You are on page 1of 8

Forest Ecology and Management 255 (2008) 468–475

www.elsevier.com/locate/foreco

Winter habitat use by red and roe deer in pine-dominated forest


J. Borkowski a,*, J. Ukalska b,1
a
Department of Forest Ecology and Wildlife Management, Forest Research Institute, Se˛kocin-Las, 05-090 Raszyn, Poland
b
Department of Biometry, Faculty of Agriculture and Biology, Warsaw University of Life Sciences, Nowoursynowska 159, 02-776 Warsaw, Poland
Received 31 October 2006; received in revised form 23 May 2007; accepted 5 September 2007

Abstract
Although pines (Pinus spp.) represent the main tree genus in Europe, most studies of habitat use by European deer (Cervidae) have been
conducted in spruce-dominated forests. This neglects the fact that the two forest types are quite different. Ground vegetation is much more
abundant in pine stands, especially when they are mature. In this study, red and roe deer pellet groups were counted in approximately 12,500 ha of
pine forests during the early spring over 4 years. Combinations of food and cover determined habitat attractiveness. Those habitat types which
offered both food and cover were used most intensely. The role of forage was important only where cover was sufficient, suggesting that cover plays
a primary role in influencing winter habitat use by deer.
Habitat use by red and roe deer was similar both among habitats and within them, but the smaller species, roe deer seemed able to satisfy their
cover requirements more easily than the larger red deer. No evidence of interspecific influence of red on roe deer (as suggested by Latham et al.,
1997) was found in this study. As predicted, mature pine stands were an attractive habitat for deer. It was concluded that, in hunted deer populations,
the presence of cover is important even in areas lacking large predators. Introduction of forest understories into mature pine forests should thus be
promoted in big game management.
# 2007 Elsevier B.V. All rights reserved.

Keywords: Central Europe; Cover; Winter; Habitat use; Pine forest; Poland; Red deer; Roe deer

1. Introduction ground vegetation in pine stands is generally much richer than


under the spruce canopy. Since old stands are among the most
Red deer (Cervus elaphus) and roe deer (Capreolus food-abundant age classes in pine-dominated forests (Bor-
capreolus) are the most numerous and widespread deer in kowski, 2003), the question arises as to how deer use old pine
Europe, and are the most extensively studied ungulates (Clutton- stands in comparison with younger stands. Old growth pine
Brock et al., 1982; Andersen et al., 1998). There are many studies forests are expected to be more attractive to deer than their spruce
of habitat use by these two species in European forests (e.g. counterparts. Moreover, most studies on red and roe deer habitat
Staines and Welch, 1984; Catt and Staines, 1987; Thirgood and use come from Scotland. Only a few were conducted in winter,
Staines, 1989; Aulak and Babińska-Werka, 1990a; Welch et al., which is a period of physiological stress, affecting important
1990; Latham et al., 1996; Tufto et al., 1996; Palmer and processes such as population dynamics (among others). Winter is
Truscott, 2003). However, most studies have been conducted in also the season damage caused by deer in forests is most
spruce-dominated (Picea) forests, even though the most concentrated (Gill, 1992). Thus, a better understanding of factors
abundant trees in Europe are pines (Pinus spp.; UN-ECE/ affecting winter habitat use by deer is important for both game
FAO, 1992). There are distinct differences between pine and and forest management.
spruce, especially in mature forests. Due to different crown In general, habitat use by deer is determined by the presence
structure and light conditions (Robakowski et al., 2004), the of both food and cover. The role of cover in habitat use may be
especially important in winter, when cervids reduce their food
intake and live to a remarkable extent off fat reserves (Putman,
1988). Two recognized cover types are thermal cover, in which
* Corresponding author. Tel.: +48 22 7150412; fax: +48 22 7150417.
E-mail addresses: boku@ibles.waw.pl (J. Borkowski),
a forest overstorey protects against weather and sun, and hiding
joanna.ukalska@agrobiol.sggw.waw.pl (J. Ukalska). (security) cover used in escape and as protection against
1
Tel.: +48 22 5932730; fax: +48 22 5932722. predators and humans (Peek et al., 1982). Although many
0378-1127/$ – see front matter # 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.foreco.2007.09.013
J. Borkowski, J. Ukalska / Forest Ecology and Management 255 (2008) 468–475 469

studies have documented the role of thermal cover (see 2. Study area
Mysterud and Østbye, 1999, for references), recent experi-
mental study demonstrated no positive effect of thermal cover The study was carried out near Rudy Raciborskie, in a Forest
on elk (C. elaphus nelsoni; Cook et al., 1998). On the contrary, District of ca. 17,500 ha in southwestern Poland. As a large
the authors found that thermal cover resulted in greater over stand-replacement fire burned about 5000 ha of the District in
winter mass loss, fat catabolism and mortality. Less information 1992, data referred to here are from unburned forest only.
is available on hiding cover, which is generally presumed to be Coniferous habitat types accounted for approximately 65% of
associated with predation risk (Mysterud and Østbye, 1999). In the study area, with Scots pine (Pinus silvestris), the main tree
fact, it is often people who influence deer behaviour and habitat species in the canopy (up to 85% of cover). As there was almost
use most. Hunting, as the extreme manifestation of human no management in the unburned forest for several years after
influence, is obviously an important factor. While animals in the fire, the proportion of new plantations, and especially of
unhunted populations are usually tolerant of human presence pre-thickets, was minor. All of the forest is managed for timber
(Schultz and Bailey, 1978; Borkowski, 2001), hunting tends to with artificial regeneration accounting for the greatest part of it.
promote flight behaviour (Altman, 1958; Behrend and Lubeck, Brown podzolic soils, together with podzols, make up 73% of
1968). Moreover, as hunting is ‘‘human predation,’’ it ensures the soils in the study area. About 20 years ago, foresters in this
that hiding cover will be an important determinant of forest area began to introduce understories as part of a nationwide
habitat use by deer, even where natural enemies are absent. This strategy to enrich homogenous pine stands. The main species
has been demonstrated for red deer by Bonenfant et al. (2004). used for this purpose were Norway spruce (Picea abies),
Kufeld et al. (1988), found that mule deer (Odocoileus common beech (Fagus sylvatica), black dogwood (Frangula
hemionus) used habitats with increasingly better cover as the alnus) and black cherry (Prunus serotina).
hunting season progressed. A reasonable assumption, is that use The winter climate of the study area is very mild. The
of old pine forest (a relatively open habitat) during the hunting average temperature between December and March is 0.2 8C,
season would be positively influenced by understorey cover with January being the coldest month (2.1 8C). The number of
conditions. days with snow covers between December and February is
To date, the role of cover in habitat use among European usually around 45, and average snow depth rarely exceeds
deer, has mostly been studied in roe deer (Mysterud and 5 cm. The area is flat with an elevation range between 180 and
Østbye, 1995), with much less attention given to red deer. 307 m above sea level.
Being much larger than roe deer, red deer may experience At the time of the study, red and roe deer population
greater difficulty finding cover under the same habitat densities (as estimated by drive counts) were of 4–6 and 6–10
conditions. For instance, habitat use by red deer in young individuals/100 ha of forest area, respectively (P. Nasiadka,
pine stands depends much more on tree height than is the case unpubl. data). Other ungulates in the area were wild boar (Sus
with roe deer (Borkowski, 2004). Similar phenomena may also scrofa) and the less-common fallow deer (Dama dama). All the
apply in older forests. species were hunted by both individuals and groups. The
In spite of the numerous studies on both deer species, hunting season varied with the species, but was most intense
relatively little information is available on comparisons of between October and February. There are no large predators in
habitat use and potential competition. However, Latham et al. the area, except for occasional occurrences of single wolves
(1996) compared red and roe deer densities in 20 localities and (Canis lupus). The study area is located in Silesia, a densely
found a negative correlation between them. In the same 20 populated, industrial region of Poland, and is commonly used
forests, the authors also found that red deer density had a for recreation.
negative influence on roe deer density, while the reverse was not
the case (Latham et al., 1997). Both these studies are from 3. Methods
Scotland and there is little information available from other
places. If there is an interspecific influence of red on roe deer, The study was conducted between 1997 and 2000, using
spatial distribution of both species may be negatively related, standing crop counts of faecal pellet groups (Neff, 1968; Collins
not only between areas but also within them. and Urness, 1981; Mayle et al., 1999). Pellet groups were
The objective of this study was to examine patterns of winter counted every year in March along 2 m wide transects. Because
habitat use by red and roe deer in pine-dominated forests. It was the disappearance rate of summer and autumn pellets is relatively
predicted that in pine-dominated forests, deer would use mature rapid in Poland (Aulak and Babińska-Werka, 1990b), all data
stands more frequently as compared to deer in spruce forests. refer to patterns of winter habitat use by red and roe deer. The
Another prediction was that a hunted population would use transects were drawn on a 1:50000 map, so that their spatial
habitats with security cover more intensely than habitats with distribution in the study area was more or less even. Each transect
less security cover, even in predator-free environments. It was started at a characteristic point (e.g. a crossroads), which could be
also predicted that security cover would influence habitat use easily found in the field. The bearing of each transect was
more in large deer species (red deer) than in small ones (roe measured on the map and then walked with a compass. In total
deer). The final prediction was that, if interspecific competition there were eight transects. Transects walked in 1997 differed
between them is a factor, than deer densities in different slightly from subsequent years, when transects were drawn such
habitats within one area would be negatively correlated. that they could be reached by car (using forest and not the main
470 J. Borkowski, J. Ukalska / Forest Ecology and Management 255 (2008) 468–475

Table 1 be preferred, e.g. grasses, raspberry (Rubus ideus), bramble (R.


The length of transect units (km) in different habitat types in year 2000
caesius) and bilberry (Vaccinium myrtillus) or unpalatable
Total Palatable Unpalatable Pal/Unpal (dominated by plants which are usually avoided or seldom
T 2.7 – – – consumed, e.g. small-reed (Calamagrostis sp.) and sedges
PS 3.3 – – – (Carex sp.)). For clarity, analysis was confined to those sections
MF 26.0 – – – in which the ground vegetation could be clearly classified as
De 2.2 – – – dominated by palatable or unpalatable plants (more than 50%
P0 2.1 0.8 0.9 0.4
of ground cover as estimated visually).
P1 4.0 1.2 2.4 0.4
P2 9.0 3.0 3.1 2.9
P3 8.7 2.7 2.5 3.5 4. Statistical methods
T: thicket; PS: pole size forest; MF: mature forest; De: deciduous forest; P0:
pine forest with no understorey; P1: pine forest with sparse understorey; P2: Generalized linear mixed models GLMM (Breslow and
pine forest with medium understorey; P3: pine forest with dense understorey; Clayton, 1993; Littel et al., 1996) were applied to the data for
Palatable: patches dominated by palatable plants; Unpalatable: patches domi- each species separately. Two mixed models were used. The first
nated by unpalatable plants; Pal/Unpal: patches with both palatable and basic model (Model 1) was used to examine whether habitat
unpalatable plants.
type, year and their interaction, influence forest use in red and
roe deer (habitat is defined as transect sections with similar
roads). Total transect length was about 32 km. Each year transect stand age and species, understorey and ground vegetation). The
length could differ slightly, for reasons such as recent forest work second model (Model 2) analysed whether deer utilization of
in a given sub-compartment (e.g. tree cutting and young mature forest with various understorey densities depends on
plantation fencing) or high groundwater level. ground vegetation and its interaction with the understorey.
Pellet groups were counted separately for each part of the Pellet group count data per section was a linear combination
transect crossing an area of similar habitat characteristics (stand of fixed and random effects described by the generalized linear
age class, tree species, understorey and ground vegetation mixed Model 1:
condition). Each such part of the transect was considered a
sampling unit, hereafter called a transect section. The length of hi jk ¼ logðli jk Þ ¼ logðsÞ þ m þ yi þ h j þ ðyhÞi j þ tk (1)
each section was measured from the map or estimated by
where hijk is the log link function log(lijk) with a Poisson error
pacing. The total number of sections in different study years
term, lijk the conditional mean (Littel et al., 1996) for the ijkth
ranged from 108 to 184. Table 1 shows the length of sections in
year–habitat–transect combination, m the intercept, yi the ith
different habitats. Average section length was 220 m, while the
year fixed effect, hj the jth habitat type fixed effect, (yh)ij the
average number of sections per transect was 18.4.
year–habitat interaction and tk is the random effect of the kth
Beside the numbers of red and roe deer pellets alongside
transect. The logarithm of the section area, log(s), is an offset
transects, section data were collected on habitat characteristics
term to allow for the different lengths of the transect sections.
with regards to dominant tree species (pine, deciduous), forest
Model 2 exclusively used data from mature pine forest where
age class (thickets: 16–25; pole stage forests: 26–50; mature
ground vegetation was observed. The form of this model was
forests: >50-year-old stands) and understorey or ground
vegetation conditions. Understorey and ground vegetation hi jkl ¼ logðli jkl Þ ¼ logðsÞ þ m þ yi þ u j þ gl þ ðugÞ jl þ tk
conditions were recorded in mature pine stands only, since both
(2)
were scarce or absent in the other habitats. Understorey (woody
and shrubby vegetation up to 4 m in height) tree/shrub species where lijkl is the conditional mean for the ijklth year–under-
data (spruce, deciduous) was augmented by density informa- storey–ground vegetation–transect combination, uj the jth
tion. The understorey in each transect section was assigned to understorey fixed effect, gl the lth ground vegetation fixed
one of four classes: no understorey (P0), sparse – composed of effect and (ug)jl is the understorey–ground vegetation interac-
single trees or shrubs (P1), medium-density – visibility distance tion and the other effects are the same as in Model 1.
longer than 100 m (P2) and dense – visibility distance below Because empirical count data are typically over-dispersed,
100 m (P3). Thus, the term ‘understorey density’ in this study (i.e., the variance is greater than the mean; Littel et al., 1996),
refers to visibility rather than to shrub/tree stem density. In the significance of fixed effects of both models was based on the
addition, understorey was classified as deciduous or spruce, F-statistics adjusted for over-dispersion. Degrees of freedom
depending on composition. Although the classification of were estimated using the Satterthwaite approximation, which
understorey density was in this sense somewhat arbitrary, the could give non-integer values. Differences between factorial
fact that a single observer collected most data (with only limited combinations were tested using Tukey–Kramer adjustment for
help from another person) assured consistency. unbalanced data (Kramer, 1956).
Data on ground vegetation (herbaceous vegetation and In order to check if cover was equally important in habitat
dwarf shrubs such as berries) was also collected within each used by the two deer species, index of selectivity was
transect section. Based on food preferences of red and roe deer calculated. Index of selectivity was defined as the ratio of
(Dzie˛ciołowski, 1969; Ge˛bczyńska, 1980), the ground vegeta- mean pellet group count values per area unit in examined
tion was classified as palatable (dominated by plants known to habitats to the general mean pellet group count value per area
J. Borkowski, J. Ukalska / Forest Ecology and Management 255 (2008) 468–475 471

unit. The assumption was that the species with more


pronounced cover requirements will ‘‘prefer more’’ covered
habitats and ‘‘avoid more’’ open ones, while the species with
less pronounced cover requirements will not show these
preferences. As a result, the range of the index will be larger for
the former than for the latter.
Correlation between pellet group count data of the two deer
species was checked using the Spearman rank correlation. For
each species, transect residuals from Model 1 were compared
by Pearson’s correlation to see if there was any unexplained
variation correlated in the same way for red and roe deer. Both
correlations were done on the scale of the study area (sections
from all the transects together) and on the habitat scale Fig. 1. Pellet group density means for red (closed bars) and roe (open bars) deer
(separately for each habitat type). per 100 m2 in different habitats. T: thicket; PS: pole stage forest; D: deciduous
forest; P0: pine forest with no understorey; P1: pine forest with sparse under-
Both GLMM models were fitted using the GLIMMIX
storey; P2: pine forest with medium understorey; P3: pine forest with dense
macro-implement in the SAS System for Windows 9.1.3 (SAS understorey; s: spruce understorey; d: deciduous understorey. Error bars show 1
Institute, 2002). Pearson’s correlation was assessed using the S.E.
CORR procedure of the SAS System.
For both species, the rankings were nearly identical and the
5. Results score of given habitat type depended very much on its cover
conditions. More meaningful however, are the comparisons of
In total, 2890 red deer and 3101 roe deer pellet groups were specific pairs of habitats (Table 2). For both deer species, pellet
recorded. Red deer pellet group density differed significantly group densities were higher in mature pine stands with dense
among the habitats (F = 16.14; d.f. = 9, 540.4; p < 0.001) and deciduous and spruce understorey (P3d and P3s), and to some
years of the study (F = 2.74; d.f. = 3, 543.7; p = 0.04), but was extent in mature pine stands with medium density spruce
not influenced by the interaction habitat  year (F = 0.71; understorey (P2s; p < 0.05 or less), than in most other habitat
d.f. = 27, 538.6; p > 0.05). Similarly, habitat type was an types (thickets, pole-sized and deciduous stands). In general,
important factor influencing roe deer pellet group density there were no differences in either red or roe deer pellet group
(F = 9.27; d.f. = 9, 545.3; p < 0.001). Roe deer pellet group densities among mature pine stands with sparse or no
density was similar among the years (F = 1.46; d.f. = 3, 541.4; understorey. Pellet group densities of both species were not
p > 0.05); interaction habitat  year (F = 1.31; d.f. = 27, 542; dependent on the kind of understorey (deciduous vs. spruce) in
p > 0.05). any of understorey density classes ( p > 0.05 in all cases,
It was possible to rank the habitats based on mean pellet Table 2).
group density (Fig. 1): In mature pine stands, in all understorey density classes
except for the sparse understorey, there were significantly more
red deer : D  PS  P1d  P0  P1s  T  P2d  P2s red deer pellet groups in transect sections with the ground
vegetation dominated by palatable plants then in those
 P3d  P3s
dominated by unpalatable plants ( p < 0.05 or less, Fig. 2).
red deer : D  P1d  P0  PS  T  P1s  P2s  P2d In the case of roe deer, more pellet groups were recorded in
sections with palatable plants than in unpalatable ones only in
 P3d  P3s P0, P2s and P3d ( p < 0.05 or less, Fig. 2).

Table 2
Significance of comparisons of red (bold font) and roe deer (underlined font) pellet group densities recorded in different habitats
T PS D P0 P1d P1s P2d P2s P3d P3s
T – ns ns ns ns ns ns ns * ***
PS ns – ns ns ns ns ns *** *** ***
D ns ns – ns ns ns ns/* * *** ***
P0 ns ns ns – ns ns ns * *** ***
P1d ns ns ns ns – ns ns ns ns ns
P1s ns ns * ns ns – ns *** *** ***
P2d ns ns * ns ns ns – ns ns ns
P2s ns * ** ns ns ns ns – ns ***
P3d ** *** *** ** ns ns/* ns ns – ns
P3s *** *** *** *** ns * ns * ns –
T: thicket; PS: pole stage forest; De: deciduous forest; P0: pine forest with no understorey; P1: pine forest with sparse understorey; P2: pine forest with medium
understorey; P3: pine forest with dense understorey; s: spruce understorey; d: deciduous understorey; ns: not significant; ns/*: 0.05 < p < 0.1, *p < 0.05, **p < 0.01,
***p<0.001.
472 J. Borkowski, J. Ukalska / Forest Ecology and Management 255 (2008) 468–475

Table 4
Correlation between red and roe deer pellet group densities in different habitats
Habitat type RS P
T 0.45 *
PS 0.46 *
De 0.51 *
P0 0.27 ns
P1d 0.22 ns
P1s 0.51 *
P2d 0.41 ns
P2s 0.29 ns
P3d 0.71 *
P3s 0.52 *
All habitats 0.57 *
T: thicket; PS: pole stage forest; MF: mature forest; De: deciduous forest; P0:
pine forest with no understorey; P1: pine forest with sparse understorey; P2:
pine forest with medium understorey; P3: pine forest with dense understorey; s:
spruce understorey; d: deciduous understorey; RS: Spearman correlation coeffi-
cient; P: probability; ns: not significant; *p < 0.05.

positive but insignificant ( p > 0.05). Similarly, there was a


positive correlation between the two sets of GLMM residuals
Fig. 2. Pellet group density means and standard errors for red and roe deer per on the scale of the study area (r = 0.44 and p < 0.001, Table 5).
100 m2 in transect sections dominated by palatale (closed bars) and unpalatable Significant ( p < 0.05 or less) positive correlations between the
plants (open bars) in mature pine forest: P0: pine forest with no understorey; P1: residuals were recorded in 7 out of 10 habitats (Table 5).
pine forest with sparse understorey; P2: pine forest with medium understorey;
P3: pine forest with dense understorey; s: spruce understorey; d: deciduous
understorey; ns: not significant. *p < 0.05, **p < 0.01 and ***p<0.001. 6. Discussion

6.1. Pellet count credibility


The index of selectivity (Table 3) for red was more than one
and half times larger than for roe deer (1.88 and 1.17, Dissimilarities in structural characteristics among habitats in
respectively). Similarly, the coefficient of variation of the index this study may have influenced deer pellet group density in
for red deer was nearly twice as large as the one recorded for roe different habitats. Pellet decomposition rates may depend on
deer (39% and 65%, respectively, Table 3). habitat conditions (Aulak and Babińska-Werka, 1990b),
On the scale of the study area, red and roe deer pellet groups particularly openness (Welch et al., 1990; Lehmkuhl et al.,
were similarly distributed (RS = 0.57 and p < 0.05, Table 4). 1994; Massei et al., 1998). However, in each of the cited
Pellet groups of both species were positively correlated in 6 out studies, habitat-related differences in decomposition rates were
of 10 habitats (Table 4), while in 4 remaining ones RS were minor in winter. Thus, it is unlikely that this factor exerted an

Table 5
Table 3 Correlation between the two sets of GLMM residuals (Model 1) for red and roe
Index of selectivity (see Section 3 for details) in red and roe deer habitat use deer in different habitats
Habitat Red deer Roe deer Habitat type RS P
T 0.94 0.83 T 0.47 ***
PS 0.49 0.75 PS 0.52 ***
D 0.25 0.41 D 0.41 *
P0 0.54 0.72 P0 0.47 **
P1d 0.50 0.68 P1d 0.02 ns
P1s 0.61 1.02 P1s 0.18 ns
P2d 1.29 1.29 P2d 0.35 ns
P2s 1.32 1.15 P2s 0.35 ***
P3d 1.93 1.58 P3d 0.65 ***
P3s 2.13 1.58 P3s 0.47 ***
Range 1.88 1.17 All habitats 0.44 ***
CV (%) 65 39
T: thicket; PS: pole stage forest; De: deciduous forest; P0: pine forest with no
T: thicket; PS: pole stage forest; De: deciduous forest; P0: pine forest with no understorey; P1: pine forest with sparse understorey; P2: pine forest with
understorey; P1: pine forest with sparse understorey; P2: pine forest with medium understorey; P3: pine forest with dense understorey; s: spruce under-
medium understorey; P3: pine forest with dense understorey; s: spruce under- storey; d: deciduous understorey; R: Pearson correlation coefficient; P: prob-
storey; d: deciduous understorey; CV: coefficient of variation. ability; ns: not significant; *p < 0.05, **p < 0.01 and ***p<0.001.
J. Borkowski, J. Ukalska / Forest Ecology and Management 255 (2008) 468–475 473

important effect. Similarly, ground vegetation is scarce at the Although a dense understorey offering cover could also
end of winter and should not have much influence on the ability attract deer through its abundant browse, this probably was not
to detect pellet groups. Ground vegetation therefore, was the main reason for the use of such patches. Spruce twigs are
assumed not to influence the results of this study. usually avoided by deer in Poland (Dzie˛ciołowski, 1969;
Borowski and Kossak, 1975), but the sites with a deciduous
6.2. Mature pine stand use understorey of higher quality browse were used as often as
those with spruce-dominated understories. Although Borowski
Our results suggest that mature pine forests (especially some and Kossak (1975) found spruce bark to be important in winter
patches within them), were much more attractive to deer, than diets, in our study area we observed no spruce de-barking.
has been reported for mature stands with spruce (Staines and
Welch, 1984; Catt and Staines, 1987; Latham et al., 1996). The 6.4. Comparison of habitat use between red and roe deer
reason for this is probably poorer food conditions under the
spruce canopy, due to the presence of much less favourable 6.4.1. Importance of ground vegetation
light conditions than in mature pine stands (Robakowski et al., Ground vegetation was an important factor modifying deer
2004). Mature spruce forests with ground vegetation have been habitat use in Rudy Raciborskie. In case of red deer, this trend
shown to sustain greater use on the part of red, and especially was recorded in nearly every understorey density class and use
roe, deer, than those lacking ground vegetation (Welch et al., intensity was positively related to understorey density. Thus,
1990). However, in this study, reference to the ground the cover seems to be a necessary primary condition for
vegetation alone could not account for the use deer made of preferred winter habitat of the studied deer species. Only when
mature stands. For instance, use of deciduous stands offering cover is sufficient will food availability influence deer
little food was similar to that of mature pine stands with little or distribution within this habitat. Although we do not have data
no understorey, that had much more abundant food resources. on animal activity in the studied habitats, these results suggest
that cover is important not only for resting but also for foraging
6.3. Influence of security cover and ground vegetation animals.
The use of mature pine stands by roe deer seemed to be less
As predicted, cover seemed to be a very important influenced by ground vegetation than was found for red deer.
determinant of winter habitat use by the red and roe deer in However, this is probably a result of the coarse classification of
Rudy Raciborskie. General habitat rankings for both species patches with palatable or unpalatable plants. Roe deer is a
very much reflected a gradient of hiding cover conditions concentrate selector (Hofmann, 1985). The classification of the
offered by various habitats. Deciduous mature stands received patches in the present study probably better reflected the
very little use, pole stage stands and mature pine stands without conditions for an intermediate or roughage eater such as red
or with sparse understorey received moderate use, and mature deer (Hofmann, 1985). Contrary to roe deer, red deer depends
pine stands with dense understorey were used most intensely. more on food biomass than on its quality. In addition, unlike red
When the number of visitors to an area is high, deer seek dense deer, roe deer winter fat reserves are relatively small and the
cover (Jeppesen, 1987) even in populations generally tolerant principal source of winter energy is food supply (Holand et al.,
of human presence (Borkowski, 2001). As already mentioned, 1998). Therefore, winter habitat use by roe deer should have
this study area is commonly used, not only by people for been more influenced by food resources than was the case for
recreation, but also for hunting. Hunting promotes intensive red deer.
cover use by deer (Lovaas, 1958; Skolvin, 1982; Kufeld et al.,
1988). However, cover alone was not able to explain habitat use 6.4.2. Importance of security cover
pattern by deer in this study. For instance, there was no In spite of similar habitat use patterns, roe deer (the smaller
significant difference in habitat use intensity between thickets species) seemed to find it easier to satisfy its cover requirements
offering favourable cover conditions, pole stage stands with than did the larger red deer. The idea of the more pronounced
neither food nor cover, or mature pine stands without cover requirements for red deer than for roe deer was supported
understorey which had abundant food but no cover. by wider ranges of the selectivity index and higher coefficient
Thus, the combination of food and cover seemed to of variation for red than for roe deer. A similar conclusion was
determine habitat attractiveness. Mature pine stands with a drawn by Lyon and Jensen (1980), who found that habitat use
dense understorey were the habitat type most preferred by both among the smaller mule deer was less influenced by security
deer species. No other habitat within the study area could have cover than was that of the larger elk. An especially clear
offered both abundant food and good cover conditions. It has relationship between species size and the need for cover was
been documented that simultaneous access to food and cover found in the relatively open post-burn area of young pine stands
can influence home range size. That is, the more fine-grained a at Rudy Raciborskie (Borkowski, 2004).
habitat mosaic, the smaller the home range that can encompass
all required resources (Moe and Wegge, 1994; Borkowski and 6.4.3. Interspecific relations
Furubayashi, 1998). Henry (1981) concluded that forest rides Latham et al. (1996) found that the densities of both deer
were the habitat type most used by roe deer because of high species in different forests of Scotland were negatively
food availability and proximity to cover. correlated. In a subsequent study, Latham et al. (1997) pointed
474 J. Borkowski, J. Ukalska / Forest Ecology and Management 255 (2008) 468–475

out that there were several climatic factors (such as remarkable damage to forests, young plantations near forest
temperature, snow depth, windspeed and rainfall), shaping patches offering security cover may be under higher pressure
the negative relation between red and roe deer densities. than those surrounded by poor cover stands.
Moreover, within individual forests studied by Latham et al.
(1997), red and roe deer used their habitats in different ways. In Acknowledgements
this study, densities of red and roe deer pellet groups were
positively correlated on the scale of the study area and within We are grateful to the Headquarters of the Rudy Raciborskie
many among available habitats. In winter especially, this should Forest District for assistance and the friendly atmosphere.
not come as a surprise, since major differences in the feeding Thanks also to Dr. Pawel Nasiadka for his help and suggestions.
habits of the two species observed in summer are likely to Prof. Rory Putman kindly commented on an earlier draft of the
disappear in winter (Staines et al., 1985; Latham et al., 1999). manuscript.
Therefore, winter habitat use by both species may differ less
than in the other seasons. Similar habitat use patterns between References
both species were also found in other studies (Staines and
Altman, M., 1958. The flight distance in free-ranging big game. J. Wildl.
Welch, 1984; Welch et al., 1990). Latham et al. (1997) Manage. 22, 207–209.
suggested that there may be an interspecific influence of red on Andersen, R., Duncan, P., Linnel, J.D.C., 1998. The European Roe Deer:
roe deer. In the present study, red and roe deer pellet group Biology of Success. Scandinavian University Press, Oslo.
densities were not negatively correlated, either on the study Aulak, W., Babińska-Werka, J., 1990a. Preference of different habitats and age
classes of forest by roe deer. Acta Theriol. 35, 289–298.
area or the habitat scale. Nevertheless, there may be negative
Aulak, W., Babińska-Werka, J., 1990b. Estimation of roe deer density based on
effect of red on roe deer performance (body mass, survival and the abundance and rate of disappearance of their faeces in the forest. Acta
reproduction), even with similar habitat preferences. As a Theriol. 35, 111–120.
result, temporal exclusion or declining body mass of roe deer Behrend, D.F., Lubeck, R.A., 1968. Summer flight behaviour of white-tailed
could have occurred. During last decade however, no decline in deer in two Adirondack forests. J. For. 48, 118–126.
roe deer population number (Nasiadka, unpubl. data) or body Bonenfant, C., Leif, E.L., Mysterud, A., Langvatn, R., Stenseth, N.C., Gaillard,
J.-M., Klein, F., 2004. Multiple causes of sexual segregation in European red
mass (Borkowski, in prep.) has been observed. In addition, deer: enlightenments from varying breeding phenology at high and low
transect section residuals from the fitted models were positively latitude. Proc. Roy. Soc. Lond. B 271, 883–892.
correlated for red and roe deer. This suggests that untested Borkowski, J., 2001. Flight behaviour and observability in human-disturbed
factors were affecting both species in the same way. In sika deer. Acta Theriol. 46, 195–206.
combination with similar habitat preferences, these results Borkowski, J., 2003. Influence of forest age on fallow deer food availability.
Sylwan 1, 88–95 (in polish with English summary).
suggest that interspecific competition was not an important Borkowski, J., 2004. Distribution and habitat use by red and roe deer following a
factor in this study. Nevertheless, the relations between the two large forest fire in South-western Poland. For. Ecol. Manage. 201, 287–293.
main and largely sympatric European deer species should Borkowski, J., Furubayashi, K., 1998. Home range size and habitat use in radio-
receive more attention from the researchers. collared female sika deer at high altitudes of Tanzawa Muntains. Jpn. Ann.
Zool. Fenn. 35, 181–186.
Borowski, S., Kossak, S., 1975. The food habits of deer in the Bialowieża
7. Conclusions primeval forest. Acta Theriol. 20, 463–506.
Breslow, N.E., Clayton, D.G., 1993. Approximate inference in generalized
In hunted areas, security cover is an important factor linear mixed models. J. Am. Stat. Assoc. 88, 9–25.
determining habitat use by deer, even though large predators are Catt, D.C., Staines, B.W., 1987. Home range use and habitat selection by red
absent. This study suggests that food also plays an important deer (Cervus elaphus) in a Sitka spruce plantation as determined by radio-
tracking. J. Zool. Lond. 211, 681–693.
role, but only where cover requirements are already met. Where Clutton-Brock, T.H., Guinness, F.E., Albon, S.D., 1982. Red Deer: Behaviour
managers are interested in high densities of deer, it is important and Ecology of Two Sexes. University of Chicago Press, Chicago.
to make sure that security cover is available. This is especially Collins, W.B., Urness, P.J., 1981. Habitat preferences of mule deer as rated by
true in habitats with abundant food and probably in areas where pellet-group distributions. J. Wildl. Manage. 45, 969–972.
Cook, J.G., Irwin, L.L., Bryant, L.D., Riggs, R.A., Thomas, J.W., 1998.
people hunt in groups using beaters. Under such circumstances,
Relations of forest cover and condition of elk: a test of the thermal cover
dense cover not only impedes shooting, but also allows some hypothesis in summer and winter. Wildl. Monogr. 141 (1), 1–61.
deer to remain undriven. This issue should be considered, Dzie˛ciołowski, R., 1969. The Quantity, Quality and Seasonal Variation of Food
especially in regions inhabited by larger deer species, since Resources Available to Red Deer in Various Environmental Conditions of
cover requirements in smaller deer seem to be less pronounced. Forest Management. Forest Research Institute, Warsaw.
Mature pine forests can serve as high-quality winter habitat Ge˛bczyńska, Z., 1980. Food of the roe deer and red deer in the Białowieża
primeval forest. Acta Theriol. 25, 487–500.
for deer, but only when they offer abundant understories. The Gill, R.M.A., 1992. A review of damage by mammals in north temperate
precise species composition of these understories is less forests. 1. Deer. Forestry 65, 145–169.
important in central Europe. The introduction of such Henry, B.A.M., 1981. Dstribution patterns of roe deer (Capreolus capreolus)
understories by foresters would thus seem important, not only related to the availability of food and cover. J. Zool. Lond. 194, 271–275.
for general biocenotic reasons, but also on account of the ‘‘side Hofmann, R.R., 1985. Digestive physiology of the deer—their morphological
specialisation and adaptation. In: Fennessy, P.F., Drew, K.R. (Eds.), Biol-
effect’’ of improved deer habitat. Such work can be seen as an ogy of Deer Production. Royal Society of New Zeland, pp. 393–407.
integrated forest and game management practice worthy of Holand, Ø., Mysterud, A., Wannag, A., Linnell, J.D.C., 1998. Roe deer in
promotion. However, in areas where large deer species cause northern environments: physiology and behaviour. In: Andersen, R.,
J. Borkowski, J. Ukalska / Forest Ecology and Management 255 (2008) 468–475 475

Duncan, P., Linnell, J.D.C. (Eds.), The European Roe Deer: The Bioogy Mysterud, A., Østbye, E., 1999. Cover as a habitat element for temperate
of Success. Scandinavian University Press, Oslo, pp. 117–137. ungulates: effects on habitat selection and demography. Wildl. Soc. Bull. 27,
Jeppesen, J.L., 1987. Impact of human disturbance on home range, movements 385–394.
and activity of red deer (Cervus elaphus) in a Danish environment. Dan. Neff, D.J., 1968. The pellet-group count technique for big game trend, census,
Rev. Game Biol. 13, 1–38. and distribution: a review. J. Wildl. Manage. 32, 597–614.
Kramer, C.Y., 1956. Extension of multiple range tests to group means with Palmer, S.C.F., Truscott, A.M., 2003. Seasonal habitat use and browsing by deer
unequal numbers of replications. Biometrics 12, 309–310. in Caledonian pinewoods. For. Ecol. Manage. 174, 149–166.
Kufeld, R.C., Bowden, D.C., Scrupp, D.L., 1988. Influence of hunting on Peek, J.M., Scott, M.D., Nelson, L.J., Pierce, J.D., 1982. Role of cover in habitat
movements of female mule deer. J. Range Manage. 41, 70–72. management for big game in northwestern United States. Trans. North Am.
Latham, J., Staines, B.W., Gorman, M.L., 1996. The relative densities of red Wildl. Nat. Res. Conf. 47, 363–373.
(Cervus elaphus) and roe (Capreolus capreolus) deer and their relationship Putman, R., 1988. The Natural History of Deer. Christopher Helm, London.
in Scottish plantation forest. J. Zool. Lond. 240, 285–299. Robakowski, P., Wyka, T., Samardakiewicz, S., Kierzkowski, D., 2004. Growth,
Latham, J., Staines, B.W., Gorman, M.L., 1997. Correlations of red (Cervus photosynthesis, and needle structure of silver fir (Abies alba Mill.) seedlings
elaphus) and roe (Capreolus capreolus) deer densities in Scottish forests under different canopies. For. Ecol. Manage. 201, 211–227.
with environmental variables. J. Zool. Lond. 242, 681–704. SAS Institute, 2002. SAS 9.1.3. SAS Institute Inc., Cary.
Latham, J., Staines, B.W., Gorman, M.L., 1999. Comparative feeding ecology Schultz, R.D., Bailey, J.A., 1978. Responses of national park elk to human
of red (Cervus elaphus) and roe deer (Capreolus capreolus) in Scottish activity. J. Wildl. Manage. 42, 91–100.
plantation forest. J. Zool. Lond. 247, 409–418. Skolvin, J.M., 1982. Habitat requirements and evaluation. In: Thomas, J.W.,
Lehmkuhl, J.F., Hansen, C.A., Sloan, K., 1994. Elk pellet-group decomposition Toweill, D.E. (Eds.), Elk of North America: Ecology and Management.
and detectability in coastal forests of Washington. J. Wildl. Manage. 58, Stackpole Books, Harrisburg, pp. 219–278.
664–669. Staines, B.W., Welch, D., 1984. Habitat selection of red (Cervus elaphus L.) and
Littel, R.C., Milliken, G.A., Stroup, W.W., Wolfinger, R.D., 1996. SAS System Roe (Capreolus capreolus) deer in a Sitka spruce plantation. Proc. Roy. Soc.
for Mixed Models. SAS Institute Inc., Cary. Edinb. 82B, 303–319.
Lovaas, A.L., 1958. Mule deer food habits and range use, Little Belt Mtns., Staines, B.W., Welch, D., Catt, D.C., Hinge, M.D.C., 1985. Habitat use and
Montana. J. Wildl. Manage. 22, 275–283. feeding by deer in Sitka spruce plantations. Institute of Terrestrial Ecology
Lyon, L.J., Jensen, C.E., 1980. Management implications of elk and deer use of Annual Report. ITE, Abbots Ripton, pp. 12–16.
clear-cuts in Montana. J. Wildl. Manage. 44, 352–362. Thirgood, S.J., Staines, B.W., 1989. Summer use of young stands of restocked
Massei, G., Bacon, P., Genov, P.V., 1998. Fallow deer and wild boar pellet group sitka spruce by red and roe deer. Scot. For. 43, 183–191.
disappearance in a Mediterranean area. J. Wildl. Manage. 62, 1086–1094. Tufto, J., Andersen, R., Linnel, J., 1996. Habitat use and ecological correlates
Mayle, B.A., Peace, A.J., Gill, R.M.A., 1999. How Many Deer? A Field Guide of home range size in a small cervid: the roe deer. J. Anim. Ecol. 65, 715–
to Estimating Deer Population Size. Forestry Commission, Edinburgh. 724.
Moe, S.R., Wegge, P., 1994. Spacing behaviour and habitat use of axis deer UN-ECE/FAO, 1992. The Forest Resources of the Temperate Zone. General
(Axis axis) in lowland Nepal. Can. J. Zool. 72, 1735–1744. Forest Resource Information, vol. 1. United Nations, New York.
Mysterud, A., Østbye, E., 1995. Bed-site selection by European roe deer Welch, D., Staines, B.W., Catt, D.C., Scott, D., 1990. Habitat usage by red
(Capreolus capreolus) in southern Norway during winter. Can. J. Zool. (Cervus elaphus) and roe (Capreolus capreolus) deer in a Scottish Sitka
73, 924–932. spruce plantation. J. Zool. Lond. 221, 453–476.

You might also like