Printer Friendly

Seasonal dynamics of forage for red deer in temperate forests: importance of the habitat properties, stand development stage and overstorey dynamics.

Populations of wild ungulates have increased rapidly over last decades in many European countries (Milner et al. 2006, Putman et al. 2011). For example in Slovakia, European red deer Cervus elaphus increased by 95% between 2000-2015 resulting in a population size that may be unprecedented for the past several centuries (Konopka et al. 2014, Pajtik et al. 2015). Commercialization of hunting (aimed at hunting for large trophies) during 1990s caused an expansion of the excessive feeding and male-biased harvest leading to skewed sex structure and predominance of females, resulting in rapid increase of the population. Currently, the red deer population continues to grow despite a threefold increase in harvest over the period 2000-2015 (Green Report 2016), which indicates an underestimation of population size and/or underharvesting, particularly of females (Clutton-Brock et al. 2004, Milner et al. 2006). Consequently, the increased impact on commercial forests is causing an extensive economic loss by reducing tree growth and lowering timber quality (Edenius et al. 2013, Manson et al. 2015) and by decreasing productivity of agricultural crops (Konopka et al. 2014). This situation has led to an emergence of tensions between groups with interests in forestry and agriculture on the one hand, and in wildlife management on the other hand. In the long-term, ungulates are pronounced ecosystem engineers and when at high densities they can trigger structural changes in plant species composition, soil properties and carbon cycle which are cascading to other species such as avian communities and forest-dwelling insects (Cote et al. 2004, Baiser et al. 2008). At the same time, density-dependent food limitation is the main driving factor in population dynamics of ungulates (Lone et al. 2014). Thus, quantifying the forage biomass and understanding of its temporal dynamics within different habitats is relevant to current efforts aimed at mitigating undesired long-term consequences of high deer densities on forestry, agriculture and nature conservation.

Hunting systems in many central and eastern European countries are based on detailed management plans (Milner et al. 2006). The system obliges holders of hunting rights (hunting clubs, private and state forests) to provide annual counts, detailed harvest plans and reports for an age- and sex-structured cull. The major responsibility for hunting-right holders is to keep deer numbers (represented by spring counts) close to normative numbers given by the state administration, which are based on foraging quality of the hunting ground. However, traditionally in Slovakia, the quality of hunting grounds reflecting forage potential for red deer is indexed solely according to the forest vegetation association (hereafter FVA; Zlatnik 1959) and adjusted by the extent of surrounding pastures (Sebeii et al. 2011). The FVAs are phytocenological units determined by climate and soil conditions and are known to provide various amounts of forage for ungulates (Katreniak et al. 1993, Seben et al. 2011). Although FVAs reflect climate and soil conditions, the approach does not account for other factors known to affect understorey vegetation such as stand development stage (Jules et al. 2008, Visscher and Merrill 2008) and structure of the overstorey (Hart and Chen 2006, Ujhazy et al. 2017).

In temperate closed-forest ecosystems, the amount of forage at a fine scale is the result of the joint effects of several environmental factors driving vegetation structure at several spatial and temporal scales (Owen-Smith et al. 2010). For example at the deer's seasonal home-range scale (5-10 [km.sup.2]; Kropil et al. 2015), the climate forms several forest vegetation belts along the altitudinal gradient. Further, at the scale of the daily range (0.9-1.3 [km.sup.2]; Kamler et al. 2007), a combination of forest vegetation belts with different soil types drive formation of the FVAs. Finally, at the scale of the foraging area (0.1-10 ha; Beguin et al. 2016), light availability is the major driving factor of vegetation growth, shaping architecture and vertical diversity of habitat patches (Kucbel et al. 2010, Drossler et al. 2016). Besides occasional disturbances such as wind storms, bark-beetle outbreaks and/or fires, forestry management apparently is the major disturbance factor driving ground vegetation development in commercial forests by manipulating with light levels (Weisberg et al. 2003, Jules et al. 2008). The intensity of understorey growth is usually driven by the magnitude of canopy disturbance caused by the particular type of management intervention (Ujhazy et al. 2017). Whereas thinning and selection cutting create only small openings in the canopy (several [m.sup.2]) and stimulate understorey vegetation only for a short period (several years), final felling of mature stands under shelterwood system (or clearcutting in some cases) creates large openings in forest stands ([less than or equal to] 3 ha; hereafter cutblocks) switching the forest stand to initiation stage (Oliver and Larson 1996). Vegetation responds to such disturbance by initiating secondary succession significantly increasing availability of forage for ungulates for several decades (Visscher and Merrill 2008). Moreover, crowns of tall trees that regularly become reachable after winter logging provide large amounts of potential forage for ungulates (Manson et al. 2015), considerably increasing the ungulate forage base (Edenius et al. 2013).

Many studies from temperate forests focus on evaluating understorey properties such as composition, species richness and diversity (Weisberg et al. 2003, Hart and Chen 2006, Jules et al. 2008, Ujhazy et al. 2017); however, only few try to predict the amount of forage available to wild ungulates seasonally. The goal of our study was to fill this gap by describing the dynamics of forage abundance available to red deer in commercial temperate forests and identify the major factors affecting forage availability. First, we compared differences in amount and composition of forage accessible to red deer (including bark of tree stems) among habitats in summer and winter. Second, we developed seasonal models to predict available forage biomass for the major habitat types using factors related to climate, soil conditions and light availability. We omitted bark of trees in estimates of forage availability for evaluating winter carrying capacity because excessive consumption of bark from tree stems has undesired long-term economic and ecological consequences. Third, we calculated amount of additional biomass produced by logging operations during winter and assessed herbivory impact on summer vegetation using data from inside and outside ungulate exclosures. Large-scale forage models are crucial for exploring mechanisms driving ungulate-vegetation interactions (Riggs et al. 2015, De Jager et al. 2017). Our study provides a long-term perspective on the forage availability to red deer in space and time for better integrating in forestry/agriculture and wildlife management, and the results may be applicable in countries with similar environmental conditions and hunting systems.

Material and methods

Study area

The study was conducted in temperate forests of the Kremnicke vrchy mountains (48[degrees]39'60"N, 19[degrees]00'00"E) located in central Slovakia (Fig. 1a). The area (ca 600 [km.sup.2]) is characteristic by a mountainous terrain within the altitudinal range of 400-1318 m determining two major climate zones used by red deer (Fig. 1; Kropil et al. 2015). First, the moderately warm and humid climate in lower parts of the area (< 700 m a.s.l.) with the average daily temperatures of 16-18[degrees]C in July and -3- -4[degrees]C in January, annual precipitation of 700-800 mm. Second, the moderately cool climate at higher altitudes (700-1100 m a.s.l.) with very humid climate and average daily temperatures of 12-14[degrees]C in July and -5- -6[degrees]C in January, annual precipitation of 900-1600 mm (Enviroportal SAZP 2017). Winter 2015/2016 was considerably mild with average snow cover 1 cm (max. 30 cm), thus we did not take snow into account in our study.

The area is mostly forested (450.0 [km.sup.2]; 76.5%) with 40% of deciduous forests, 39% of mixed forests and 21% of coniferous forests. In total, 15.0 [km.sup.2] of the forested area is in the initiation stage (3.3%). Pastures (105.7 [km.sup.2] = 18%) and crop fields (14.4 [km.sup.2] = 2.4%) are mainly concentrated at low altitudes in a warmer climate zone surrounding human settlements (18.4 [km.sup.2] = 3.1%). Forests are composed of European beech Fagus silvatica (37% of the forest area), oak Quercus spp. (19%) and the European hornbeam Caprinus betulus (10%) being a secondary successional species with a mixture of other valuable deciduous species for red deer such as maple Acer spp., ash Fraxinus excelsior and silver fir Abies alba (7%). Norway spruce Picea abies (10%) naturally occurs from ca 1000 m a.s.l. and dominates at 1200 m a.s.l. Forests in the Kremnicke vrchy Mts are commercially exploited; however, there are three nature preserves (Boky, Mlacik and Badin Primeval Forest) where hunting and forest management are banned.

According to management plans of all 22 hunting grounds in the area, the total estimated population of red deer, the main game species, was ca 1300 for the hunting season of 2014-2015, 1300 individuals of European roe deer Capreolus capreolus and 700 individuals of wild boar Sus scrofa. Hunters report stable populations of large carnivores such as brown bear Ursus arctos, grey wolf Canis lupus and Eurasian lynx Lynx lynx in the area.

Data collection

We followed previously used methods (Sachro et al. 2005, Hebblewhite and Merrill 2009) to sample the above ground vegetation biomass available to red deer during summer 2014 and winter 2016. We sampled 30 randomly selected plots throughout the altitudinal gradient of the Kremnicke vrchy Mts during summer and winter within each of the five major habitat types: deciduous, coniferous, and mixed forests, cutblocks and meadows/pastures. Within a geographic information system (GIS), locations of sample plots were selected randomly within the five habitats types in road-accessible areas (i.e. within 500 m) known to be used by red deer (Kropil et al. 2015). Crop fields were not sampled because they were known to have comparable amounts of biomass as pastures (Borowik et al. 2013), comprised only 2.4% of the study area, and were considered unimportant as habitat for red deer in this area (Kropil et al. 2015). When the GIS location of sample plot was not in the intended habitat type, the area was searched for a similar site within a radius of 200 m or the site was not used.

Vegetation was sampled in summer 2014 at the peak of vegetation biomass (June--August) and subsequently during winter 2016 (January--March) on the same locations. Vegetation was measured in a hierarchical manner (Fig. 1c) along 30-m transects laid initially perpendicular to the dominant slope or at a random direction in flat terrain in summer 2014 and in opposite directions in winter 2016. Three systematically selected subplots of 0.5 x 0.5 m (using a wooden frame) located at 5, 15 and 25 m were used to clip all available above-ground vegetation biomass within the red deer reachable height ([less than or equal to] 2 m). Because red deer is a typical intermediate feeder (Hofmann 1989) consuming a mixed diet composed of grasses and concentrate food items throughout the year (Gebert and Verheyden-Tixier 2001), we collected grasses, forbs and deciduous and/or coniferous shoots (< 3 mm in diameter) in both seasons. Ferns (Driopteris filix-mas and Athyrium filix femina) and subshrubs (Rubus spp.) were included in forbs, and shrubs (Rosa canina, Sambucus spp.) were included in deciduous trees. Further, we sampled 11 freshly created cutblocks (max. one month after logging) to add the contribution of the winter logging to the total forage base for red deer (Edenius et al. 2013, Manson et al. 2015). We also clipped biomass inside and outside four long-term exclosures (5 X 5 m) and used the percent difference in vegetation to assess herbivory impact on the ecosystem. All collected biomass was dried at 70[degrees]C for 48 h or until reaching dry weight and weighed to the nearest 0.1 g. For each sample plot, we calculated the mean dry biomass weight of vegetation in grams by averaging the weights of three sample subplots (0.75 [m.sup.2]). To calculate the dry collected biomass in grams per [m.sup.2], we multiplied dry biomass weight by the factor of 1.33 (and added dry bark biomass in winter).

Trees within the 'critical' diameter of 5-25 cm were counted on transect within a 30 x 10 m area (300 [m.sup.2]). Trees of this size are considered the most vulnerable to bark striping (Findo and Petras 2007). Because we were not permitted to remove bark from stems of economically valuable tree species, we estimated a tree's volume measured from tree height and diameter at breast height (1.3 m), then used stem volume relationships (in bark versus without bark) to estimate fresh bark biomass of the 1800 measured trees (Petras and Pajtik 1991). To calculate the proportion of the bark within reach of red deer ([less than or equal to]2 m height), we integrated tree morphological curves, which mathematically describe the relationship between tree height and its diameter for deciduous and coniferous (Fabrika and Pretzsch 2011), based on measured tree heights. Finally, we used species-specific coefficients to convert fresh bark volume into dry weight (Sebik 1989), summed bark estimates (g) for all trees counted in a plot, and divided by the plot size (300 [m.sup.2]) to obtain estimates of tree bark biomass (g [m.sup.-2]) accessible to red deer.

Environmental variables

Environmental and topographic characteristics that were used to predict seasonal forage biomass (Table 1) were measured either on site, derived from remote sensing data (NDVI), or values were taken from pre-existing raster GIS layers. Canopy closure was measured in the field at three subplots along the transect where biomass was sampled as the mean of the spherical densiometer reading and recorded as one of five classes: 1) 30-50% closure, 2) 60-70% closure, 3) 80% closure, 4) 90% closure and 5) 100% closure. All forest stands with closure < 30%, which is the canopy coverage of the last stage of the stand restoration in the shelterwood system, were considered already being at stand initiation stage. Ground cover of the new tree growth on cutblocks was estimated visually for the whole sample plot (300 [m.sup.2]) as mean of the three partial estimates (10 X 10 m). Age of a clear-cut was defined as time since harvest. We obtained seasonal indexes of forage stocks (IFS), which is a quantitative indicator of forage availability to red deer estimated for each FVA within the National Forest Inventory and Monitoring Programme 2005-2006 (Seben et al. 2011), from available GIS raster layers of FVAs. In summer models, we also used the normalised difference vegetation index (NDVI; pixels 250 x 250 m; Tucker 1979) centred on the site because NDVI is known to be correlated with green vegetation biomass in both open (positive relationship) and forested (negative relationship) habitats in temperate systems (Santin-Janin et al. 2009, Borowik et al. 2013). Further, average annual precipitation and average annual temperature (30-years averages for 1970-2000) were obtained from the WorldClim ver. 2.0 database at spatial resolution 1 [km.sup.2] (Fick and Hijmans 2017) and interpolated to spatial resolution 250X250 m using kriging in ArcMap 10.3 (ESRI 2014). Soil depth to bedrock was obtained from the SoilGrids ver. 0.5.1 database (ISRIC--World Soil Information; Shangguan et al. 2017) at spatial resolution 250 X 250 m and soil types (eutric, dystric and stagni-eutric cambisols; Table 1) were obtained from the Enviroportal database (Enviroportal SAZP 2017). The altitude was calculated using a digital elevation model with a pixel size of 30 m.

Statistical analyses

We used the t-test to test for differences in the amount of forage (g [m.sup.-2]) available in plots during summer and winter by habitat type. Normality of the biomass distribution was tested prior to the t-tests using Shapiro-Wilk test and the log-transformation was used when needed. We evaluated sets of candidate models for predicting forage biomass in both summer and winter based on environmental variables. We used a specific set of explanatory variables for each habitat class to focus the hypotheses tested. Collinearity of predictor variables was tested using Pearson's correlation, and the Akaike's information criterion corrected for small sample sizes ([AIC.sub.c], Anderson and Burnham 2002) was used to select models best predicting forage abundance in a habitat class. Subsequently, ANOVA was used to calculate the amount of the explained variation for each of the tested variables. The significance level of the explanatory variables was set at P [less than or equal to] 0.05. All statistical analyses were performed in the R program (ver. 3.2.5. <>).


In total, we sampled 320 vegetation plots throughout the Kremnicke vrchy Mts, 154 during the summer 2014 (including four exclosures) and 161 during the winter 2016 including 11 freshly felled plots. This represented about 4 plots per [km.sup.2] of the study area. Amount of forage for ungulates varied significantly between seasons (t = -7.43; p[less than or equal to] 0.001) and between habitats in both summer (F = 146.28; [less than or equal to]0.001) and winter (F = 94.74; p[less than or equal to]0.001). The habitat type in summer with the highest forage availability were cutblocks (250.3 [+ or -] 18.5 g [m.sup.-2], mean[+ or -]SE), followed by pastures (230.7 [+ or -] 21.3 g [m.sup.-2]), and both of these types were dominated by grasses (Table 2). Deciduous, coniferous and mixed forests did not differ in the total amount of forage (F = 1.794; p = 0.185) and provided on average only 22.7 [+ or -] 2.2 g [m.sup.-2] comprised primarily of shade tolerant forbs (Table 2). Based on total biomass inside (34.5 [+ or -] 8.3 g [m.sup.-2]) compared to outside (5.6 [+ or -] 1.5 g [m.sup.-2]) the four exclosures, herbivory reduced forage availability by 83.8%.

Available forage biomass decreased 65-98% across habitat types in winter (Fig. 2, Table 3). As in summer, cutblocks had the highest availability of forage (87.8 [+ or -] 11.6 g [m.sup.-2]), with forage consisting primarily of deciduous and coniferous vegetation. Moreover, an additional 36.7 [+ or -] 0.6 g [m.sup.-2] was available on freshly logged cutblocks. Forest types again did not differ in the amount of forage (F = 1.794; p = 0.185) and provided on average 4.6 [+ or -] 0.7 g [m.sup.-2] of mostly forb biomass (20.1% of the summer forage availability). The greatest decrease in biomass from summer (98%) occurred in pastures where biomass averaged only 2.45 [+ or -] 0.86 g [m.sup.-2], and 63% of the 35 sampled plots had minimal or no biomass. While farmed pastures that were mowed and grazed by cattle provided only 0.1 [+ or -] 0.02 g [m.sup.-2] of forage for red deer, the pastures without farming provided considerably more forage (13.6 [+ or -] 5.1 g [m.sup.-2]). Hence, we did not perform further analyses on winter pastures. Bark was a dominant forage component in maturing forests, comprising 50.0-81.5% of the total available forage biomass (Table 3).

We log-transformed forage biomass values prior to modelling forage--environment relationships because they were not normally distributed. We grouped the three forest types (deciduous, coniferous and mixed) into one forest class for further modelling because forage biomass did not differ between them in either season. Due to a high correlation (r > 0.7) between several variables (temperature x precipitation x elevation x soil depth) we did not use these variables in the same candidate model.

The best model for forested habitats indicated a negative effect of CCI on understorey biomass in both seasons (Table 5, Fig. 3a). CCI explained much of the understorey biomass in the closed-canopy forests in both summer (25.2% of 48% of variance in summer model) and in winter (22.4% of 46.5%, Table 5). We also found forage availability during summer was negatively related to NDVI (Fig. 3b), but not to IFS (t = 0.730, p = 0.468). In contrast, forage availability was positively related to IFS in winter, as well as to soil depth with higher productivity of stagnieutric cambisols (Table 5).

Amount of forage biomass in cutblocks was similarly related to stand age, canopy cover and elevation in both seasons, but soil type also explained forage abundance in winter (Table 5). After logging, summer forage biomass developed nonlinearly with the peak in summer biomass (forbs and grasses) at about ~8 years since timber harvest and winter biomass (mostly woody forage) developed somewhat slower peaking at about ~10 years, (Table 5, Fig. 4a) declining thereafter as canopy cover of newly forming stand increased. Coverage by a new tree growth had a positive relationship with forage biomass that was stronger in winter compared to summer (Table 5). Forage availability also was related to an altitudinal gradient over seasons. Total forage biomass was higher on cutblocks at low altitudes in summer, whereas it was higher at high altitudes during winter (Fig. 4b). In contrast to forests, winter forage biomass in cutblocks was more abundant on dystric and eutric cambisols compared to those associated with stagni-eutric cambisols (Table 5). IFS was not related to forage biomass on cutblocks in either season (summer t = 0.915, p = 0.368; winter t = 1.540, p = 0.137). Finally, forage biomass on summer pastures was positively related with NDVI (Fig. 3b), negatively related with increasing altitude, and those pastures located on eutric cambisols had higher biomass when compared to those on stagni-eutric cambisols (Table 5).


Our study shows significant spatial and temporal differences in forage availability for red deer between seasons and across habitats within seasons. Cutblocks were the most productive and diverse habitat type in both seasons producing 11 times more biomass in summer and 19 times more biomass in winter than mature forests (Table 2, 3.). A similar productivity in summer (245.7 g [m.sup.-2]) and winter (58.0-78.0 g [m.sup.-2]) was observed on cutblocks in the same area (Katreniak 1991, Garaj and Garaj 2007). Development of forage biomass on cutblocks exhibited the typical post-disturbance shape reported in other ecosystems (Visscher and Merrill 2008) with peak biomass reached by 8-10 years after logging (Fig. 4a). Even after 20 years clearcuts still provided 5 times more forage in summer and 8.5 times more forage for red deer in winter than in mature forests. Interestingly, we found higher forage biomass in summer on low-altitude cutblocks compared to those at high altitudes (threshold 700 m a.s.l.), whereas the opposite trend was true in winter (Fig. 4b). It is likely that because tree coverage was 14% lower on low-altitude cutblocks, growth of grasses (Carex spp., Calamagrostis arundinacea) was stimulated earlier and plants overall had higher light availability during summer. At the same time, the vegetative dieback in winter may have decreased the overall forage availability of these cutblocks. Regardless, we can conclude that areas where forest is at initiation stage have great importance in terms of the overall carrying capacity of the landscape.

Our study confirms that canopy cover plays a key role in understorey dynamics in closed-canopy temperate forests (Table 4, Fig. 3a). Development of the canopy limits the light transmission to lower layers, negatively affecting growth of ground vegetation (Weisberg et al. 2003, Hart and Chen 2006, Jules et al. 2008), especially on shallow soils. Our results also confirmed that NDVI can be used as a reliable proxy for summer forage biomass in closed-canopy forests and open pastures. That we found NDVI and ground vegetation biomass negatively related, whereas the same relationship on pastures was positive, is consistent with other studies (Pettorelli et al. 2005, Wessels et al. 2006, Borowik et al. 2013). Using the NDVI might be advantageous in the long-term, particularly for modelling effects of climate change on the spatial patterns of forage availability over time (Santin-Janin et al. 2009, Borowik et al. 2013).

New cutblocks freshly created during winter provided a considerable amount of additional forage from crowns of tall trees, which was about eight times that of forage available in mature forests. Such areas, when harvested in fall or early winter, can be used to increase ungulate use reducing thus browsing damage elsewhere (Edenius et al. 2013, Manson et al. 2015). Bark from trees is a common food source consumed by red deer in winter (Jamrozy 1980, Suter et al. 2004, Krojerova-Prokesova 2010); however it is important to differentiate between bark from tree shoots and bark from tree stems. While consumption of shoots is considered a 'bearable' damage, extensive consumption of bark from stems of economically valuable trees is unacceptable in forest management. Thus, the ecological carrying capacity of an ecosystem is much greater than it is economically or socially acceptable (Fig. 2; Arrow et al. 1995). We provide models that can be used to predict forage availability that indexes a 'bearable' level of browsing damage by ungulates, which would be more useful for practical wildlife management than prediction of total biomass.

Although light availability associated with succession-mediated changes in canopy cover appears to be a major mechanism in spatial variability in forage biomass, the 84% difference in forage biomass inside and outside exclosures points to a second mechanism: ungulate browsing could alter the spatial and temporal patterns in forage biomass, i.e. the 'ecosystem engineer' hypothesis. In other words, at high deer densities successional patterns in understory vegetation may be altered by the habitat and foraging preferences of deer. However, the small number of reference plots (i.e. four exclosures) limits our ability to incorporate the effects of herbivory into our statistical models (Table 5). Future studies that use deer density as an explanatory variable may improve our predictions of forage biomass across the landscape (De Jager et al. 2017). Indeed, long-term, excessive herbivory pressure may not only alter forage availability to deer, but change forest trajectories during succession altering forest composition and diversity with negative long-term ecological and economic consequences (Cote et al. 2004, Beguin et al. 2016). Hence, a pressing challenge wildlife and forestry managers are facing is to develop new management techniques aimed at successfully establishing forest regeneration under high deer populations (Beguin et al. 2016).

Management implications

Traditionally in Slovakia, the quality of hunting grounds reflecting forage potential for red deer is indexed solely according to the FVAs and adjusted by the extent of surrounding pastures. However, this approach does not take into account the developmental stage nor does the structure of the forest stands and thus lacks the predictive potential as shown in our study. We showed that from an ungulate perspective, cutblocks were the most productive habitats throughout the year irrespective of the FVA. Thus, all have a great importance and should be considered separately from mature forests when estimating the overall forage potential. Further, we showed that in mature closed-canopy forests, light is the major driving factor of understorey growth over site quality and should be taken into consideration when predicting forage biomass for ungulates. Farmed pastures provide very limited quantity of forage during winter compared to pastures without summer mowing or grazing, and thus farmed pastures should not be taken into consideration as winter forage resource. Our results suggest that temporal adjustments of logging towards winter might be especially advantageous because it is being conducted on a regular basis over large areas, hence creating potential synergies through integration of wildlife management into standard forestry (Edenius et al. 2013).

Gradual increase of red deer population in short period generated a substantial herbivory pressure on forest ecosystems resulting in considerable economic losses and potential conflicts between interest groups. However, to achieve the agreement between interest groups, a 'bearable' damage of commercial forests and agricultural crops needs to be discussed between foresters, farmers and wildlife managers (Pajtik et al. 2015). New approaches has been proposed using biological deterrents such as "hunting for fear" (Kuijper et al. 2010, Beguin et al. 2016). This concept involves increasing disturbance in regenerating forests through more aggressive hunting practices (encouraging harvest of females and juveniles) or/and through using top--down control of natural predators on ungulates (Ripple et al. 2003). The most effective solution in a long-term however, would be the development of an integrated adaptive management system of forestry, agriculture, wildlife and conservation management where all involved management practices are interconnected and mutually coordinated (Beguin et al. 2016).

Acknowledgements--We thank the University Forest Enterprise and the Urban Forests of Kremnica for the field assistance. We also thank the National Forest Centre for providing GIS data on FVAs. Special thanks to Dr. E. Merrill, Dr. E. Krizova, Dr. M. Fabrika, Dr. J. Vido and Dr. M. Kodrik for their advices and to P. Gibas and L Ferlica for their dedicated field assistance.

Funding--This paper was produced within the project APVV-14-0637 supported by the Ministry of Education, Science, Research and Sport of the Slovak Republic and projects ITMS 26220120006 and ITMS 26220120049 supported by the Operational Programme Research and Development within the European Regional Development Fund.


Anderson, D. R. and Burnham, K. P. 2002. Avoiding pitfalls when using information-theoretic methods.--J. Wildl. Manage. 66: 912-918.

Arrow, K. et al. 1995. Economic growth, carrying capacity and the environment.--Ecol. Econ. 15: 91-95.

Baiser, B. et al. 2008. A perfect storm: two ecosystem engineers interact to degrade deciduous forests of New Jersey.--Biol. Invas. 10: 785-795.

Beguin, J. et al. 2016. Management of forest regeneration in boreal and temperate deer--forest systems: challenges, guidelines and research gaps.--Ecosphere 7: 1-16.

Borowik, T. et al. 2013. Normalized difference vegetation index (NDVI) as a predictor of forage availability for ungulates in forest and field habitats.--Eur. J. Wildl. Res. 59: 675-682.

Clutton-Brock, T. H. et al. 2004. Red deer stocks in the Highlands of Scotland.--Nature 429: 261-262.

Cote, S. D. et al. 2004. Ecological impacts of deer overabundance. --Annu. Rev. Ecol. Evol. Syst. 35: 113-147.

De Jager, N. R. et al. 2017. Simulating ungulate herbivory across forest landscapes: a browsing extension for LANDIS-II.--Ecol. Model. 350: 11-29.

Drossler, L. et al. 2016. What happens after the gap?--Size distributions of patches with homogenously sized trees in natural and managed beech forests in Europe.--Open J. For. 6: 177-190.

Edenius, L. et al. 2013. Ungulate-adapted forest management: effects of slash treatment at harvest on forage availability and use.--Eur. J. For. Res. 133: 191-198.

Enviroportal SAZP 2017. Atlas Krajiny Slovenskej Republiky Slovesnka Agentura Zivotneho Prostredia < >.

ESRI 2014. ArcGIS Desktop: rel. 10.3.--Environmental Systems Res. Inst.

Fabrika, M. and Pretzsch, H. 2011. Analyza a modelovanie lesnych ekosystemov.--Technicka Univerzita vo Zvolene, Slovakia, in Slovak.

Fick, S. E. and Hijmans, R. J. 2017. WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas.--Int. J. Climatol. 37: 4302-4315.

Findo, S. and Petras, R. 2007. Ekologicke zaklady ochrany lesa proti poskodzovaniu zverou.--Narodne lesnicke centrum Zvolen, Slovakia, in Slovak.

Garaj, P. jr. and Garaj, P. 2007. Trophic availability of dendromass of beech communities in non-vegetation period for ruminant game in relation to silvicultural system.--Folia Venatoria 36-37:7-18, in Slovak.

Gebert, C. and Verheyden-Tixier, H. 2001. Variations of diet composition of red deer (Cervus elaphus) in Europe.--Mammal Rev. 31: 189-201.

Green Report 2016. Ministry of Agriculture and Rural Development of the Slovak Republic.

Hart, S. A. and Chen H. Y. H. 2006. Understory vegetation dynamics of North American boreal forests.--Crit. Rev. Plant Sci. 25: 381-397.

Hebblewhite, M. and Merrill, E. 2009. Tradeoffs between predation risk and forage differ between migrant strategies in a migratory ungulate.--Ecology 90: 3445-3454.

Hofmann, R. R. 1989. Evolutionary steps of ecophysiological adaptation and diversification of ruminants: a comparative view of their digestive system.--Oecologia 78: 443-457.

Jamrozy, G. 1980. Winter food resources and food preferences of red deer in Carpathian Forests.--Acta Theriol. 25: 221-238.

Jules, M. J. et al. 2008. Assessing the relationships between stand development and understory vegetation using a 420-year chronosequence.--For. Ecol. Manage. 255: 2384-2393.

Kamler, J. et al. 2007. Factors affecting daily ranges of red deer Cervus elaphus in Bialowieza Primeval Forest, Poland.--Acta Theriol. 52: 113-118.

Katreniak, J. 1991. Supply of food for ruminating game in the summer and autumn period in the first to fourth forest vegetation belt.--Folia Venatoria 21: 21-34, in Slovak.

Katreniak, J. et al. 1993. Biologicke zaklady rajonizacie, bonitacie a planovania v polovnictve.--Lesnicky vyskumny ustav, Zvolen, Slovakia, in Slovak.

Konopka, J. et al. 2014. Koncepcia rozvoja polovnictva v Slovenskej Republike. Polovnicke studie 12. Narodne lesnicke centrum --Lesnicky vyskumny ustav, Zvolen, Slovakia, in Slovak.

Krojerova-Prokesova, J. 2010. Feeding patterns of red deer in the Bohemian Forest: effect of habitat and season.--Wildl. Biol. 16: 173-184.

Kropil, R. et al. 2015. Home range and migration patterns of male red deer Cervus elaphus in western Carpathians.--Eur. J. Wildl. Res. 61: 63-72.

Kucbel, S. et al. 2010. Canopy gaps in an old-growth fir-beech forest remnant of western Carpathians.--Eur. J. For. Res. 129: 249-259.

Kuijper, D. P. J. et al. 2010. Bottom--up versus top--down control of tree regeneration in the Bialowieza Primeval Forest Poland. --J. Ecol. 98: 888-899.

Lone, K. et al. 2014. Improving broad scale forage mapping and habitat selection analyses with airborne laser scanning: the case of moose.--Ecosphere 5: 144.

Manson, J. et al. 2015. Food plots as a habitat management tool: forage production and ungulate browsing in adjacent forest. --Wildl. Biol. 21: 246-253.

Milner, J. M. et al. 2006. Temporal and spatial development of red deer harvesting in Europe--biological and cultural factors. --J. Appl. Ecol. 43:721-734.

Oliver, C. D. and Larson, B. C. 1996. Forest stand dynamics College of Forest Resources, Univ. of Washington.--Wiley.

Owen-Smith, N. et al. 2010. Foraging theory upscaled: the behavioural ecology of herbivore movement.--Phil. Trans. R. Soc. B 365: 2267-2278.

Pajtik, J. et al. 2015. Modelling forage potential for red deer: a case study in post-disturbance young stands of rowan.--Ann. For. Res. 58: 1-17.

Petras, R. and Pajtik, J. 1991. Sustava cesko-slovenskych objemovych tabuliek drevin.--For. J. 37: 49-56, in Slovak.

Pettorelli, N. et al. 2005. Using the satellite-derived normalized difference vegetation index (NDVI) to assess ecological effects of environmental change.--Trends Ecol. Evol. 20: 503-510.

Putman, R. et al. 2011. Ungulate management in Europe: problems and practices.--Cambridge Univ. Press.

Riggs, R. A. et al. 2015. Biomass and fire dynamics in a temperate forest-grassland mosaic: Integrating multi-species herbivory, climate, and fire with the FireBGCv2/GrazeBGC system. --Ecol. Model. 296: 57-78.

Ripple, W J. et al. 2003. Trophic cascades among wolves, elk and aspen on Yellowstone National Park's northern range.--Biol. Conserv. 102: 227-234.

Sachro, L. L. et al. 2005. Prescribed burning effects on summer elk forage availability in the subalpine zone, Banff National Park, Canada.--J. Environ. Manage. 77: 183-193.

Santin-Janin, H. et al. 2009. Assessing the performance of NDVI as a proxy for plant biomass using non-linear models: a case study on the Kerguelen archipelago.--Polar Biol. 32: 861-871.

Seben, V. et al. 2011. Sources of food for game, according to National forest inventory and monitoring of the Slovak Republic 2005-2006.--Folia Venatoria 40-41:29-52, in Slovak.

Sebik, L. 1989. Nauka o produkcii. - Technicka Univerzita vo Zvolene, Slovakia, in Slovak.

Shangguan, W. et al. 2017. Mapping the global depth to bedrock for land surface modelling.--J. Adv. Model. Earth Syst. 9: 65-88.

Suter, W. et al. 2004. Spatial variation of summer diet of red deer Cervus elaphus in the eastern Swiss Alps.--Wildl. Biol. 10: 43-50.

Tucker, C. J. 1979. Red and photographic infrared linear combinations for monitoring vegetation.--Remote Sens. Environ. 8: 127-150.

Ujhazy, K. et al. 2017. Overstorey dynamics controls plant diversity in age-class temperate forests.--For. Ecol. Manage. 391: 96-105.

Visscher, D. R. and Merrill, E. H. 2008. Temporal dynamics of forage succession for elk at two scales: implications of forest management.--For. Ecol. Manage. 257: 96-106.

Weisberg, P. J. et al. 2003. Predicting understorey vegetation cover from overstorey attributes in two temperate mountain forests. --For. Res. 122: 273-286.

Wessels, K. J. et al. 2006. Relationship between herbaceous biomass and 1-[km.sup.2] advanced very high radiometer (AVHRR) NDVI in Kruger National Park, South Africa.--Int. J. Remote Sens. 27: 951-973.

Zlatnik, A. 1959. Overview of Slovak forests by forest vegetation associations. - LF VSZ Brno, Czech rep., in Czech.

Peter Smolko, Alexandra Veselovska and Rudolf Kropil

P. Smolko (, A. Veselovska and R. Kropil, Dept of Applied Zoology and Wildlife Management, Faculty of Forestry, Technical Univ. in Zvolen, T. G. Masaryka 20, SK-960 53 Zvolen, Slovakia. PS also at: Diana--Carpathian Wildlife Research, Banska Bystrica, Slovakia.

This work is licensed under the terms of a Creative Commons Attribution 4.0 International License (CC-BY) <>. The license permits use, distribution and reproduction in any medium, provided the original work is properly cited.

doi: 10.2981/wlb.00366
Table 1. Definitions of variables used in models for predicting
seasonally available forage biomass.

Variable   Range of values  Unit        Description

CCI        1-5              -           canopy closure index
IFS        0.7-2.0          -           index of forage stocks
NDVI       0.1-1.0          -           normalised difference
                                        vegetation index
AGE        1-20             years       time since final felling
COVER      10-100           %           ground cover of the thicket
                                        within a cutblock
ELEV       488-1031         m a.s.l.    altitude
PRECIP     673-1129         mm year-1   annual rainfall
AVG_TEMP   3.1-8.6          [degrees]C  average annual temperature
SOIL_DPTH  748-1992         cm          absolute depth to bedrock
SOIL_TP    1                -           soil type 1 - Eutric Cambisols
           2                -           soil type 2 - Dystric Cambisols
           3                -           soil type 3 - Stagni-Eutric

Table 2. Abundance and composition of the summer vegetation biomass (g
[m.sup.-2]) available to red deer Cervus elaphus in different habitats
throughout the Kremnicke vrchy Mts, Slovakia.

Habitat     Grasses                Forbs

Deciduous     4.25 [+ or -] 1.56    9.34 [+ or -] 2.25
Coniferous    2.08 [+ or -] 1.95   14.80 [+ or -] 2.26
Mixed         0.48 [+ or -] 0.31   17.04 [+ or -] 2.80
Cutblock     66.64 [+ or -] 10.79  68.32 [+ or -] 8.05
Pasture     153.97 [+ or -] 15.97  76.73 [+ or -] 10.61

Habitat     Deciduous             Coniferous

Deciduous    8.26 [+ or -] 2.52    0.03 [+ or -] 0.02
Coniferous   4.50 [+ or -] 2.07    4.49 [+ or -] 2.48
Mixed        4.88 [+ or -] 1.77    0.63 [+ or -] 0.49
Cutblock    92.24 [+ or -] 19.86  23.10 [+ or -] 10.48
Pasture      -                    -

Table 3. Winter vegetation biomass (g [m.sup.-2]) by forage class in
different habitats available to red deer Cervus elaphus throughout the
Kremnicke vrchy Mts, Slovakia.

Habitat     Grasses             Forbs                Deciduous

Deciduous   0.46 [+ or -] 0.33  0.81 [+ or -] 0.22   3.10 [+ or -] 1.48
Coniferous  0.08 [+ or -] 0.05  1.47 [+ or -] 0.36   1.61 [+ or -] 1.06
Mixed       0.76 [+ or -] 0.61  0.78 [+ or -] 0.25   1.74 [+ or -] 0.72
Cutblock    0.98 [+ or -] 0.35  7.57 [+ or -] 1.38  52.18 [+ or -] 10.34
Pasture     0.96 [+ or -] 0.52  1.49 [+ or -] 0.69   -

Habitat      Coniferous          Bark

Deciduous    0.02 [+ or -] 0.01  19.31 [+ or -] 6.69
Coniferous   2.10 [+ or -] 1.42   5.26 [+ or -] 1.75
Mixed        0.35 [+ or -] 0.30  14.96 [+ or -] 4.79
Cutblock    30.30 [+ or -] 8.50   7.89 [+ or -] 3.74
Pasture      -                    -

Table 4. Summary of model selection results for candidate models
predicting forage biomass (g [m.sup.-2]) in selected habitat types in
different seasons based on explanatory variables from Table 1.
Presented are number of model variables (k), log-likelihood (LL),
change in [AIC.sub.c] from the top model ([DELTA][AIC.sub.c]), and
Akaike's weights (w).

Model                                    k   LL      [DELTA][AIC.sub.c]

CCI + NDVI                               4  -23.47   0.00
IFS + CCI + NDVI + AVG_TEMP + SOIL_TP    8  -21.97   6.29
IFS + CCI + NDVI + PRECIP + SOIL_TP      8  -22.39   7.14
IFS + CCI + NDVI + SOIL_DPTH + SOIL_TP   8  -22.21   6.77
NULL                                     2  -46.52  41.76
IFS + CCI + SOIL_DPTH + SOIL_TP          7  -60.11   0.00
IFS + CCI + PRECIP + SOIL_TP             7  -65.26  10.32
IFS + CCI + AVG_TEMP + SOIL_TP           7  -66.27  12.34
NULL                                     2  -86.68  41.92
AGE + [AGE.sup.2] + [AGE.sup.3] +
COVER + ELEV                             7    9.74   0.00
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + ELEV + SOIL_TP                10   11.15   9.67
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + AVG_TEMP + SOIL_TP            10   11.03   9.91
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + PRECIP + SOIL_TP              10   10.89  10.19
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + SOIL_DPTH + SOIL_TP           10    9.13  13.71
NULL                                     2   -4.92  14.68
AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + ELEV + SOIL_TP                 9   13.27   0.00
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + ELEV + SOIL_TP                10   11.96   5.20
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + PRECIP + SOIL_TP              10    9.18  10.75
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + AVG_TEMP + SOIL_TP            10    6.76  13.59
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + SOIL_DPTH + SOIL_TP           10    7.19  14.73
NULL                                     2  -13.33  28.65
NDVI + ELEV + SOIL_TP                    5   18.16   0.00
NDVI + SOIL_DPTH + SOIL_TP               5   16.14   4.02
NDVI + AVG_TEMP + SOIL_TP                5   16.14   4.03
NDVI + PRECIP + SOIL_TP                  5   16.09   4.13
NULL                                     2    5.17  17.91

Model                                   w

CCI + NDVI                              0.90
IFS + CCI + NDVI + PRECIP + SOIL_TP     0.03
NULL                                    0.00
IFS + CCI + SOIL_DPTH + SOIL_TP         0.99
IFS + CCI + PRECIP + SOIL_TP            0.01
IFS + CCI + AVG_TEMP + SOIL_TP          0.00
NULL                                    0.00
AGE + [AGE.sup.2] + [AGE.sup.3] +
COVER + ELEV                            0.98
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + ELEV + SOIL_TP                0.01
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + AVG_TEMP + SOIL_TP            0.01
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + PRECIP + SOIL_TP              0.01
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + SOIL_DPTH + SOIL_TP           0.00
NULL                                    0.00
AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + ELEV + SOIL_TP                0.93
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + ELEV + SOIL_TP                0.07
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + PRECIP + SOIL_TP              0.00
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + AVG_TEMP + SOIL_TP            0.00
IFS + AGE + [AGE.sup.2] + [AGE.sup.3]
+ COVER + SOIL_DPTH + SOIL_TP           0.00
NULL                                    0.00
NDVI + ELEV + SOIL_TP                   0.72
NDVI + SOIL_DPTH + SOIL_TP              0.10
NDVI + AVG_TEMP + SOIL_TP               0.10
NDVI + PRECIP + SOIL_TP                 0.09
NULL                                    0.00

Table 5. Parameter estimates (p [+ or -] SE) for linear models
explaining seasonal forage availability for red deer Cervus elaphus in
the Kremnicke vrchy Mts, western Carpathians. Presented are t and
p-values of the tests, total variation explained by the top candidate
models (in bold), and partial variations explained by particular
variable ([R.sup.2]).

Variables        [beta]     SE                 t       p       [R.sup.2]

Forest/summer                                                  47.8
Intercept         9.3678    1.3348              7.018  <0.000
CCI              -0.2305    0.0377             -6.102  <0.001  25.2
NDVI             -8.6555    1.5118             -5.725  <0.001  22.6
Forest/winter                                                  46.5
CCI              -0.2412    0.0424             -5.684  <0.001  22.4
IFS               3.8488    1.2285              3.133   0.002   8.3
SOIL_DPTH         0.0014    0.0003              4.176  <0.001   9.9
SOIL_TP1         -5.0372    1.4850             -1.907   0.060   5.9
SOIL_TP2         -4.9545    0.1385              0.597   0.552
SOIL_TP3         -4.1720    0.3018              2.867   0.005
Cutblock/summer                                                57.8
Intercept         1.2201    0.1027             10.984  <0.001
AGE               0.3452    0.0840              1.756   0.092  46.3
[AGE.sup.2]      -0.0270    0.0120              3.621   0.001
[AGE.sup.3]       0.0006    0.0002              3.964  <0.001
COVER             0.0034    0.0015              2.206   0.036   4.6
ELEV             -0.0006    0.0003             -2.138   0.041   6.9
Cutblock/winter                                                73.8
AGE               0.3047    0.2130              0.041   0.968  41.3
[AGE.sup.2]      -0.0224    0.0068              3.701  <0.001
[AGE.sup.3]       0.0004    0.0001              2.093   0.046
COVER             0.0079    0.0020              3.617   0.002  22.1
ELEV              0.0012    0.0005              2.495   0.020   8.0
SOIL_TP1         -0.7599    0.8609             -0.883   0.387   2.4
SOIL_TP2         -0.8679    0.1165             -0.927   0.364
SOIL_TP3         -1.2851    0.2582             -2.033   0.050
Pasture/summer                                                 52.9
NDVI              1.1260    0.4995              2.922   0.007  29.2
ELEV              1.04E-05  6.00 X [10.sup.6]  -1.172   0.039  10.9
SOIL_TP1          1.4600    0.4998              4.922  <0.001  12.8
SOIL_TP3          0.9450    0.1056             -2.608   0.015
COPYRIGHT 2018 Nordic Board for Wildlife Research
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Smolko, Peter; Veselovska, Alexandra; Kropil, Rudolf
Publication:Wildlife Biology
Article Type:Report
Geographic Code:4EXSV
Date:Jan 1, 2018
Previous Article:Estimating sex-ratio, survival, and harvest susceptibility in greater sage-grouse: making the most of hunter harvests.
Next Article:Population estimation, distribution, and habitat preference of Irrawaddy dolphins Orcaella brevirostris (Owen in Gray, 1866) in the Brunei Bay,...

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters