Printer Friendly

LANDSCAPE FRAGMENTATION AND FOREST COMPOSITION EFFECTS ON GROUSE BREEDING SUCCESS IN BOREAL FORESTS.

SAMI KURKI [1,5]

ARI NIKULA [2]

PEKKA HELLE [3]

HARTO LIND[acute{E}]N [4]

Abstract. We examined the breeding success of forest grouse in relation to anthropogenic forest fragmentation in Finland. Employing Geographic Information Systems (GIS) and grouse data derived from Finnish wildlife triangle censuses conducted during 1989-1994, we combined the locations of 2267 Black Grouse (Tetrao tetrix) and 1060 Capercaillie (T. urogallus) females after the breeding season in mid-August with landscape data. The indicators of breeding success were the proportion of grouse hens with a brood and brood size. Two study areas (each 45 000 [km.sup.2]) in the boreal zone were selected for investigation.

The breeding success of grouse was negatively correlated with both fragmentation of forest area per se by farmland and the decreasing proportion of older forest as a result of clear-cutting. The extent of landscape accounting best for variation in nesting success was an order of magnitude larger ([sim]100 [km.sup.2]) than the area most probably used by a grouse female and her brood during the summer, which suggests that landscape-scale factors may override local factors such as track size and distance from edge. The proportion of grouse hens with a brood was lower in heavily fragmented landscapes than in more continuous forest landscapes whereas only minor differences in brood size were detected. We suggest that the most likely cause of the observed spatial correlation was higher nest predation by generalist predators in fragmented forest landscapes. The effects of landscape composition on the breeding success of grouse were more marked in northern than in southern Finland, probably because predator populati ons are more food-regulated in the north. The diminished breeding success of forest grouse as a result of increased forest fragmentation is a probable cause of population declines in forest grouse species during the past decades in Fennoscandia.

Key words. Black Grouse; Capercaillie; Finland; forest fragmentation; Geographic information Systems (GIS); landscape ecology; landscape structure; nest loss; predation; Tetrao tetrix; Tetrao urogallus; tetraonids.

INTRODUCTION

By influencing natural community structure, changes caused in landscape composition by human activities can affect the viability of populations indirectly through interspecific interactions (Danielson 1991, Angelstam 1992, Dunning et al. 1992, Wiens et al. 1993, Hanski 1995, Didham et al. 1996, Fahrig and Grez 1996). Elevated rates of brood parasitism and nest predation on forest birds as a result of forest fragmentation have probably received the most attention in this respect (e.g., Paton 1994, Andr[acute{e}]n 1995, Donovan et al. 1995, Marini et al. 1995, Robinson et al. 1995, Huhta 1996). The vast majority of studies have focused on the spatial variation in predation rate at forest patch level, i.e., the effects of forest patch size and distance to the nearest forest edge on the predation risk of avian nests (reviewed by Andr[acute{e}]n 1995). Whether these effects can be detected, however, depends on the habitat use of potential predators in the landscape (Andr[acute{e}]n 1995, Marmni et al. 1995). If t he density of the predator population is generally related to landscape composition, but individual predators use the entire landscape mosaic, i.e., the landscape is heterogeneous but undivided from their standpoint (Addicott et al. 1987), then the variation in predation pressure occurs on a larger, landscape scale rather than in forest patches of different size.

In Fennoscandia, two hypotheses have been advanced to explain how human-caused alterations in boreal landscape structure may increase predation rates on nests of ground-nesting forest birds. First, predation pressure may be higher in forest areas fragmented and interspersed with agricultural land (Andr[acute{e}]n et al. 1985, Andr[acute{e}]n 1992, Kurki and Lind[acute{e}]n 1995, Huhta 1996), as higher densities of generalist predators are supported by landscapes containing agricultural patches (Andr[acute{e}]n 1992, Kurki et al. 1998). This hypothesis has been developed largely through experimental studies using artificial nests. The lack of an incubating female and the artificial selection of the nest site, however, make it difficult to apply the results to real nests (Storaas 1988, Willebrand and Marcstr[ddot{o}]m 1988). Second, during the past four decades the widespread fragmentation of old forests by clear-cutting has increased the proportion of young succession stages in the Fennoscandian boreal landsc apes (Esseen et al. 1992, Hansson 1992, Anonymous 1996). Clear-cuttings and plantations, often with grass-dominated undergrowth, are suitable habitats for voles of the genus Microtus (e.g., Hansson 1994). Microtus voles, in turn, can attain much higher densities than original forest rodent species (e.g., Clethrionomys voles) in a given area (Henttonen 1989). Furthermore, because Microtus voles are the preferred prey for many generalist predators (Henttonen 1989), fragmentation of old forest has been suggested to benefit these predators (Henttonen 1989, Lindstr[ddot{o}]m 1989a, Angelstam 1992, Kurki et al. 1998). Although predators may not actively search for the ground nests of birds (Angelstam 1986, Vickery et al. 1992), the likelihood of incidental predation will be increased because of the higher predator densities (Henttonen 1989, Rolstad and Wegge 1989, Angelstam 1992). Despite the interest it has aroused, this hypothesis has not yet been verified with empirical data.

We tested these two hypotheses and the role of landscape scale by investigating the spatial variation in the breeding success of two ground-nesting tetraonids, Black Grouse (Tetrao tetrix L.) and Capercaillie (T. urogallus L.). During the past three-four decades the populations of these forest grouse species have clearly declined in Finland (Lind[acute{e}]n and Rajala 1981, Helle and Helle 1991) and the increased nest and brood predation pressure due to forest fragmentation has been suggested as one of the causes (e.g., Henttonen 1989).

Using data from Finnish wildlife triangle censuses (1989-1994), we combined the locations of 2267 Black Grouse and 1060 Capercaillie females with land use and forest resource data by employing Geographic Information Systems (GIS). The proportion of grouse hens with a brood and the brood size in mid-August were used to indicate breeding success. To examine spatial correlation between landscape composition and grouse breeding success, we applied multivariate logistic and linear regression analyses. The study was performed in two areas in Finland located in different boreal subzones, and so differed to some extent in their predator communities and landscape compositions.

METHODS

Study species

Black Grouse and Capercaillie are forest-dwelling grouse species with lekking behavior, ground nests, and precocial chicks. Black Grouse prefer younger and more open forests whereas Capercaillie favor mature forests most of the year (Angelstam 1983, Swenson and Angelstam 1993, Rolstad and Wegge 1989). The cryptically colored females of both species begin to breed in their first spring and raise one brood solitarily during the summer, but casual renesting may occur after early loss of nest. In Finland, the mean clutch size of Black Grouse (7.9) is about one egg larger than that of Capercaillie (7.1) (Rajala 1974). According to studies of radio-tagged birds, virtually all Black Grouse and Capercaillie females attempt to breed (Angelstam 1983, Rolstad et al. 1988, Willebrand 1988, Marjakangas 1996; P. Helle and P. Kumpu unpublished data). They do not seem to show a clear preference in nest site selection for a particular successional stage of forest stand, nests of both species being found in clear-cuttings and plantations as well as in mature forests (Storans and Wegge 1987).

Wildlife triangle data

The wildlife triangle scheme has been described in detail by Lind[acute{e}]n et al. (1996), and we will only summarize it briefly here. The nationwide monitoring program, organized by the Finnish Game and Fisheries Research Institute and run by volunteer hunters is based on permanent transect lines located randomly. The basic unit in the system is an equilateral triangular route (12 km). A triangle should be located within one topographic map sheet (10 X 10 km), and no map sheet should contain more than one triangle. The whole monitoring network currently consists of some 1500 triangles countrywide (Fig. 1).

Grouse censuses along the routes of triangles are conducted after the most intensive breeding season, in mid-August. Tetraonids (Capercaillie, Black Grouse, Hazel Grouse (Bonasa bonasia L.) and Willow Grouse (Lagopus lagopus L.)) are censused from a 60 m wide belt by a chain of three people. The total census area covered by one triangle is thus 0.72 [km.sup.2]. All observations of grouse are pinpointed and marked on a topographic map (1:20 000). Black Grouse and Capercaillie females are still with their brood throughout Finland in mid-August, and the wildlife triangle census produces two parameters of breeding success: percentage of grouse hens with a brood (later %HWB; in statistical analyses treated as a categorical variable 0/1) and brood size (later BS) (Rajala 1974, Lind[acute{e}]n 1981). These indices are possible because the sex and age (old or young of the year) of Black Grouse and Capercaillie can be determined in the census. The %HWB and BS have commonly been used in field studies (Rajala 1974, Lind[acute{e}]n 1981, Lindstr[ddot{o}]m et al. 1987, Marcstr[ddot{o}]m et al. 1988). We used the observations of brood size in the analyses only if the number of chicks, according to census-takers, could be counted reliably. The recorded %HWB may be an overestimate owing to the lower probability of observations of broodless hens in the census (Britras and Karlbom 1990), but temporal and spatial analyses of the variable are nonetheless justified. Because an observation in our data requires the existence of an adult grouse female, the environment cannot be entirely unsuitable for grouse. We stress, however, that we are focusing here on spatial variation in breeding success, not on density.

Study areas

The boreal forest zone in Finland has been divided into latitudinal subzones based on plant species composition (Ahti et al. 1968). The length of the growing season, the proportion of agricultural land and the density of the human population increase whereas the period of snow cover decreases southwards in this country. Further, the degree of cyclicity of vole populations decreases southwards throughout Fennoscandia (Hansson and Henttonen 1985, 1988). Finally, most likely due to the general reduction in predation pressure, the breeding success of Black Grouse and Capercaillie improves in general northwards in Finland (Kurki et al. 1997). There might be latitudinal variation also in the importance of different predator species as nest and brood predators, thus causing variation in the effects of landscape structure. To illustrate this latitudinal component, we selected two study areas (Fig. 1; the northern and southern study areas), each one covering [sim]45 000 [km.sup.2] (300 X 150 km; Fig. 1). They were loc ated to include variation in landscape composition, particularly in the percentage of agricultural land, while minimizing the regional variation related to vole population cyclicity and predator community. The final selection of study areas, however, was also constrained by the availability of landscape data. From the study areas we then selected all wildlife triangles censused at least in three years during 1989-1994 (Fig. 1).

The northern study area lies in the middle and northern boreal zones (Ahti et al. 1968). In this area, 201 wildlife triangles were selected for investigation, and 1518 Black Grouse and 796 Capercaillie females were located during line-transect counts in a six-year study period (Table 1, Fig. l). The landscape was dominated by different-aged coniferous forests (Scots pine, Pinus sylvestris L. and Norwegian spruce, Picea abies Karst.) with a mixture of deciduous species (mainly Betula spp. and a sparse population of aspen Populus tremula L.). Open bogs and lakes are common. The percentage of agricultural land is low in the east, but increases westward. Vole populations (both Clethrionomys and Microtus) are cyclic, with 3-4 yr periodicity (Hansson and Henttonen 1985). Potential avian nest and brood predators in the northern area are the Common Raven Corvus corax L., Hooded Crow C. corone cornix L., European Jay Garrulus glandarius L., and Black-billed Magpie Pica pica L. Potential mammalian predators include re d fox Vulpes vulpes L., pine marten Martes martes L., and stoat Mustela erminea L. The raccoon dog Nyctereutes procyonoides Gray and badger Meles meles L. occur in low densities mainly in the west of the study area. Furthermore, a number of birds of prey, particularly Goshawks Accipiter gentilis L., affect the breeding success of grouse by killing females and chicks during the breeding season (e.g., Angelstam 1984).

The southern study area lies mainly in the southern boreal zone (Ahti et al. 1968); 749 Black Grouse and 264 Capercaillie females were located in the 169 wildlife triangles included in the database (Table 1, Fig. 1). Agricultural land and settlements cover a higher proportion of the land than in the northern study area, but open bogs are less common. The southern study area is located in a transition zone between areas of clearly cyclic and noncyclic vole populations (Hansson and Henttonen 1988). Potential predator species are the same as in the northern study area, but the population densities of many predators are considerably higher (Hanski et al. 1991, Helle and Kauhala 1991, Helle et al. 1996, Kurki et al. 1997).

Landscape data

The land use and forest resource data for the study derive from classified composite Landsat TM 5 images (pixel 25 X 25 m) taken during 1987-1991. For the southern area, data were classified by the National Land Survey (NLS) of Finland (Vuorela 1997). Each pixel in the NLS classified images can belong to one of originally about 70 land use and forest classes. Digital maps of nonforest lands (water area, agricultural land, roads, and settlements) are used to improve the accuracy of classification. For forest areas, each pixel is classified according to total timber volume (50 [m.sup.3]/ha intervals) and further according to dominant tree species or mixed tree species composition. Clearcuttings and plantations are also given as separate classes (Vuorela 1997).

For the northern area, we used classified images produced by the National Forest Inventory, NFI (Tomppo 1993, 1996). The NFI also uses digital maps of other land use forms to improve the accuracy of classification. For each pixel on forest land, the NFI produces an estimate of growing stock separately for Scots pine, Norwegian spruce, and deciduous species. This classification of forest land leads to a multilayer image in which each layer gives an estimate of the growing stock for one of the tree species (Tomppo 1993, 1996).

Because our aim was to study the effects of forest fragmentation resulting from human activity, we reclassified primary landscape data as shown in Table 2. For the southern area, we simply combined the classes of original data to correspond with those of Table 2. For the northern area, we combined information on the volume of separate tree species (layers) to produce the respective classes. For both study areas, we used a separate digital mask for agricultural land, digitized from 1:50 000 topographic maps. Timber volume correlates very well with the age of the forest stand (Anon 1996) and is therefore applicable in landscape-level analyses of forest fragmentation. Because of the slightly different criteria used to form land use classes in the NLS and NFI classification (Table 2), however, the loading of all habitat classes is not exactly the same in study areas. Statistical comparisons are therefore justified only within study areas, not between them.

Analyses of landscape composition

Using GIS, we formed a map layer containing digitized locations of grouse females. The final spatial accuracy of locations is [sim]100 m. The breeding ecology of Black Grouse and Capercaillie is relatively well known; in efforts to explain the %HWB and BS in mid-August, landscape, in theory, can be divided into three scales based on important ecological processes: (1) the scale of habitat selection of individuals in mid-August, (2) the scale of nest location and home range of grouse females and their broods during the summer, and (3) the scale of predator density impact. To convert these ecological scales into hectares and square kilometers, information on movements of grouse females during the breeding season is required. In an unpublished data set collected by A. Marjakangas, the mean distance between nest site and the location of the same hen in August was 550 m (N = 20, range 80-1540 m) and 1025 m (N = 18, range 20-2430 m) for radio-tagged Black Grouse hens with and without a brood, respectively. There was, however, also one broodless hen, not included in the mean, that had dispersed 8500 m. Other authors, too, have reported the movements or home range sizes of radio-tagged Black Grouse and Capercaillie females during the breeding season (Angelstam 1983, Rolstad et al. 1988, Willebrand 1988; P. Valkeaj[ddot{a}]rvi, personal communication). On the basis of this information, we formed a circular landscape with a radius of 1500 m around each grouse female located and assumed that the resulting area ([sim]700 ha) will, in nearly all cases, include the nest site and the home range of females and their brood during the summer. Additionally, to depict the differences, if any, in habitat selection between hens with and without a brood, we formed a smaller landscape with a radius of 500 m (Fig. 2). The scale of habitat selection could have been even smaller, but it was restricted by the accuracy of locations (l00 m). To predict the density of predators within the area used by a grouse female and her brood during the summer, the extent of the landscape should be larger (Kurki et al. 1998). We therefore formed two circular landscapes with radii of 5000 m and 10 000 m, respectively, for each wildlife triangle, using the triangle center point. With these radii, the composition of the landscape obtained represents the situation fairly well for all grouse female locations on the same triangle route (Fig. 2). The predator species with the largest home ranges are red fox, pine marten, and raven (see Study areas). Kurki et al. (1998) examined the effects of landscape composition on the relative densities of red fox and pine marten using three radii (3000, 5000, and 10 000 m) of circular landscape. With the radius of 5000 in, the landscape composition explained the variation in winter track densities best in both species.

The percentages of different habitat classes (Tables 2 and 3) and the edge density (length of edge per hectare) of agricultural land with a radius of 1500 m were computed with FRAGSTATS (McGarigal and Marks 1995). For further details about the integrative use of wildlife triangle and landscape data, see Helle and Nikula (1996).

Statistical analyses

Before testing the effects of landscape composition, we removed the annual variation in breeding success indices. Spatial asynchrony in breeding success within study areas may, however, still impair analyses. In both study areas, we therefore examined the spatial--temporal synchrony of breeding success by testing the significance of the interaction terms YEAR X LATITUDE and YEAR X LONGITUDE. We further divided the study areas into different subareas (3-4) and tested the significance of the interaction YEAR X SUBAREA.

No spatial variation in synchrony of breeding success was detected in the southern study area (P [greater than] 0.1 in all cases). Because the annual variation in %HWB was not significant for either Capercaillie or Black Grouse, we pooled the data from different years. In the southern area, we analyzed the brood size of Black Grouse only because of the low number of observations of Capercaillie broods (Table 1). The brood size of Black Grouse was standardized (mean = 0, SD = 1) for each year.

Likewise, within the northern study area the annual variation in brood size tended to be spatially constant for both species (P [greater than] 0.2 in all cases), but the effect of the interaction YEAR) X LONGITUDE on %HWB was significant (P [less than] 0.01) for both. Closer examination revealed that the lower level of %HWB in 1990 and 1991, most likely caused by the increased predation pressure after the vole crash during winter 1989-1990 (Marjakangas 1996, Kurki et al. 1997; A. Kaikusalo, unpublished data), and the hi gh levels of %HWB in 1989 and 1992 (vole peaks), only applied to the eastern part of the northern study area. In the west, the annual variation was not significant. The pattern was similar for both grouse species. To overcome the confusing effect of spatial asynchrony, we first divided the northern study area into western and eastern parts. We found a fairly distinct (within [sim]50 km) border where the annual variation in %HWB began to appear. Next we created a new variable, YEAREFFECT, to remove the annual variation in %HWB in both sides of the border separately. We forced YEAREFFECT into models before testing the effect of landscape composition on %HWB in the northern study area. Also in the northern study area, the brood size of both grouse species was standardized (mean = 0, SD = 1) for each year.

Because the percentages of different habitat classes in the landscape sum to 100%, they are not independent of each other (the unit-sum constraint). Aitchison (1986) recommended the use of log ratios between habitat classes instead of percentages, a method known as compositional analysis. These log ratios can then be analyzed with standard multivariate statistics. Aebischer et al. (1993) described how compositional analysis can be used for investigating the habitat use of radio-tagged animals, i.e., when habitat composition is considered as a dependent variable. The relationship between the dependent variable and habitat composition as an independent variable, however, can be tested by applying stepwise multiple regression methods (e.g., Robertson 1994). We computed all log-ratio combinations between different habitat classes (Table 2) with different extents of landscape, and used them as potential explanatory variables in forward stepwise logistic regression for %HWB (Hosmer and Lemeshow 1989, SAS 1995), and forward stepwise multiple linear regression for brood size. Because it is impossible to increase one habitat type in the landscape without decreasing the others, the log ratio a/b is interpreted as the manner in which an increase in habitat type a at the cost of habitat type b will affect the dependent variable. It should be noted, however, that only log ratios with the same denominator are independent of each other. To enable transformations to be carried out, zero proportions were replaced with 0.001, which was an order of magnitude smaller than the smallest proportion detected (Aebischer et al. 1993).

Because we used the biogeographical variation along longitude to attain variation in landscape composition, LONGITUDE and also LATITUDE were included as a potential explanatory variable. The proportions of different habitat classes, and therefore, log ratios from different scales also correlate because the larger landscapes include smaller ones (see Table 4). We assume, however, that the scale on which the target ecological process is most heavily affected by landscape composition should account best for the variation and therefore should be selected in the model. A significance level of 0.05 was used when variables were included or excluded in modeling. Statistical analyses were made separately for grouse species and study areas. All statistical analyses were performed with the SAS statistical package (SAS 1989).

RESULTS

Before applying stepwise regression procedures, we examined the effects of interaction between census year or YEAREFFECT (in the case of %HWB in the northern study area) and landscape composition on breeding success indices. No evidence of significant interaction in either northern or southern study areas was found (P [greater than] 0.05 in all cases). Thus, the breeding success of grouse tended to be high or low in the same kind of landscapes in all study years.

Northern study area

In the northern study area, after INTERCEPT and YEAREFFECT had been included in the models, the log ratio OLFOR/YOUFOR(10 000 m) clearly accounted best for the spatial variation in %HWB in both grouse species (Table 5). The same log ratio was also significantly positively associated with %HWB on all other scales in both species, but the statistical significance increased with increasing landscape extent. At the second step, the negative effects of the log ratios AGRI/CCPLM(10 000 m) and AGRI/YOUFOR(l0 000 m) accounted best for the remaining variation in Capercaillie and Black Grouse, respectively (Table 5). This was clearly due to AGRI in the log ratios, because most of the log ratios on the two largest scales containing AGRI had significant effects. Nevertheless, the negative effect of agricultural land was evident only on the two largest scales. At step three, the positive effects of log ratios containing agricultural land on the two smallest scales appeared, and tended to account best for the remaining var iation. With Capercaillie, the positive effect of AGRIEDGE(1500 m) was included in the model. Likewise, with Black Grouse, the positive effect of AGRIEDGE(1500) was the most significant of the remaining variables (P = 0.085), but it was not included in the model because of the significance criterion used (P = 0.05).

In general, the models attained (Table 5) fit well with the data (Fig. 3), although they could account for only small amount of spatial variation in %HWB (Table 5). The results indicate that, on a larger scale, both the increasing percentage of agricultural land and the decreasing percentage of older forests at the cost of younger forests in the landscape decrease the probability of an observed grouse female being with a brood. On a smaller scale, however, grouse hens with a brood were more likely to be found near agricultural patches than were those without a brood.

Landscape composition had only a minor effect on the brood size of grouse. With Black Grouse, none of the variables accounted for the variation well enough to be included in the model. The positive effects of log ratios OLFOR/WATER(10 000 m) and OLFOR/YOUFOR(10 000 m) together accounted for 6.5% of the spatial variation in the brood size of Capercaillie (Table 6).

Southern study area

In the southern study area, the number of Capercaillie females observed (264) was much smaller than that of Black Grouse (749). None of the landscape variables significantly explained the spatial variation in the %HWB of Capercaillie (Table 5). With Black Grouse, the positive effect of the log ratio of OLFOR/WATER(5000 m) entered the model at the first step, and the positive effect of log ratio OLFOR/YOUFOR(5000 m) accounted best for the remaining variation at the second step. Finally, the negative effect of the log ratio AGRI/YOUFOR(l0 000 m) was included (Table 5). However, the model obtained did not predict the spatial variation in %HWB as well as it did in the northern study area (Fig. 3). The negative effect of log ratio OLFOR/AGRI(5000 m) accounted for 3.9% of the spatial variation in the brood size of Black Grouse (Table 5).

DISCUSSION

We demonstrated a spatial correlation between the composition of boreal forest landscape and the breeding success of grouse (Table 5, Figs. 3 and 4). Both fragmentation of forest area per se by agricultural land and the decreasing percentage of older forest were negatively correlated with the breeding success of grouse. However, a number of additional patterns can be recognized: (1) the connection between landscape composition and breeding success of grouse seems to be stronger in the northern than in the southern study area; (2) landscape composition affects the probability of an observed grouse hen being with a brood, but does not so clearly affect brood size in mid-August; and (3) the area of landscape best accounting for spatial variation in breeding success is larger than the home range of the grouse female and her brood during the summer.

Causal factors

The spatial correlation between breeding success of grouse and landscape composition is not directly informative about the underlying mechanisms. We can assume, however, that if the correlation is produced by local factors such as (1) differences in habitat selection between grouse hens with and without a brood (e.g., Rolstad et al. 1988), (2) the direct effects of habitat quality on survival of broods, or (3) an edge-related increase in predation pressure (Andr[acute{e}]n and Angelstam 1988), then the extent of landscape that best accounts for the variation in breeding success should not be larger than the one used by the grouse female and her brood during the summer. In both study areas, the correlation was strongest with an area clearly larger than the one used by grouse females and their broods. We suggest, therefore, that either: (1) landscape composition affects the mating success of females by impairing the viability of grouse leks, or (2) landscape composition may increase densities of generalist pred ators with large home ranges in the landscape. The mating success hypothesis can be true only for Capercaillie, which require large, continuous old-forest areas for lekking sites, whereas Black Grouse lekking sites are usually located in open areas such as open bogs, fields, or lake ices (Rolstad and Wegge l989, Swenson and Angelstam 1993). Because the effects of landscape composition on the breeding success of both grouse species were so similar (Table 5), we feel that predation is the most likely explanation.

Further evidence for the predation hypothesis and potential predators

Northern area.--In the northern study area, the landscape traits that correlated negatively with the breeding success of grouse in our study correlated positively with the local abundance of red fox, and accounted for 26% of the spatial variation in fox abundance (Kurki et al. 1998). The relative density of pine marten, in turn, was negatively affected by the fragmentation of forest area due to agricultural land, but positively related to an increasing percentage of younger forest in the landscape (Kurki et al. 1998). Furthermore, Kurki et al. (1997) combined the annual red fox densities and breeding success indices of grouse in a study area partly overlapping that used by us, and found that the increase in fox density significantly decreased the probability of a grouse female being with a brood in low vole years, but the relationship between brood size and fox density was not significant. This is probably because grouse chicks learn to fly at the age of [sim]10 days and are thereafter able to escape mammali an predators; before that, fox predation may cause total loss of nest or brood. The red fox has also been shown to be an important grouse nest predator in other studies conducted in Fennoscandia (Lindstr[ddot{o}]m et al. 1987, 1994, Marcstr[ddot{o}]m et al. 1988).

Andr[acute{e}]n (1992) examined the densities of different corvid species in relation to forest fragmentation in central Sweden. He found that the total density of corvids increased as the forest area became fragmented and interspersed with agricultural land. The main predation pressure caused by corvids, however, seemed to be at forest/field edges (Andr[acute{e}]n and Angelstam 1988). Therefore, the percentage of agricultural land, and especially the length of edge between forest and agricultural land, should explain spatial variation best on a smaller scale (r = 1500 in), which was not the case.

We conclude that the spatial relationship observed between breeding success and landscape composition on a larger scale was most likely caused by the effect of predation in the northern study area. Evidence that red fox is a potential predator is strong, but we do not exclude the probable additional effect of other generalist predators. Increased predation affects the survival of nests and broods immediately after hatching, when the chance of total loss is highest.

Southern area.--Before attempting to interpret the results, we should note some important differences between the northern and southern study areas. First, the abundance of potential nest and brood predators is higher in the southern than in the northern study area (Hanski et al. 1991, Helle and Kauhala 1991, Helle et al. 1996, Kurki et al. 1997). Second, the degree of cyclicity of vole populations is lower and the number of potential alternative prey species for generalist predators is higher in the southern study area (Angelstam et al. 1984, Hansson and Henttonen 1988, Kurki et al. 1997). Third, the hypotheses concerning landscape-dependent predator densities (see Introduction) assume that the predator populations are mainly food regulated. This may well be true in the northern study area, but it is less likely in the southern study area, where the predator densities are generally much higher and the role of social regulation is more important (e.g., Englund 1970, Lindstr[ddot{o}]m 1989b).

In the southern study area, the probability of a Black Grouse hen being with a brood was significantly related to landscape composition, whereas that of Capercaillie was not. We suggest that this is rather due to the lower number of Capercaillie observations (Table 1) than to real differences between grouse species. The fragmentation of older forest and the increasing percentage of agricultural land were negatively associated with the probability of a Black Grouse hen being with a brood (Table 5). However, in the southern study area, the increasing percentage of water area also had a negative effect (Table 5). Kurki et al. (1998) found that fox density in the southern area was negatively correlated with the percentage of older forest, and was highest in the landscapes having [sim]20-30% of agricultural land. We found, however, no evidence for nonlinearity in breeding success of grouse in relation to the amount of agricultural land. In the same area, fox abundance could not significantly account for spatial v ariation in grouse breeding success in annual data (Kurki et al. 1997), indicating a reduced role of red fox in that predator community compared with the one in the northern study area (Kurki et al. 1997). One important predator in the southern study area is raccoon dog (Helle and Kauhala 1991), which has a preference for lake shores, and it may explain the negative effect of water area on the breeding success of Black Grouse (Table 5).

We further suggest that the correlation detected between the %HWB of Black Grouse and landscape composition in the southern study area was at least partly because of predation. Predation pressure, however, is generally higher in the south than in the north (Kurki et al. 1997), and both the predator and alternative prey guilds are more diverse. Moreover, because the predator populations are probably not as clearly regulated by food resources as they are in the northern study area, the predictive power of landscape composition is lower in the south (Fig. 3; see Kurki et al. 1998). However, because the habitat classification varies slightly between study areas (Table 2), we may have been more "successful" in habitat classification in the northern than in the southern area.

Small-scale processes

After the effect of landscape composition on a larger scale was included in the models, the positive effect of agricultural land on %HWB on a smaller scale appeared in the northern study area. Although the length of edge between agricultural and forest land on the 1500-m scale was included only in the model for Capercaillie explaining the probability of an observed hen being with a brood (Table 5), it also accounted best for the remaining variation in Black Grouse. An indication of a similar pattern was found by Kurki and Lind[acute{e}]n (1995). Clearly, agricultural lands are not distributed at random in the rather poor soils in the north, but have been cleared in areas with high natural productivity. Thus, the productivity and structure of the surrounding forest stands may also differ, providing a better brooding habitat for grouse. It is difficult to say whether the observed relationship is then caused by the difference in habitat selection between hens with and without a brood, or by the better survival of broods in those areas. If the latter is true, however, we would expect a response in brood size too, and this was not detected.

Conclusions

Information from this and our other studies suggests that the most likely cause of the spatial correlation between the composition of forest landscape and the breeding success of grouse is the elevated predation pressure caused by generalist predators. Predator populations, however, are also affected by factors other than landscape composition (e.g., Kurki et al. 1998); hence, landscape composition cannot entirely explain the spatial variation in predation pressure on grouse nests and broods (Tables 5 and 6).

In general, our findings support the hypothesis that human-caused changes in boreal landscape composition have altered the role of the interaction between forest grouse and their nest and brood predators in Fennoscandia. In northern Fennoscandia, the red fox is probably a key species among potential predators, whereas other predator species may also be involved in the south. Our results stress the importance of landscape extent being clearly larger than the area determined from the standpoint of an individual grouse female. Therefore, when trying to understand how anthropogenic forest fragmentation affects predation pressure in general, we need to use both landscape- and patch-level approaches (Brown and Litvaitis 1994, Andr[acute{e}]n 1995). On a broader, regional scale, however, variation in the structure of predator communities and factors that regulate predator populations may further modify the responses of predator populations to human caused-changes in landscape composition (Fig. 5).

Tetraonids have experienced significant population declines in Fennoscandia during the past four decades (e.g., Wegge 1979, Lind[acute{e}]n and Rajala 1981, Helle and Helle 1991). One plausible reason for this trend is diminished breeding success due to increased old-forest fragmentation by clear-cuttings (Esseen et al. 1992, Anonymous 1996). With that in mind, the similar population trends noted in all grouse species become more understandable (Lind[acute{e}]n and Rajala 1981).

ACKNOWLEDGMENTS

This study was financially supported by the Maj and Tor Nessling foundation and the Academy of Finland (FIBRE program). The data processing required input from many people and we would like to especially thank A. Marjakangas, V. Nivala, E. Tomppo, and M. Wikman for their invaluable help during the project. Y. Haila, H. Hakkarainen, E. Haukioja, E. Korpim[ddot{a}]ki, E. Lehikoinen, S. K. Robinson, and two anonymous referees provided comments on earlier versions of this paper, which is a contribution to the research program on Finnish grouse.

(1.) Section of Ecology, Department of Biology, University of Turku, FIN-20014 Turku, Finland

(2.) Finnish Forest Research Institute, Rovaniemi Research Station, P.O. Box 16, FJN-96301 Rovaniemi, Finland

(3.) Finnish Game and Fisheries Research Institute, Oulu Research Station, FIN-90570 Oulu, Finland

(4.) Finnish Game and Fisheries Research Institute, P.O. Box 202, FIN-00151 Helsinki, Finland

(5.) E-mail: sami.kurki@sjoki.uta.fi

LITERATURE CITED

Addicott, J. F., J. M. Aho, M. F. Antolin, D. K. Padilla, J. S. Rickhardson, and D. A. Soluk. 1987. Ecological neighborhoods: scaling environmental patterns. Oikos 49:340-346.

Aebicher, N. J., P. A. Robertson, and R. E. Kenward. 1993. Compositional analysis of habitat utilization from animal radiotelemetry data. Ecology 74:1313-1325.

Ahti, T., L. H[ddot{a}]met-Ahti, and J. Jalas. 1968. Vegetation zones and their sections in northwestern Europe. Annales Botanigici Fennici 5:169-211.

Aitchison, J. 1986. The statistical analysis of compositional data, Chapman and Hall, London, UK.

Andr[acute{e}]n, H. 1992. Corvid density and nest predation in relation to forest fragmentation: a landscape perspective. Ecology 73:794-804.

Andr[acute{e}]n, H. 1995. Effects of landscape composition on predation rates at habitat edges. Pages 225-255 in L. Hansson, L. Fahrig, and G. Merriam, editors. Mosaic landscapes and ecological processes. Chapman and Hall, London, UK.

Andr[acute{e}]n, H., and P. Angelstam. 1988. Elevated predation rates as an edge effect in habitat islands: experimental evidence. Ecology 69:544-547.

Andr[acute{e}]n, H., P. Angelstam, E. Lindstr[ddot{o}]m, and P. Wid[acute{e}]n. 1985. Differences in predation pressure in relation to habitat fragmentation: an experiment. Oikos 45:273-277.

Angelstam, P. 1983. Population dynamics of tetraonids, especially the black grouse Tetrao tetrix L., in boreal forests. Dissertation. Uppsala University, Uppsala, Sweden.

Angelstam, P. 1984. Sexual and seasonal differences in mortality of the Black Grouse Tetrao tetrix in boreal Sweden. Ornis Scandinavica 15:123-133.

Angelstam, P. 1986. Predation on ground-nesting birds' nests in relation to predator densities and habitat edge. Qikos 47:365-373.

Angelstam, P. 1992. Conservation of communities--the importance of edges, surroundings and landscape mosaic structure. Pages 9-70 in L. Hansson, editor. Ecological principles of nature conservation. Elsevier Applied Science, London, UK.

Angelstam, P., E. Lindstr[ddot{o}]m, and P. Wid[acute{e}]n. 1984. Role of predation in short-term population fluctuations of some birds and mammals in Fennoscandia. Oecologia 62:199-208.

Anonymous. 1996. Statistical Yearbook of Forestry 1996. The Finnish Forest Research Institute, Helsinki, Finland.

Brittas, R., and M. Karlbom. 1990. A field evaluation of the Finnish 3-man chain: a method for estimating forest grouse numbers and habitat use. Ornis Fennica 67:18-23.

Brown, A. L., and J. A. Litvaitis. 1994. Habitat features associated with predation of New England cottontails: What scale is appropriate? Canadian Journal of Zoology 73:1005-1011.

Danielson, B. J. 1991. Communities in a landscape: the influence of habitat heterogeneity on the interactions between species. American Naturalist 138:1105-1120.

Didham, R. K., J. Ghazoul, N. E. Stork, and A J. Davis. 1996. Insects in fragmented forests: a functional approach. Trends in Ecology and Evolution 11:255-260.

Donovan, T. M., F R. Thompson, III, J. Faaborg, and J. R. Probst. 1995. Reproductive success of migratory birds in habitat sources and sinks. Conservation Biology 9:1380-1395.

Dunning, J. B., B. J. Danielson, and H. R. Pulliam. 1992. Ecological processes that affect populations in complex landscapes. Oikos 65:169-175.

Englund, J. 1970. Some aspects of reproduction and mortality rates in Swedish foxes (Vulpes vulpes) 1961-1963 and 1966-1969. Viltrevy 8:1-82.

Esseen, P. -A., B. Ehnstr[ddot{o}]m, L. Ericson, and K. Sj[ddot{o}]berg. 1992. Boreal forests--the focal habitats of Fennoscandia. Pages 252-325 in L. Hansson, editor. Ecological principles of nature conservation. Elsevier Applied Science, London, UK.

Fahrig, L., and A. A. Grez. 1996. Population spatial structure, human-caused landscape changes and species survival. Revista Chilena de Historia Natural 69:5-13.

Hanski, I. 1995. Effects of landscape pattern on competitive interactions. Pages 204-224 in L. Hansson, L. Fahrig, and G. Merriam, editors. Mosaic landscapes and ecological processes. Chapman and Hall, London, UK.

Hanski, I., L. Hansson, and H. Henttonen. 1991. Specialist predators, generalist predators, and the microtine rodent cycle. Journal of Animal Ecology 60:353-367.

Hansson, L. 1992. Landscape ecology of boreal forests. Trends in Ecology and Evolution 7:299-302.

Hansson, L. 1994. Vertebrate distributions relative to clear-cut edges in a boreal forest landscape. Landscape Ecology 9:105-115.

Hansson, L., and H. Henttonen. 1985. Gradients in density variations of small rodents: the importance latitude and snow cover. Oecologia 67:394-402.

Hansson, L., and H. Henttonen. 1988. Rodent dynamics as community processes. Trends in Ecology and Evolution 3:195-200.

Helle, E., P. Helle, H. Lind[acute{e}]n, and M. Wikman. 1996. Wildlife populations in Finland during 1990-1995, based on wildlife triangle data. Finnish Game Research 49:12-17.

Helle, E., and K. Kauhala 1991. Distribution history and present status of the raccoon dog in Finland. Holarctic Ecology 14:278-286.

Helle, P., and T. Helle 1991. How do changes in forest structure explain recent changes in Finnish grouse populations? Suomen Riista 37:56-66. (In Finnish with English summary.)

Helle, P., and A. Nikula. 1996. Usage of geographic information system (GIS) in analyses of wildlife triangle data. Finnish Game Research 49:26-36.

Henttonen, H. 1989. Does an increase in the rodent and predator densities, resulting from modern forestry, contribute to the long-term decline in Finnish tetraonids? Suomen Riista 35:83-90. (In Finnish with English summary.)

Hosmer, D. W., and S. Lemeshow. 1989. Applied logistic regression. Wiley, New York, New York, USA.

Huhta, E. 1996. Effects of forest fragmentation on reproductive success of birds in boreal forests. Dissertation. University of Jyv[ddot{a}]skyl[ddot{a}], Jyv[ddot{a}]skyl[ddot{a}], Finland.

Kurki, S., P. Helle, H. Lind[acute{e}]n, and A. Nikula. 1997. Breeding success of black grouse and capercaillie in relation to mammalian predator densities on two spatial scales. Oikos 79: 301-310.

Kurki, S. and H. Lind[acute{e}]n. 1995. Forest fragmentation due to agriculture affects the reproductive success of ground-nesting black grouse (Tetrao tetrix L.). Ecography 18:109-113.

Kurki, S., A. Nikula, P. Helle, and H. Lind[acute{e}]n. 1998. Abundance of red fox and pine marten in relation to the composition of boreal forest landscapes. Journal of Animal Ecology 67:874-886.

Lind[acute{e}]n, H. 1981. Estimation of juvenile mortality in the capercaillie, Tetrao urogallus, and the black grouse, Tetrao tetrix, from indirect evidence. Finnish Game Research 39: 35-51.

Lind[acute{e}]n, H., E. Helle, P. Helle, and M. Wikman. 1996. Wildlife triangle scheme in Finland: methods and aims for monitoring wildlife populations. Finnish Game Research 49:4-11.

Lind[acute{e}]n, H., and P. Rajala. 1981. Fluctuations and long-term trends in the relative densities of tetraonid populations in Finland. Finnish Game Research 39:13-34.

Lindstr[ddot{o}]m, E. 1989a. The role of medium-sized carnivores in the Nordic boreal forest. Finnish Game Research 46:53-63.

Lindstr[ddot{o}]m, E. 1989b. Food limitation and social regulation in a red fox population. Holarctic Ecology 12:70-79.

Lindstr[ddot{o}]m, E., H. Andr[acute{e}]n, P. Angelstam, G. Cederlund, B. H[ddot{o}]rnfeldt, L. J[ddot{a}]derberg, P.-A. Lemnell, B. Martinsson, K. Sk[ddot{o}]ld, and J. E. Swensson. 1994. Disease reveals the predator: sarcoptic mange, red fox predation, and prey populations. Ecology 75:1042-1049.

Lindstr[ddot{o}]m, E., P. Wid[acute{e}]n, P. Angelstam, and H. Andr[acute{e}]n. 1987. Do predators synchronize vole and grouse fluctuations? An experiment. Oikos 48:121-124.

Marcstr[ddot{o}]m, V., R. Kenward, and E. Engren. 1988. The impact of predation on boreal tetraonids during vole cycles: an experimental study. Journal of Animal Ecology 57:859-872.

Marini, M. A., S. K. Robinson, and E. J. Heske. 1995. Edge effects on nest predation in the Shawnee National Forest, southern Illinois. Biological Conservation 74:203-213.

Marjakangas, A. 1996. Nest losses in a boreal black grouse population. Suomen Riista 42:25-31. (In Finnish with English summary.)

McGarigal, K., and B. J. Marks. 1995. FRAGSTATS: spatial pattern analysis program for quantifying landscape structure. U.S. Forest Service General Technical Report, PNW 351.

Paton, P. W. C. 1994. The edge effect on avian nest success: how strong is the evidence? Conservation Biology 8:17- 26.

Rajala, P. 1974. The structure and reproduction of Finnish populations of capercaillie, Tetrao urogallus, and black grouse, Lyrurus tetrix, on the basis of late summer census data from 1963-66. Finnish Game Research 35:1-51.

Robertson, P. A. 1994. Pheasants and fragmentation: positive effects of landscape interspersion. Pages 169-175 in J. W. Dower, editor. Proceedings of the Third Annual IALE(UK) Conference, Fragmentation in Agricultural Landscapes. The UK Region of the International Association for Landscape Ecology, London, UK.

Robinson, S. K., F. R. Thompson, III, D. R. Donovan, J. Whitehead, and J. Faaborg. 1995. Regional forest fragmentation and the nesting success of migratory birds. Science 267:1987-1990.

Rolstad, J., and P. Wegge. 1989. Capercaillie (Tetrao urogallus L.) populations and modern forestry--a case for landscape ecological studies. Finnish Game Research 46:43-52.

Rolstad, J., P. Wegge, and B. B. Larsen. 1988. Spacing and habitat use of capercaillie during summer. Canadian Journal of Zoology 66:670-679.

SAS Institute. 1989. SAS/STAT user's guide. Fourth edition. Volumes 1-2. SAS Institute, Cary, North Carolina, USA.

SAS Institute. 1995. Logistic regression examples using the SAS system. Version 6, first edition. Cary, North Carolina, USA.

Storaas, T. 1988. A comparison of losses in artificial and naturally occurring Capercaillie nests. Journal of Wildlife Management 52:123-126.

Storaas, T., and P. Wegge. 1987. Nesting habitats and nest predation in sympatric populations of Capercaillie and Black Grouse. Journal of Wildlife Management 51:167-172.

Swenson, J. E., and P. Angelstam. 1993. Habitat separation by sympatric forest grouse in Fennoscandia in relation to boreal forest succession. Canadian Journal of Zoology 71:1303-1310.

Tomppo, E. 1993. Multi-source National Forest Inventory of Finland. In Aarne Nyyss[ddot{o}]nen, Simo Poso, and Johanna Rautala, editors. Proceedings of Ilvessalo Symposium on National Forest Inventories. Finland 17-21 August 1992. The Finnish Forest Research Institute. Research Papers 444:52-59.

Tomppo, E. 1996. Multi-source National Forest Inventory of Finland. In R. P[ddot{a}]ivinen, J. Vanclay, and S. Miina, editors. New Thrusts in Forest Inventory. Proceedings of the Subject Group S4.02-00 "Forest Resource Inventory and Monitoring" and Subject Group S4.12-00 "Remote Sensing Technology." Volume I. IUFRO (International Union of Forest Research Organizations) XX World Congress, 6-12 August 1995, Tampere, Finland. EFI (European Forest Institute) Proceedings 7:27-41.

Vickery, P. D., M. L. Hunter, J. W. Wells. 1992. Evidence of incidental nest predation and its effects on nests of threatened grassland birds. Oikos 63:281-288.

Vuorela, A. 1997. Satellite-based land cover and forest classification of Finland. In Risto Kuittinen, editor. Finnish--Russian Seminar on Remote Sensing 29.8.-1.9.1994. Reports of the Geodetic Institute 97(2):41-51.

Wegge, P. 1979. Status of Capercaillie and Black Grouse in Norway. Proceedings of the International Grouse Symposium 1:17-26.

Wiens, J. A., N. Chr. Stenseth, B. Van Home, and R. A. Ims. 1993. Ecological mechanisms and landscape ecology. Qikos 66:369-380.

Willebrand, T. 1988. Demography and ecology of a black grouse (Tetrao tetrix L.) population. Dissertation. Uppsala University, Uppsala, Sweden.

Willebrand, T., and V. Marcstr[ddot{o}]m. 1988. On the danger of using dummy nests to study predation. Auk 105:378-379.
COPYRIGHT 2000 Ecological Society of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2000 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:KURKI, SAMI; NIKULA, ARI; HELLE, PEKKA; LINDEN, HARTO
Publication:Ecology
Geographic Code:4EUFI
Date:Jul 1, 2000
Words:8274
Previous Article:ANALYSIS OF POPULATION TRENDS FOR FARMLAND BIRDS USING GENERALIZED ADDITIVE MODELS.
Next Article:PREDATOR EFFECTS ON AN ASSEMBLAGE OF CONSUMERS THROUGH INDUCED CHANGES IN CONSUMER FORAGING BEHAVIOR.
Topics:

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