Introduction

The group of several ant species commonly inhabiting forests in temperate and boreal regions called red wood ants (RWAs) is considered to be an especially important component of forest ecosystems. The species that belong to the RWA group are territorial and aggressive, and form large communities. It has been proven that RWAs have great influence on the distribution and abundance of other ant species, and many other epigeic and arboreal invertebrates (e.g. spiders, carabids; Hawes et al. 2002; Koivula and Niemelä 2003; Reznikova and Dorosheva 2004; Ohashi et al. 2007; but see also Lenoir et al. 2003). It was also observed that these ants may affect the behaviour of carabids, for example forcing them to change their foraging and moving strategies (e.g. Dorosheva and Reznikova 2006). Moreover, as RWAs significantly reduce the abundance of many invertebrate taxa, they may indirectly affect the breeding success of some insectivorous birds (Aho et al. 1999; Jantti et al. 2007). Therefore, it may be assumed that factors affecting the distribution and abundance of RWAs also affect the distribution of many other groups of animals. In the light of the great biocenotic importance of RWAs, knowledge concerning the variability of their densities is crucial for understanding the overall patterns of biodiversity distribution in forest ecosystems (Punttila et al. 1991).

Since only 7% of European forests is protected (Verkerk et al. 2008) and remaining woodland is managed for timber production, one might expect that the distribution and abundance of RWAs are driven by habitat transformations that are related to timber production (clear-cutting, artificial replanting etc.; Domisch et al. 2000; Sorvari and Hakkarainen 2005, 2007). The effect of forest management on RWAs may indirectly affect the spatial distribution of invertebrate communities that are sensitive to predation by and competition with the ants. However, little is known about response of RWAs to clear-cutting in Central European forests. In this study, the distribution of RWAs in clear-cut areas of managed forests was investigated and its ecological consequences are discussed.

Materials and methods

The study was conducted in the spring and summer of 2007, in a managed forest of Western Poland, Central Europe (52°48′N, 14°20′E). The study area was located in the temperate climate zone and its main potential vegetation type was broadleaved forest dominated by oaks (Quercus spp.), hornbeams (Carpinus betulus) and limes (Tilia spp.). The region is affected by mild and wet oceanic climate of Western Europe and by harsh and dry continental climate of Eastern Europe and Asia. The forest was ca. 14,000 ha in size and was managed by National Forest Holding for timber production. As a result of long-lasting forest management tree composition of the stand was dominated by Scots pine (Pinus sylvestris), with considerably lower proportion of oaks and birch (Betula pendula). The forest comprised semi-rich and poor habitats, with sandy soils and understorey vegetation of the forest floor dominated by mosses (Polytrichum spp.), grasses (Calamagrostis spp., Deshampsia flexuosa) and shrubs (Rubus spp., Vaccinium vitis-idaea). Plant coverage on clear-cut areas was composed mainly of Calamagrostis epigejos, Convolvulus arvensis, Rubus idaeus, Senecio vulgaris and Dryopteris carthusiana. Stand age was highly diversified and ranged from 0 years on fresh clear-cuts to 100–150 years in the oldest patches. On the basis of old maps of the region it can be concluded that this forest complex has been managed for timber production for at least 120 years and its area has retained the same shape until nowadays (Żmihorski, unpublished data).

Sixteen transects were selected in 4 clear-cut areas not older than 3 years (3–5 transects per clear-cut area, at least 30 m away from each other). The transects started at the edge of clear-cut area and led towards its centre. Each transect was ca. 30 m long and consisted of ca. 15 separate squares where RWAs were sampled. However, because of some local disturbances some squares were omitted. Squares were distributed evenly every 2 m along the transect; the last square (fifteenth) was 30 m away from the clear-cut area edge. The boundary of each square was delineated by a 1 × 1 m wooden frame. Individuals of RWA in each square were directly counted during 1–3 min. The RWA group in the study area was composed mainly of Formica rufa Linnaeus, F. sanguinea Latreille, F. exsecta Nylander and F. polyctena Förster (Włodarczyk et al. 2009; Włodarczyk, unpublished data). Great care was taken to avoid double counting. Moreover, for each square, plant coverage (0–100%) was visually assessed as a factor potentially affecting the ants. All counts were conducted in good weather conditions, mainly in afternoon. In total, RWAs were counted in 196 squares.

The number of RWAs in the sampled squares was not normally distributed, due to a considerable proportion of squares where no RWAs were recorded. Therefore, two statistical analyses were performed. First, mean distance to clear-cut edge was compared between squares with and without red wood ants by using t test. Second, for all squares with RWAs (n = 53) a univariate analysis of covariance was conducted, with clear-cut area as a random categorical factor, distance to the edge of clear-cut area and plant coverage as covariates, and log-transformed number of RWAs as the dependent variable. Variance homogeneity was checked with Levene’s test, which tests the null hypothesis that the error variance of the dependent variable is equal across groups. All statistical analyses were performed with SPSS 13.0 (SPSS 2004).

Results

Squares containing ants were found to be closer to clear-cut area edges (mean 10.57 m ± 7.24 SD; n = 53) than squares without ants (mean 17.70 m ± 8.62; n = 143), and the difference was statistically significant (t test, t 194 = 5.39, P < 0.0001). The abundance of RWAs differed significantly among four clear-cut areas (P = 0.005), whereas the effect of plant coverage was not statistically significant (P = 0.095, Table 1). The number of ants recorded in the squares was significantly affected by distance to clear-cut area edge (Table 1; Fig. 1). The latter relationship was best explained by an exponential function (Fig. 1).

Table 1 Results of univariate analysis of covariance testing of the effect of clear-cut area, distance to clear-cut area edge and plant coverage on density of red wood ants in 1-m2 squares in a managed forest in Western Poland
Full size table
Fig. 1
figure 1

Number of red wood ants (RWAs) in 1-m2 squares in managed forest clear-cut areas in Western Poland, plotted against distance to clear-cut area edge. Note that the abundance of RWAs is not adjusted for the clear-cut area effect (Table 1)

Full size image

Discussion

Because the presented results are based on ca. 200 squares distributed on 4 clear-cut areas, the results should be treated with caution. Also, the analysis concerns only the youngest clear-cut areas, whereas patterns of RWAs abundance and distribution in older ones may be different. Despite these shortcomings, the study generated significant and reliable results. It should be mentioned that in the study area forest management, habitat features, replanting techniques and clear-cut characteristics are typical of most managed forests in Poland. Therefore, it can be suspected that the obtained pattern of RWAs distribution in clear-cut areas is similar in other Polish managed forests. Since managed forests comprise nearly 80% of the overall forested area in Poland (Budna et al. 2006), the results can have wide application.

It was found that the density of RWAs decreased significantly from the edge of the clear-cut area toward its centre. This result seems to be consistent with other findings, which showed negative effects of anthropogenic changes of forest habitat on Formica species (Domisch et al. 2000; Sorvari and Hakkarainen 2005, 2007). As a consequence, the clear-cut area centre lacks RWAs. This pattern can be explained by several factors. First, abiotic conditions, such as insolation, ambient temperature amplitude, soil and air moisture are significantly more variable in clear-cut areas than in closed-canopy forest (e.g. Aussenac 2000). Because of these specific microclimate conditions, clear-cut areas may be avoided by most ant species that belong to the RWA group (Rosengren and Pamilo 1978; Domisch et al. 2000). Second, the activity of RWAs in clear-cut areas may be limited by food resources, for instance scarcity of aphids (Punttila et al. 1991; Sorvari and Hakkarainen 2007).

Factors driving the decrease of RWAs density towards the centres of the analysed clear-cut areas remain unknown. Nevertheless, because RWAs have a strong impact on many other invertebrates (Hawes et al. 2002; Koivula and Niemelä 2003; Reznikova and Dorosheva 2004; Dorosheva and Reznikova 2006; Ohashi et al. 2007), their distribution pattern can have serious consequences. The decrease of the number of RWAs towards the clear-cut area centre may result in a change of the density of other invertebrates that suffer from predation by and/or competition with RWAs. If this is the case, clear-cut area centres may act as refugia for some epigeic invertebrates. On the other hand, some other species may decrease on the clear-cut areas because the release from RWAs predation may simplify the invertebrate community through strengthened interspecific competition that RWAs usually alleviate and through loss of niches that are usually provided by RWAs (e.g. Laakso and Setälä 1997). The predictions need further research, since their verification may help to explain the effect of timber harvesting on forest biodiversity.