; Vercauteren and Hygnstrom, 1998). Increasing the range size, for example, would benefit the animal by incorporating alternative habitat. Animals can also shift the range centre to incorporate either the recently modified fields or alternative habitat, or to avoid agricultural machinery and fields that have recently been managed. These possible movement reactions may be accompanied by changes in energy expenditure, which has often been stated (Harestad and Bunnel, 1979; Rühe and Hohmann, 2004; Schai-Braun et al., 2014) but has historically been difficult to measure (Brown et al., 2013). Another level of complexity in how animals adjust their movement behaviour in response to different management events is added by the different production cover types (crops and grasslands). For example, previous studies only considered cereal fields because they were the main crop type in the respective study areas (Marboutin and Aebischer, 1996; Schai-Braun et al., 2014). However, with agricultural intensification maize and rapeseed have become more numerous over the last decades and by now cover substantial parts of agricultural

landscapes (MEL Bundesministerium für Ernährung und Landwirtschaft, 2014; Sauerbrei et al., 2014). Further, the consolidation of fields leads to low compositional and configurational landscape heterogeneity (Fahrig et al., 2011). Hence, the landscape consists mainly of extensive, homogeneous areas of agricultural matrix with a reduced number of and large distances between (semi-) natural landscape elements such as hedgerows, tree stands and other non-crop vegetation (Batáry et al., 2017). In these landscapes, animals must travel longer distances between the habitats they use (e.g. resting sites and foraging patches), often resulting in increased home range sizes (SchaiBraun and Hackländer, 2014; Ullmann et al., 2018). This increase in movement can lead to higher energy expenditure (Daan et al., 1996; Harestad and Bunnel, 1979; Petrovan et al., 2017) and reduce the animals’ reproductive output (Doherty and Driscoll, 2018). In this study, we investigated the adjustments in movement behaviour and energy expenditure of European brown hares to agricultural management events with and without resource changes and while accounting for differences in landscape configuration and a variety of production cover types (wheat, grassland, maize and rapeseed). We collared hares with GPS tags with internal accelerometers in Northeast Germany (here after NE Germany), representative of a simple landscape and in South Germany (here after S Germany), representing a complex landscape with higher heterogeneity of production cover types and (semi-)natural landscape elements. We hypothesize that hares react differently to the two types of agricultural management events. After events without resource changes we expect hares to decrease their ranges, increase their energy expenditure and shift their range centres away from the focal field to avoid the agricultural machinery and/or the measures applied on it (e.g. fertilizer or plant protection products). In contrast, after management events with resource changes, we expect hares to incorporate the harvested and mown fields into their ranges as they experience a release from the spatial restriction of high standing vegetation. This would allow them to reach distant habitat patches by crossing the harvested/mown fields, while simultaneously exploring the recently modified fields for alternative foraging possibilities and improved predator detection. We predict an increase in range size and energy expenditure, and that hares might shift their entire range towards harvested and mown fields. We furthermore hypothesize that hare movement reactions are stronger in NE Germany compared to S Germany, as fewer alternative habitats are available and lie further apart. 2. Methods 2.1. Study area We selected two study areas which differed in landscape configuration of arable fields and non-arable vegetation (Fig. 1). The study area representing the simple landscape, NE Germany, was located in Brandenburg, around 100 km north of Berlin (centred at 53° 35′ N; 13° 68′ E). This area was part of the long-term research platform AgroScapeLab Quillow (Agricultural Landscape Laboratory Quillow) of the Leibniz Centre for Agricultural Landscape Research (ZALF) and the Biomove research training group (www.biomove.org/about-biomove/ study_area/). The 213 km² area is characterized by intensive agriculture with an average field size of 27.5 ± 1.1 ha and an average field edge length of 20.3 ± 0.7 km per 1 km² (mean ± SE; calculated based on maps provided by the Landesvermessung und Geobasisinformation Brandenburg (InVeKoS, 2014)). The agricultural landscape shows a lower edge density than the S German study site. Only few (semi-) natural landscape elements like hedgerows, trees and fallow land (∼20 % of the total area available, European Environment Agency, 2018b) remained after field consolidation practices in East Germany in the 1950s (Batáry et al., 2017). The NE Germany study area is covered up to 73 % by arable land (European Environment Agency, 2018b) which 2

Agriculture, Ecosystems and Environment 292 (2020) 106819

W. Ullmann, et al.

Fig. 1. Representative extracts of the two study areas in A) NE Germany and B) S Germany (GADM http://gadm.org/) and their location in C) Europe and D) Germany. The satellite images (Google Maps, 2017) in A) and B) have the same scale (1:12,000).

consists mainly of wheat, maize and rapeseed (Landesvermessung und Geobasisinformation Brandenburg (InVeKoS, 2014)). The study area representing the complex landscape, S Germany, was located in Bavaria 50 km north of Munich (centred at 48° 48′ N; 11° 86′ E). The 256 km² area is characterized by small-scale agriculture with an average field size of 2.9 ± 0.04 ha and an average field edge length of 44.2 ± 2.7 km per 1 km² (mean ± SE; calculated based on maps provided by the Bayerische Vermessungsverwaltung, 2014). The amount of (semi-) natural landscape elements like hedgerows, tree stands and fallow land comprises ∼32 % of the total area available (European Environment Agency, 2018b). Arable land covers 62 % of the study area (European Environment Agency, 2018b) and the main crop types are wheat, maize and grassland (Bayerisches Landesamt für Statistik und Datenverarbeitung, 2016). For more detailed characteristics of the two study areas please see the supplementary material (Mulitmedia Component 4).

weighed, sexed and collared following the recommendations of Rühe and Hohmann (2004). The 69 g collars (Model A1, e-obs GmbH, Munich – Germany, www.e-obs.de) consisted of a GPS unit and a tri-axial acceleration sensor, which provides the possibility to use acceleration informed GPS duty cycles. During active periods GPS fixes were taken every full hour, during inactive periods GPS fixes were recorded every four hours. Inactivity was determined automatically by the acceleration sensor when three consecutive acceleration samples did not surpass a variance threshold of 700 (e-obs raw values without unit). The acceleration sensor was programmed to record a movement burst every 4 min. Each burst was recorded at 33 Hz for 3.27-seconds, receiving 110 samples per burst per axis. All tracking and acceleration data are stored at www.movebank.org (Wikelski and Kays, 2015). 2.3. Movement parameters We used four different movement parameters: the number of GPS fixes on the focal field, range size, range shift and energy expenditure, to describe hare movement behaviour 4 days before and 4 days after agricultural management events (see section 2.4 for the explanation about management events and the focal field). The number of GPS points on the focal field was extracted by overlaying the topological field boundary polygons (Bayerische Vermessungsverwaltung 2014, InVeKoS, 2014) with the GPS fixes using the R packages rgdal (Bivand et al., 2014), rgeos (Bivand and Rundel, 2016) and raster (Hijmans and

2.2. Animal tracking We equipped 36 adult hares with GPS collars from April to July in 2014 and 2015 in both study areas simultaneously (for detailed information and deployment times see supplementary material Mulitmedia Component 3): 21 hares were collared in NE Germany (14 individuals in 2014 and 7 in 2015) and 15 hares in S Germany (9 individuals in 2014 and 6 in 2015). Hares were driven into woollen nets, 3

Agriculture, Ecosystems and Environment 292 (2020) 106819

W. Ullmann, et al.

days before and after a day without a management event (a total of 9 days in which no agricultural management event occurred). The baseline samples did not include a focal field and therefore we were not able to calculate the number of GPS points on the focal field. We included the baseline into our analysis to account for possible changes in movement parameters that might occur naturally during a 9-day period, irrespective of a management action. Each hare provided various data table entries, as each home range (seen over the course of the entire deployment period) covers several agricultural fields. Each of these fields experienced management events with or without resource changes or no management (i.e. the baseline) over time and thereby contributed to the data set. Sometimes various fields of the same production cover type were managed at the same day with the same management type. In this case we combined all respective fields and pooled them to one data entry. The amount of recorded management events per hare ranged from 1 to 17 and averaged around 4.5 ± 3.6 events (mean ± SD, for further details see supplementary material Mulitmedia Component 3). The home ranges of three hares in NE Germany and two hares in S Germany partly overlapped and therefore sometimes included the same focal field. Hence, a total of 11 management events were recorded for more than one hare (10 in NE Germany and 1 in S Germany).

Van Etten, 2014). Ranges were calculated using the R package adehabitatHR (Calenge, 2006). We use the term “range” (to refer to range size and range shift) instead of “home range”, because agricultural management events last for short periods of time, with the ranging behaviour of animals only being temporarily affected. A home range, on the other hand, is defined as the area used by an animal during its normal activities including foraging, mating and caring for young (Burt, 1943). We used 95 % minimum convex polygons (MCP) to calculate the 4-day ranges before and after each management event. We only included ranges that were based on at least 15 GPS fixes per day (> 60 % of the data). The MCP area of each 4-day range was used to measure range size. The distance between the centres of the “before” and “after” ranges was used to calculate the range shift, using the R package rgeos (Bivand and Rundel, 2016). The overall dynamic body acceleration (ODBA) – a proxy for energy expenditure – was calculated as described by Scharf et al. (Scharf et al., 2016). We first calculated ODBA (the values have no units) for each single day of the 4-day time period. Subsequently, the mean was used to gain one ODBA value for each 4day time period. We only calculated the ODBA for the time from 22:00 to 02:00 at night as hares are often active during night time. Hares shift their activity time with advancing sunset/sunrise (Schai-Braun et al., 2012), so we avoided the daylight shift by using only hours that were always set in the dark period within the study areas and over the course of the study period. For the number of GPS fixes on the focal field, range size and energy expenditure we accounted for seasonality by calculating the change between the “before” movement parameter and the corresponding “after” movement parameter, and thus we subtracted the “before” values from the “after” values. Otherwise our data may have been biased towards management events occurring early in the year, as hare range sizes are larger in spring and winter compared to summer and fall (Smith et al., 2004).

2.5. Statistical analyses We used linear mixed-effects models (R package lme4 (Bates et al., 2014)) to test the effect of agricultural management events with and without resource changes on the four movement parameters. For all four movement parameters we analysed each of the production cover types (wheat, grassland, maize, and rapeseed) in a separate model. Animal ID was the random effect for all models. We used the following fixed effects: the management type (baseline, with or without resource change), the two different study areas (NE and S Germany), the study year (2014 and 2015) and the date of the management event (Julian date). Further, we included an interaction between the management type and the study area, as we expected stronger reactions of movement parameters to agricultural management events in NE Germany than in S Germany. There was no need to apply transformations to the explanatory variables, but Julian date was standardized to a zero mean and 0.5 SD to avoid estimates of very different scales between explanatory variables (Grueber et al., 2011). We assessed the assumption of normality visually by simulating scaled residuals with the R package DHARMa (Hartig, 2017). Range shift was log-transformed to assure normality and homoscedasticity. Model selection was done by constructing a set of all possible submodels from the global models (Dochtermann and Jenkins, 2011). We applied an information theoretic approach based on AICc values due to small sample sizes (Burnham and Anderson, 2002) and selected all models within the range of ΔAICc < 2 from the most parsimonious model (the model with the lowest AICc). We then retrieved model averaged parameter estimates by averaging over the competing models (Johnson and Omland, 2004). The R package MUMIn (Barton, 2013) was used for the model selection and averaging process. We used 95 % confidence intervals that did not overlap zero as a measure to show consistent effects (Grueber et al., 2011). The effect plots are based on the most parsimonious models (R package effects (Fox and Weisberg, 2018)). Unless indicated otherwise, we present consistent variable estimates together with standard errors. The statistical analyses were conducted using R3.6.1 (R Core Team, 2016).

2.4. Management assessment Local famers were asked for information about the agricultural management measures within each hare’s range. We collected information for 73 % of all fields and only those were included in the analysis. We excluded samples in which hares displayed less than four GPS fixes on the focal field during the 4 days before or after the management event. We recorded hare movement parameters for management events with resource change (harvesting and mowing), events without resource change (e.g. fertilizer application), and for periods without management (hereafter referred to as baseline). We refer to harvesting when crop fields are concerned and to mowing when the management event with resource changes occur on grasslands. For details on the management practices please see supplementary material Mulitmedia Component 5. Movement parameters for the two management types were calculated for 4 days before and 4 days after a day with a management event. We call the field on which the management event took place “the focal field”. We recorded the main production cover type (wheat, grassland, maize, and rapeseed) for each managed field (Table 1). Movement parameters for the baseline were calculated for 4 Table 1 The number of agricultural management events with and without resource changes and the baseline (no management as a control) in NE and S Germany that were used for statistical analyses. The numbers given for wheat, grassland, maize, and rapeseed include management types with and without resource changes. In contrast, the numbers given for the baseline only include samples without any kind of management. Study area

Baseline

Wheat

Grassland

Maize

Rapeseed

NE Germany S Germany Total

84 61 145

40 34 74

4 15 19

19 17 36

18 9 27

3. Results Visually inspecting hare movement behaviour reveals various responses to agricultural management events (Fig. 2 and Table 2). The individual variability ranged from increases in range size with simultaneous large range shifts to no change in range size and shift, but 4

Agriculture, Ecosystems and Environment 292 (2020) 106819

W. Ullmann, et al.

Fig. 2. Hares showed individually different reactions to management events. Here we show the movement reactions of four hares to the harvests of different crop types. Examples for these individual differences are: A) large range shift and large change in range size, B) large range shift and small change in range size, C) small range shift and large change in range size and D) small range shift and small change in range size. The name of the crop overlays the harvested field (outlined in light orange). The “before” ranges and GPS points are depicted in blue, while the “after” ranges are in red. The large circles show the range centre, whereas the white line shows the range shift. NE Germany is represented in panels A) and C), while S Germany is shown in panels B) and D). The satellite images (Bing Maps, 2019) in all panels show the same scale (1:9000). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

an increased amount of GPS points on the focal field.

showing that hares spend more time (17.3 ± 6.1 GPS points) on the focal field after harvest (Fig. 3, Table 2). Further, we found that the number of GPS fixes on rapeseed fields increased after harvest by 11.3 ± 4.8 GPS fixes (Fig. 3, Table 2).

3.1. Number of GPS fixes On average, we received 82 ± 8 GPS fixes (mean ± standard deviation) for each 4-day period before and after a management event, where a maximum of 96 fixes was possible. We detected four models in the confidence set for wheat fields, two for grassland, three for maize and one model for rapeseed (Mulitmedia Component 6). The averaged model for wheat fields contained the variables management type (part of the entire confidence set), study area, study year and Julian date (Table 2, Mulitmedia Component 6), and showed that the number of GPS fixes decreased slightly by 5.4 ± 4.9 after management events without resource changes (Fig. 3, Table 2). Management type was not contained in the averaged model for grassland, but was part of the averaged model for maize fields

3.2. Range size The average 4-day range size in NE Germany was 23.9 ± 18.2 ha and ranged from 0.3 ha to 98 ha. The 4-day ranges in S Germany fluctuated between 0.7 ha and 50 ha with an average range size of 8.0 ± 7.8 ha. We found three models for wheat fields, which all contained the interaction term between management type and study area, the averaged model included all explanatory variables. None of the models for grasslands, maize and rapeseed included management type (Table 2, ST3). Hares on wheat fields in NE Germany increased their range sizes 5

Agriculture, Ecosystems and Environment 292 (2020) 106819

W. Ullmann, et al.

Table 2 Estimates (β), standard error (SE), confidence intervals (CI) and variable importance (Imp.) are shown after model averaging. Models contained the variables: management type (baseline, with, without resource change), study area (S Germany, NE Germany) and their interaction (indicated by “:”), as well as study year (2014, 2015) and the date of the management event. Predictor variables for which the confidence intervals did not include zero are shown in bold and italic. Predictors that slightly overlap zero are shown in bold. MT: management type (b - baseline, w - with, wo - without). SA: Study area (S - S Germany, NE - NE Germany). SY: Study year (2014, 2015). Jdate: Julian date, for the date at which the management event occurred. The reference levels are: aMT(wo), bMT(b), cSA(S) and dSY(2014). Additionally we show the marginal R² values (mR²) and the conditional R² values (cR²), which were derived from the best model of the confidence set using the R-package MuMIn (Barton, 2013). GPS fixes wheat Fixed effects (Intercept) MT(w)a SA(NE)c SY(2015)d Jdate MT(w):SA(NE) R²values

β −5.54 12.74 −3.32 −8.82 −1.76 – mR²:

grassland SE 3.12 5.40 4.75 6.11 2.32 – 0.13

CI (-11.76 - 0.7) (1.96–23.52) (-12.82 - 6.19) (-21.04 - 3.41) (-6.39 - 2.88) – cR²:

Imp. 1.00 1.00 0.15 0.32 0.15 – 0.16

β 1.21 – – – −4.87 – mR²:

maize

SE 3.88 – – – 3.78 – 0.00

CI (-7.07–9.48) – – – (-12.96 - 3.23) – cR²:

Imp. 1.00 – – – 0.29 – 0.33

β 2.50 15.69 – 9.28 – – mR²:

rapeseed SE 6.56 7.26 – 8.96 – – 0.11

CI (−10.74 - 15.73) (0.87–30.5) – (-9.04–27.6) – – cR²:

Imp. 1.00 0.75 – 0.22 – – 0.51

β −3.35 14.68 – – – – mR²:

SE 2.64 5.48 – – – – 0.22

CI (-8.71 - 2.01) (3.52–25.85) – – – – cR²:

Imp. 1.00 1.00 – – – – 0.22

SE 1.44 – – – 2.84 1.54 – – 0.00

CI (-2.34 - 3.35) – – – (-3.46 - 7.79) (-1.76 - 4.34) – – cR²:

Imp. 1.00 – – – 0.24 0.25 – – 0.00

SE 0.23 0.38 0.25 0.25 0.26 0.12 – – 0.17

CI (3.41–4.29) (0.28–1.76) (-0.01 - 0.99) (0.09–1.08) (-0.2 - 0.82) (-0.37 - 0.09) – – cR²:

Imp. 1.00 1.00 1.00 1.00 0.28 0.29 – – 0.32

SE 0.63 2.40 1.37 0.95 1.01 0.52 – – 0.00

CI (-1.46 - 1.03) (-1.61 - 7.85) (-0.38 - 5.04) (-1.29 - 2.45) (-1.26 - 2.73) (-1.3 - 0.75) – – cR²:

Imp. 1.00 0.45 0.45 0.20 0.22 0.10 – – 0.00

Range size wheat Fixed effects (Intercept) MT(w)b MT(wo)b SA(NE)c SY(2015)d Jdate MT(w):SA(NE) MT(wo):SA(NE) R²values

β −0.39 −2.56 2.72 1.16 3.45 0.75 18.79 −8.74 mR²:

grassland SE 2.56 4.51 3.75 3.05 2.52 1.18 7.39 4.81 0.10

CI (-5.45 - 4.68) (-11.48 - 6.36) (-4.7–10.13) (-4.87 - 7.18) (-1.54 - 8.42) (-1.57 - 3.07) (4.17–33.4) (-18.24 - 0.77) cR²:

Imp. 1.00 1.00 1.00 1.00 0.37 0.18 1.00 1.00 0.10

β 0.13 – – 1.43 1.76 1.98 – – mR²:

SE 1.45 – – 2.35 2.44 1.33 – – 0.00

maize CI (-2.76 – – (-3.25 (-3.08 (-0.66 – – cR²:

- 3.01) - 6.1) - 6.59) - 4.62)

Imp. 1.00 – – 0.22 0.24 0.50 – – 0.00

β −0.25 – – −3.12 −4.29 – – – mR²:

rapeseed SE 2.34 – – 3.14 3.12 – – – 0.00

CI (-4.88 - 4.39) – – (-9.37 - 3.15) (-10.49 - 1.93) – – – cR²:

Imp. 1.00 – – 0.18 0.54 – – – 0.00

β 0.51 – – – 2.17 1.29 – – mR²:

Range shift wheat Fixed effects (Intercept) MT(w)b MT(wo)b SA(NE)c SY(2015)d Jdate MT(w):SA(NE) MT(wo):SA(NE) R²values

β 4.14 0.41 0.45 0.28 – −0.06 – – mR²:

grassland SE 0.16 0.27 0.18 0.17 – 0.09 – – 0.07

CI (3.84–4.44) (-0.12 - 0.92) (0.11 - 0.79) (-0.05 - 0.61) – (-0.23 - 0.12) – – cR²:

Imp. 1.00 1.00 1.00 0.69 – 0.20 – – 0.07

β 4.08 0.63 0.51 0.48 0.17 −0.17 – – mR²:

maize SE 0.23 0.32 0.48 0.24 0.23 0.13 – – 0.09

CI (3.65–4.52) (-0.01–1.25) (-0.45 - 1.45) (0.02 - 0.94) (-0.28 - 0.61) (-0.42 - 0.09) – – cR²:

Imp. 1.00 0.38 0.38 0.80 0.16 0.36 – – 0.86

β 3.92 0.44 0.69 0.68 – – – – mR²:

rapeseed SE 0.17 0.25 0.21 0.20 – – – – 0.24

CI (3.59–4.25) (-0.05 - 0.93) (0.29–1.09) (0.29–1.09) – – – – cR²:

Imp. 1.00 1.00 1.00 1.00 – – – – 0.34

β 3.85 1.02 0.49 0.58 0.32 −0.15 – – mR²:

ODBA wheat Fixed effects (Intercept) MT(w)b MT(wo)b SA(NE)c SY(2015)d Jdate MT(w):SA(NE) MT(wo):SA(NE) R²values

β −0.19 – – 0.74 – – – – mR²:

grassland SE 0.69 – – 1.12 – – – – 0.00

CI (-1.55 - 1.18) – – (-1.46 - 2.93) – – – – cR²:

Imp. 1.00 – – 0.31 – – – – 0.00

β −0.06 – – −0.48 0.40 – – – mR²:

maize SE 0.58 – – 0.95 0.97 – – – 0.00

CI (-1.19 - 1.08) – – (-2.34 - 1.39) (-1.51 - 2.31) – – – cR²:

by 17 ± 5.6 ha after harvests, while they showed a trend to decrease the ranges by 5.0 ± 2.3 ha after management events without resource change (Fig. 4, Table 2).

Imp. 1.00 – – 0.22 0.23 – – – 0.00

β −0.17 1.29 2.72 – 1.01 −0.27 – – mR²:

rapeseed SE 0.61 1.92 1.43 – 1.00 0.49 – – 0.00

CI (-1.35 (-2.49 (-0.12 – (-0.97 (-1.22 – – cR²:

- 1.03) - 5.06) - 5.54) - 2.97) - 0.7)

Imp. 1.00 0.39 0.39 – 0.32 0.12 – – 0.00

β −0.22 3.13 2.34 0.58 0.74 −0.28 – – mR²:

Mulitmedia Component 1). For wheat fields we found that range shifts increased after management events without resource changes from 69 ± 8 m (baseline) to 108 ± 16 m (Fig. 5, Table 2), and that hares in S Germany showed a trend of shorter (72.8 ± 9.3 m) range shifts than in NE Germany (107.6 ± 10.8 m, Mulitmedia Component 1). On grasslands range shifts increased slightly from 66 ± 9 m (baseline) to 128 ± 44 m after mowing (Fig. 5, Table 2), and hares showed larger range shifts in NE Germany (98.7 ± 16.2 m) than in S Germany (57.4 ± 9.5 m, Mulitmedia Component 1). Further, we found that hares on maize fields shifted their ranges by 73 ± 9 m in the baseline

3.3. Range shift We detected three models for wheat, eight for grassland, one for maize and three for rapeseed in the respective confidence sets (Table 2, Mulitmedia Component 6). All averaged models for the single production cover types included management type and study area (Table 2, 6

Agriculture, Ecosystems and Environment 292 (2020) 106819

W. Ullmann, et al.

after management events without resource changes (Fig. 5, Table 2). Hares shifted their ranges on rapeseed fields on average by 56.9 ± 11.8 m in S Germany and by 101.1 ± 16.4 m in NE Germany (Mulitmedia Component 1). 3.4. Energy expenditure (ODBA) We detected two models within the confidence set for wheat, eight for grasslands, five for maize and seven for rapeseed, few of them (2–3 per production cover type) included management type (Mulitmedia Component 6). However, even if management type or other explanatory variables were retained in one of the averaged models, the confidence intervals always included zero. Hence, there was no apparent effect of agricultural management, the two study areas and production cover type on energy expenditure. 4. Discussion

Fig. 3. Effects plots showing the change in the number of GPS fixes on the focal field after management events with and without resource changes for the different models: when all production cover types are analysed in combination, and only for wheat, maize and rapeseed fields. There is no graph for grasslands, as there was no effect of management type on the number of GPS fixes on grasslands. Shown are the group means for the number of GPS points ( ± 95 % confidence intervals) averaged over the respective other categorical explanatory variables in the model, e.g. study year (2014 and 2015) or study area (S Germany and NE Germany). See Table 2 and Mulitmedia Component 6 for details on explanatory variables that were used to calculate the group means.

Our results show that hares adjust their movement behaviour and space use in response to management type, study area and production cover type. We found for example more GPS fixes on crop fields, in general, after harvest and fewer GPS fixes on wheat fields after management events without resources changes. Range sizes were only affected in NE Germany on wheat fields and increased after the harvest, while they decreased after management events without resource change. Range shifts became larger after mowing grasslands, harvesting rapeseed and maize fields, and after management events without resource changes on wheat, maize and rapeseed fields. 4.1. Management events with resource changes European brown hares shifted their ranges towards the focal field after management events with resource changes and also spent more time on most of the crop fields. This might have two non-exclusive reasons. First, hares forage on the harvested field for fallen grains, corn stalks and freshly sprouted weeds (Späth, 1989). Second, they incorporate the newly gained habitat into their range because they favour areas with low vegetation height, most likely to improve vigilance and to spend less energy to move through neighbouring dense and high crop fields (Godt et al., 2010; Marboutin and Aebischer, 1996; Mayer et al., 2018; Smith et al., 2004). Standing crops are too high for hares, which prefer open habitat to easily perceive predators (Tapper and Barnes, 1986; Rühe, 1999). In summer most of the crops are high and the area available with low vegetation is limited (Schai-Braun et al., 2014). Therefore, the first harvests of the year (usually barley and rapeseed) provide the primary opportunity for hares to cross and use recently impassable areas. Schai-Braun et al. (2014) showed that hares in a complex agricultural landscape did not shift, but instead increased their ranges after cereal harvest. The authors argued that hares increased their range size to incorporate alternative habitat, as the harvested fields could no longer be used e.g. for cover. Our findings show no effect on range size in the complex landscape, but increased range shifts. However, both of these movement parameters (i.e. increased range size and shift) might just be seen as different methods to incorporate alternative habitats or to actually use the modified field. Grasslands seemed to be the production cover type least influencing hare movement behaviour, except that mowing increased the range shift. Despite the wide spread view of grasslands as (semi-) natural landscape elements, many grasslands are managed intensively and are not often preferred, or even avoided, by hares (Schai-Braun et al., 2013 Mayer et al., 2018). Hence, conservation management should specifically take the type of grassland (intensive vs. extensive) into account.

Fig. 4. Effects plots showing the change in hare range size ( ± 95 % confidence intervals) on wheat fields after management events with and without resource changes compared to baseline events (no management event) in NE Germany and in S Germany.

treatment, while they increased the range shifts after agricultural management events without resource change to 145 ± 27 m and showed a trend to increase the shifts to 114 ± 29 m after management events with resource changes (Fig. 5, Table 2). Range shifts on maize fields were larger in NE Germany (127 ± 18.5 m) than in S Germany (64.3 ± 10.5 m, Mulitmedia Component 1). On rapeseed fields hares shifted their ranges by 68 ± 9 m in the baseline, while they increased the shift to 188 ± 81 m after management events with resource changes, and showed a trend to increase their range shift to 110 ± 27 m 7

Agriculture, Ecosystems and Environment 292 (2020) 106819

W. Ullmann, et al.

Fig. 5. Effects plots showing the range shift from the “before” to the “after” range centre, for models analysing all production cover types in combination, and separately: wheat, grasslands, maize and rapeseed fields. Management types: baseline, with and without the change of resources on the focal field. Please note the logarithmic y-axis.

4.2. Management events without resource changes

quality matrix increases the need for (semi-)natural habitat patches to be large and close by to sustain metapopulations in agricultural landscapes (Bender and Fahrig, 2005). Animals in simple landscapes with large fields would have to move longer distances to reach adequate landscape elements and additionally have a lower matrix quality between the distant habitat patches. These features can translate into larger home ranges (Smith et al., 2005, 2004; Ullmann et al., 2018), followed by decreasing individual body condition (Daan et al., 1996; Mace and Harvey, 1983; McNab, 1963) and a subsequent drop in reproductive output, hence decreased fitness (Doherty and Driscoll, 2018). Despite the increase in home range size, fields can be too large to cross when the crops are high. This barrier function temporarily disrupts animal movements between habitats – especially, in simple landscapes where it is infeasible to circumvent large fields or pass through them. High and dense standing crops are not used by hares as the energy needed to pass through and the detection by predators are high (Mayer et al., 2018; Rühe, 1999). In complex landscapes on the other hand, hares may easily circumvent small fields, when the production cover type is high and dense, and thus keep easy connections between different habitats (Mayer et al., 2018; Ullmann et al., 2018).

Hares’ movement reaction to agricultural management without resource changes depended on the production cover type. Hares avoided wheat fields by shifting their ranges away from the managed fields. In NE Germany hares additionally decreased their range sizes on wheat fields, probably looking for safe places to decrease the risk of being overrun. Cimino and Lovari (2003) showed similar effects on the movement behaviour of roe deer, which shift their ranges away from ploughed fields. Rühe (2002) on the other hand showed that hares do not necessarily avoid fields after the application of plant protection products. In our study, hares did not avoid grasslands, maize and rapeseed fields after management events without resource changes. Hence, animals show avoidance behaviour towards some but not all production cover types and management types. Therefore, the actual movement reaction of hares might depend on a more specific practice of the management event, which was not considered in this study. For example, while plant protection products might not be a problem for hares (Rühe, 2002), no studies exist (to our knowledge) investigating the effect of organic versus inorganic fertilizer on hare movement behaviour and whether the different management techniques pose barriers to the hares’ movement or not. We also showed that hares shift their ranges after management events without the change of resources (except for grasslands). On maize and rapeseed fields they did not simultaneously increase or decrease the number of GPS fixes on the focal field, hence they might shift their range from one end of the focal field to another to avoid disturbances by the agricultural machinery.

4.4. Energy expenditure Animals that are disturbed by human activity often suffer a net loss in their energy budget resulting in poor body conditions (Bechet et al., 2004; Hertel et al., 2016) and reduced reproductive output (French et al., 2011; Strasser and Heath, 2013). Surprisingly, we could find no evidence of increased energy expenditure after agricultural management events. Hence, in our case it seems that hares follow Blumstein’s (2016) theory of not fleeing but being more vigilant. However, if the animal runs away from agricultural machinery only once or twice per day during the management event, the measured energy expenditure (ODBA) might not increase significantly. In our study we found that daily ODBA values are very similar, regardless of the hare running four or six times per day (see supplementary material Mulitmedia Component 2). In any case, we would have expected that an increase in range size (as was shown for wheat fields in the simple landscape) would produce an increase in energy expenditure (Daan et al., 1996; Harestad and Bunnel, 1979). However, some of the animals that contributed to the sample of wheat fields in NE Germany increased their range after harvest, but their ODBA stayed almost the same (supplementary material Mulitmedia Component 2). There was only one hare that increased its range as well as its energy expenditure. Thus, intraspecific differences might account for the discrepancy between energy expenditure and range size in our study, leading to no response in the

4.3. The two study areas and their landscape structure We believe that the landscape structure, i.e. the size and intrinsic spatial configuration of production cover types, is one of the most important differences between our two study areas. Our results support this hypothesis, suggesting that landscape structure plays a role in animal movement behaviour in agricultural landscapes. We showed that agricultural management affected range sizes only on wheat fields in NE Germany, the representative of the simple landscape, and that range sizes and shifts were usually higher in NE Germany than in S Germany. In other words, hares in structurally simple landscapes show more or stronger movement reactions to management events. The increased range sizes and shifts in simple landscapes point towards the need to move longer distance to find alternative habitats. Complex landscapes, with many different kinds of landscape elements on the other hand, provide more alternatives to forage and to find cover in case of resource changes (Beasley et al., 2007). Therefore, they also pose a higher matrix quality between (semi-)natural habitat patches (Fahrig, 2007). A low 8

Agriculture, Ecosystems and Environment 292 (2020) 106819

W. Ullmann, et al.

average movement behaviour. Moreover, hares might also compensate for increased range sizes by executing different fractions of behavioural categories, like resting, foraging, running, feeding. They might move more to explore but also find more food in a newly gained patch and thus compensate with smaller foraging movements, in the fresh patch, that do not require much energy.

Appendix A. Supplementary data

5. Conclusion

Báldi, A., Faragó, S., 2007. Long-term changes of farmland game populations in a postsocialist country (Hungary). Agric. Ecosyst. Environ. 118, 307–311. Barton, K., 2013. MuMIn: Multi-model Inference. R Package Version 1.9. 5 [WWW Document]. Batáry, P., Gallé, R., Riesch, F., Fischer, C., Dormann, C.F., Mußhoff, O., Császár, P., Fusaro, S., Gayer, C., Happe, A.-K., 2017. The former Iron curtain still drives biodiversity–profit trade-offs in German agriculture. Nat. Ecol. Evol. 1, 1279. Bates, D., Maechler, M., Bolker, B., Walker, S., 2014. lme4: Linear Mixed-effects Models Using Eigen and S4. R Packag. Version 1. pp. 1–23. Bayerisches Landesamt für Statistik und Datenverarbeitung, 2016. Erntemengenanteile der Fruchtartgruppen in Bayern 2015 in Prozent [WWW Document]. https://www. statistik.bayern.de/statistik/landwirtschaft/#. Bayerische Vermessungsverwaltung, 2014. Geobasisdaten zur tatsächlichen Nutzung. Beasley, J.C., Devault, T.L., Rhodes Jr., O.E., 2007. Home-range attributes of raccoons in a fragmented agricultural region of northern Indiana. J. Wildl. Manage. 71, 844–850. Bechet, A., GIROUX, J., Gauthier, G., 2004. The effects of disturbance on behaviour, habitat use and energy of spring staging snow geese. J. Appl. Ecol. 41, 689–700. Bender, D.J., Fahrig, L., 2005. Matrix structure obscures the relationship between interpatch movement and patch size and isolation. Ecology 86, 1023–1033. Benton, T.G., Vickery, J.A., Wilson, J.D., 2003. Farmland biodiversity: is habitat heterogeneity the key? Trends Ecol. Evol. 18, 182–188. https://doi.org/10.1016/S01695347(03)00011-9. Bing Maps, 2019. Obtained Via QGIS (version 3.1). DigitalGlobe (© 2019 Microsoft) Source: http://ecn.t3.tiles.virtualearth.net/tiles/a{q}.jpeg?g=1 [Accessed 13 November 2019]. . Bivand, R., Keitt, T., Rowlingson, B., 2014. Rgdal: Bindings for the Geospatial Data Abstraction Library. R Package Version 0.8-16 [WWW Document]. Available at. http://CRAN.R-project.org/package=rgdal. Bivand, R., Rundel, C., 2016. Rgeos: Interface to Geometry Engine - Open Source (GEOS). R Package Version 0.3-21 [WWW Document]. URL. https://cran.r-project.org/ package=rgeos. Blumstein, D.T., 2016. Habituation and sensitization: new thoughts about old ideas. Anim. Behav. 120, 255–262. Brinkman, T.J., Deperno, C.S., Jenks, J.A., Haroldson, B.S., Osborn, R.G., 2005. Movement of female white‐tailed deer: effects of climate and intensive row‐crop agriculture. J. Wildl. Manage. 69, 1099–1111. Brown, D.D., Kays, R., Wikelski, M., Wilson, R., Klimley, A.P., 2013. Observing the unwatchable through acceleration logging of animal behavior. Anim. Biotelem. 1, 20. Burnham, K.P., Anderson, D.R., 2002. Model selection and multimodel inference: a practical information-theoretic approach. Springer Sci. Bus. Media. Burt, W.H., 1943. Territoriality and home range concepts as applied to mammals. J. Mammal. 24, 346–352. Calenge, C., 2006. The package “adehabitat” for the R software: a tool for the analysis of space and habitat use by animals. Ecol. Modell. 197, 516–519. Cimino, L., Lovari, S., 2003. The effects of food or cover removal on spacing patterns and habitat use in roe deer (Capreolus capreolus). J. Zool. 261, 299–305. Daan, S., Deerenberg, C., Dijkstra, C., 1996. Increased daily work precipitates natural death in the kestrel. J. Anim. Ecol. 539–544. Dochtermann, N.A., Jenkins, S.H., 2011. Developing multiple hypotheses in behavioral ecology. Behav. Ecol. Sociobiol. 65, 37–45. Doherty, T.S., Driscoll, D.A., 2018. Coupling movement and landscape ecology for animal conservation in production landscapes. Proc. R. Soc. B 285, 20172272. Drygala, F., Zoller, H., 2013. Spatial use and interaction of the invasive raccoon dog and the native red fox in Central Europe: competition or coexistence? Eur. J. Wildl. Res. 59, 683–691. European Environment Agency, 2018a. Land cover and change 2000-2018 based on corine land cover classes (clc). European Environment Agency, 2018b. Corine land cover (clc) 2018 version 20. Eurostat, 2019. Archive:Agri-environmental indicator - intensification - extensification in 2013. Fahrig, L., 2007. Non‐optimal animal movement in human‐altered landscapes. Funct. Ecol. 21, 1003–1015. Fahrig, L., Baudry, J., Brotons, L., Burel, F.G., Crist, T.O., Fuller, R.J., Sirami, C., Siriwardena, G.M., Martin, J., 2011. Functional landscape heterogeneity and animal biodiversity in agricultural landscapes. Ecol. Lett. 14, 101–112. Fox, J., Weisberg, S., 2018. An R Companion to Applied Regression. Sage Publications. French, S.S., González-Suárez, M., Young, J.K., Durham, S., Gerber, L.R., 2011. Human disturbance influences reproductive success and growth rate in California sea lions (Zalophus californianus). PLoS One 6, e17686. Godt, J., Lang, J., Kugelschafter, K., 2010. Dichteentwicklung von Feldhasen (Lepus europaeus) nach Veränderungen des Bewirtschaftungssystems und zusätzlicher Aufwertung eines größeren Landwirtschaftsbetriebes in einer intensiv genutzten Börderlandschaft. In: Lang, J., Godt, J., Rosenthal, G. (Eds.), Fachtagung Feldhase. Ergebnisse der “Fachtagung Feldhase - Der aktuelle Stand der Hasenforschung” 19–20 März 2010. lutra Verlags- und Vertriebsgesellschaft, Tauer, Kassel, pp. 121–131. Google Maps, 2017. Map of Nordwestuckermark and Freising. [online]. Google.

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.agee.2020.106819. References

To our knowledge this is the first study that investigates how a large set of animal movement parameters is affected by different production cover types and the synergistic effects of agricultural machinery and sudden resource changes due to harvest and mowing. We show that both types of agricultural management events (with and without resource change) affect animals’ movement behaviour. After management events without resource changes, hares shift their ranges and avoid wheat fields, but not grassland, maize, and rapeseed fields. On the other hand, they profit from harvested fields, spending more time on them and incorporating them into their range. Hence, in comparison to other species, such as white-tailed deer and red foxes, hares may benefit from crop harvests (Brinkman et al., 2005; Drygala and Zoller, 2013). However, ensuing studies should investigate the effects of the specific differences between management practices, e.g. differences in movement behaviours after the application of mineral fertilizer versus organic fertilizer. Furthermore, the effects of the different chemical plant protection products on animal well-being and reproductive success are still largely unknown (Rühe, 2002). We also showed that range sizes and shifts were larger in NE Germany, the representative of a simple landscape. Consolidated fields—common in NE Germany—are one consequence of intensified agriculture, leading to a strong reduction in biodiversity on all trophic levels (Benton et al., 2003; Lee and Goodale, 2018; Meichtry-Stier et al., 2014). We recommend to provide smaller field sizes, and thus higher structural heterogeneity and sufficient alternative habitat patches throughout the year. This would help to increase hare population numbers, stabilize and improve other farmland wildlife populations of conservation concern and assure the continuous connectivity between habitat patches by mobile linkers, and can help to improve humanwildlife coexistence. Declaration of interest statement The authors have no personal relationships or affiliations with or involvement in any organization or entity with any financial interest in the material discussed in this manuscript. Acknowledgments This study was conducted in cooperation with and funds from the Leibniz Centre for Agricultural Landscape Research (ZALF), the longterm research platform “AgroScapeLab Quillow” (Leibniz Centre for Agricultural Landscape Research (ZALF) e.V.) and within the DFG funded research training group ‘BioMove’ (RTG 2118-1). Part of the telemetry material was also funded by the European fund for rural development (EFRE) in the German federal state of Brandenburg. We thank the employees of the ZALF research station in Dedelow for their help and technical support. We also thank the Leibnitz Institute for Zoo and Wildlife Research Berlin – Niederfinow and Jochen Godt from the University of Kassel for providing the nets to catch hares. We also thank all students and hunters that helped with the hare trapping and the land owners for allowing us to work on their land. Further, we thank Michael S. Crawford for correcting the spelling and grammar in the manuscript. All procedures for the research were obtained in accordance with the Federal Nature Conservation Act (§ 45 Abs. 7 Nr. 3) and approved by the local nature conservation authority (reference number LUGV V32347-22-2013 and 55.2-1-54-2532-229-13). 9

Agriculture, Ecosystems and Environment 292 (2020) 106819

W. Ullmann, et al. Available from: https://www.google.de/maps/place/Nordwestuckermark/@53. 3161736,13.6173236,12z/data= !3m1!4b1!4m5!3m4!1s0x47aa29f485f939db:0x42120465b5e6e40!8m2!3d53. 2973849!4d13.7247244 [Accessed 30 June 2017] and https://www.google.de/ maps/place/Freising/@48.3899113,11.6464432,12z/data=!3m1!4b1!4m5!3m4!1s0x479e6adfada5bee9:0x81dace3d9e56222!8m2!3d48.4028796!4d11.7411846 [Accessed 30 June 2017]. . Grueber, C.E., Nakagawa, S., Laws, R.J., Jamieson, I.G., 2011. Multimodel inference in ecology and evolution: challenges and solutions. J. Evol. Biol. 24, 699–711. Harestad, A.S., Bunnel, F.L., 1979. Home range and body weight-a reevaluation. Ecology 60, 389–402. Hartig, F., 2017. DHARMa: Residual Diagnostics for Hierarchical (multi-level/mixed) Regression Models. R Package Version 0.1. 5. Hertel, A.G., Zedrosser, A., Mysterud, A., Støen, O.-G., Steyaert, S.M.J.G., Swenson, J.E., 2016. Temporal effects of hunting on foraging behavior of an apex predator: Do bears forego foraging when risk is high? Oecologia 182, 1019–1029. Hijmans, R.J., Van Etten, J., 2014. Raster: Geographic Data Analysis and Modeling. R Package Version 2.2-31 [WWW Document]. http://CRAN.R-project.org/package= raster. InVeKoS, 2014. Integriertes Verwaltungs- und Kontrollsystem—Landesvermessung und Geobasisinformation Brandenburg. Johnson, J.B., Omland, K.S., 2004. Model selection in ecology and evolution. Trends Ecol. Evol. 19, 101–108. Lee, M.-B., Goodale, E., 2018. Crop heterogeneity and non-crop vegetation can enhance avian diversity in a tropical agricultural landscape in southern China. Agric. Ecosyst. Environ. 265, 254–263. Mace, G.M., Harvey, P.H., 1983. Energetic constraints on home-range size. Am. Nat. 121, 120–132. Marboutin, E., Aebischer, N.J., 1996. Does harvesting arable crops influence the behaviour of the European hare Lepus europaeus? Wildlife Biol. 2, 83–91. Mayer, M., Ullmann, W., Sund, P., Fischer, C., Blaum, N., 2018. Habitat selection by the European hare in arable landscapes: the importance of small-scale habitat structure for conservation. Ecol. Evol. 8 (23), 11619–11633. McNab, B.K., 1963. Bioenergetics and the determination of home range size. Am. Nat. 97, 133–140. Meichtry-Stier, K.S., Jenny, M., Zellweger-Fischer, J., Birrer, S., 2014. Impact of landscape improvement by agri-environment scheme options on densities of characteristic farmland bird species and brown hare (Lepus europaeus). Agric. Ecosyst. Environ. 189, 101–109. MEL Bundesministerium für Ernährung und Landwirtschaft, 2014. World Development Indicators Cereal Production. Mrlik, V., 1990. Disturbance of the roe deer (Capreolus capreolus) in agrocoenoses of Southern Moravia [Czechoslovakia]. Folia Zool. 39, 25–35. Padié, S., Morellet, N., Cargnelutti, B., Hewison, A.J.M., Martin, J.-L., Chamaillé-Jammes, S., 2015. Time to leave? Immediate response of roe deer to experimental disturbances using playbacks. Eur. J. Wildl. Res. 61, 871–879. Petrovan, S.O., Dixie, J., Yapp, E., Wheeler, P.M., 2017. Bioenergy crops and farmland biodiversity: benefits and limitations are scale-dependant for a declining mammal, the brown hare. Eur. J. Wildl. Res. 63, 49. R Core Team, 2016. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria http://www.R-project.org/. Rühe, F., 2002. Do brown hares (Lepus europaeus) avoid arable crops treated with plant protection chemicals? In: Proc XXIV IUGB Congress, Thessaloniki. Greece. pp. 483–485.

Rühe, F., 1999. Effect of stand structures in arable crops on brown hare (Lepus europaeus) distribution. Gibier Faune Sauvag. Game Wildl. Sci. 16, 317–337. Rühe, F., Hohmann, U., 2004. Seasonal locomotion and home-range characteristics of European hares (Lepus europaeus) in an arable region in central Germany. Eur. J. Wildl. Res. 50, 101–111. https://doi.org/10.1007/s10344-004-0049-9. Saïd, S., Servanty, S., 2005. The influence of landscape structure on female roe deer home-range size. Landsc. Ecol. 20, 1003–1012. Sauerbrei, R., Ekschmitt, K., Wolters, V., Gottschalk, T.K., 2014. Increased energy maize production reduces farmland bird diversity. GCB Bioenergy 6, 265–274. Schai-Braun, S.C., Hackländer, K., 2014. Home range use by the European hare (Lepus europaeus) in a structurally diverse agricultural landscape analysed at a fine temporal scale. Acta Theriol. (Warsz) 59, 277–287. https://doi.org/10.1007/s13364013-0162-9. Schai-Braun, S.C., Peneder, S., Frey-Roos, F., Hackländer, K., 2014. The influence of cereal harvest on the home-range use of the European hare (Lepus europaeus). Mammalia 78, 497–506. https://doi.org/10.1515/mammalia-2013-0099. Schai-Braun, S.C., Rödel, H.G., Hackländer, K., 2012. The influence of daylight regime on diurnal locomotor activity patterns of the European hare (Lepus europaeus) during summer. Mamm. Biol. für Säugetierkd. 77, 434–440. Schai-Braun, S.C., Weber, D., Hackländer, K., 2013. Spring and autumn habitat preferences of active European hares (Lepus europaeus) in an agricultural area with low hare density. Eur. J. Wildl. Res. 59, 387–397. Scharf, A.K., LaPoint, S., Wikelski, M., Safi, K., 2016. Acceleration data reveal highly individually structured energetic landscapes in free-ranging fishers (Pekania pennanti). PLoS One 11 (e0145732). Smith, R.K., Jennings, N.V., Robinson, A., Harris, S., 2004. Conservation of European hares Lepus europaeus in Britain: Is increasing habitat heterogeneity in farmland the answer? J. Appl. Ecol. 41, 1092–1102. https://doi.org/10.1111/j.0021-8901.2004. 00976.x. Smith, R.K., Vaughan Jennings, N., Harris, S., 2005. A quantitative analysis of the abundance and demography of European hares Lepus europaeus in relation to habitat type, intensity of agriculture and climate. Mamm. Rev. 35, 1–24. Späth, V., 1989. Untersuchungen zur Populationsökologie des Feldhasen (Lepus europaeus) in der Oberrheinebene. Selbstverlag des Instituts für Forstzoologie, Freibg. im Breisgau. Stankowich, T., 2008. Ungulate flight responses to human disturbance: a review and meta-analysis. Biol. Conserv. 141, 2159–2173. Strasser, E.H., Heath, J.A., 2013. Reproductive failure of a human‐tolerant species, the American kestrel, is associated with stress and human disturbance. J. Appl. Ecol. 50, 912–919. Tapper, S.C., Barnes, R.F.W., 1986. Influence of farming practise on the ecology of the brown hare (Lepus europaeus). J. Appl. Ecol. 23, 39–52. Ullmann, W., Fischer, C., Pirhofer-Walzl, K., Kramer-Schadt, S., Blaum, N., 2018. Spatiotemporal variability in resources affects herbivore home range formation in structurally contrasting and unpredictable agricultural landscapes. Landsc. Ecol. 1–13. Vasseur, C., Joannon, A., Aviron, S., Burel, F., Meynard, J.M., Baudry, J., 2013. The cropping systems mosaic: how does the hidden heterogeneity of agricultural landscapes drive arthropod populations? Agric. Ecosyst. Environ. 166, 3–14. https://doi. org/10.1016/j.agee.2011.10.012. Vercauteren, K.C., Hygnstrom, S.E., 1998. Effects of agricultural activities and hunting on home ranges of female white-tailed deer. J. Wildl. Manage. 280–285. Wikelski, M., Kays, R., 2015. Movebank: Archive, Analysis and Sharing of Animal Movement Data. [WWW Document]. World Wide Web Electron. Publ.

10