Ann. Zool. Fennici 37: 79–91 ISSN 0003-455X Helsinki 29 June 2000 © Finnish Zoological and Botanical Publishing Board 2000 The spacing of mature forest habitat in relation to species-specific scales in managed boreal forests in NE Finland Sakari Mykrä, Sami Kurki & Ari Nikula Mykrä, S. & Kurki, S., Section of Ecology, Department of Biology, University of Turku, FIN-20014 Turku, Finland Nikula, A., Finnish Forest Research Institute, Rovaniemi Research Station, P.O. Box 16, FIN-96301 Rovaniemi, Finland Received 13 September 1999, accepted 16 May 2000 Mykrä, S., Kurki, S. & Nikula, A. 2000: The spacing of mature forest habitat in relation to species-specific scales in managed boreal forests in NE Finland. — Ann. Zool. Fennici 37: 79–91. Modern forestry alters the structure of boreal forest landscapes, and this affects several forest species with different habitat requirements. Quantitative analyses of the structure of real landscapes are scarce, although this information will be needed in landscape- level planning and management applications. We investigated the occurrence and con- figuration of mature forest (> 80–100 year) in relation to five different extents (ranging from 20 to 7 854 ha) of circular landscapes around random points in managed forest environment in NE Finland. The habitat type was coarsely chosen according to require- ments of certain resident birds and mammals. We present an approach where species- specific home-range scales are related to the observed environmental heterogeneity pattern for assessing the potential of actual landscapes to maintain these species. As a result we found that there is a minority of landscapes where total area of mature forest was comparable to species’ home ranges. Irrespective of the scale examined, the aver- age proportion of habitat in a landscape was approximately 15%. Besides that the ma- ture forest was physically fragmented already on a scale relevant to individual animals, the variation in habitat proportion between landscapes decreased rapidly when the land- scape extent was increased. This decrease of variation occurs on surprisingly small scales; the forest landscape structure is unintentionally altered due to small-patterned land ownership, even age distribution of managed stands, and small variation in regen- eration patch size. In terms of related landscape metrics, we also compare our results with patterns achieved in randomly generated neutral landscape models. Mykrä et al. • ANN. ZOOL. FENNICI Vol. 3780 1. Introduction The central task in the conservation of biological diversity is to comprehensively maintain viable populations of naturally existing species in the foreseeable future. This applies also to managed forest environments; although, inevitably, in man- aged forests, solutions must acknowledge intrin- sic characteristics of areas and diverge spatially (Mykrä & Kurki 1998). Even though forest structure at stand level and its alteration seem to be most important for the majority of boreal forest species (e.g. Esseen et al. 1992, Raivio 1992, Haila 1994, Siitonen & Martikainen 1994, Petterson 1997, Wikars 1997), the population viability of a number of species is also critically connected to the structure of forest landscapes (Angelstam 1992, Hansson 1992, Andrén 1994, Kurki et al. 2000). Composition and configuration of forest landscape may affect these species either directly, through distribution of re- sources and habitats, or indirectly by altering in- ter-specific interactions in communities modified by changed landscape structure (e.g. Angelstam 1992, Dunning et al. 1992, Wiens et al. 1993, Kurki et al.1998). Although the ecological effects of boreal for- est management were studied intensively in Fennoscandia during the past two decades, quan- titative information on present forest landscape structure and also reliable estimates of natural large scale variation in former forest landscapes seem to be scarce. The characteristics of natural forest landscapes in Fennoscandia are well de- scribed in the recent review of Esseen et al. (1997). However, they also state that there is no quantita- tive data available regarding landscape level het- erogeneity in the past. The knowledge of natural stand level characteristics has accumulated splen- didly during recent years, but it is also necessary to understand the larger scale conditions to which forest species have adapted. The best documented changes in Fennoscandian forests caused by com- mercial forestry are the loss and fragmentation of natural old-growth forests together with the si- multaneous proportional increase of younger suc- cessional stages, and the rarity of forests with a significant old deciduous component (Hansson 1992, Östlund et al. 1997). In Finland at present, about 20.3% of productive forest land are forests older than 100 years, and 9.1% are older than 140 years. When the two northern provinces of Oulu and Lappi (which represent 42.6% of forest land in the country) are excluded, the corresponding percentages for southern Finland are 12.5 and 1.6, respectively (Anon. 1996). The loss of forests with old-growth character- istics has been shown to have deleterious effects on many forest-dwelling species in different taxa in Fennoscandia (Helle & Järvinen 1987, Virkkala 1987, 1991, Rassi et al. 1992, Angelstam & Miku- sinski 1994, Siitonen & Martikainen 1994, Edenius & Elmberg 1996, see also Esseen et al. 1997). In this paper, we examine the managed for- est landscape structure in relation to coarsely de- fined requirements of certain resident forest dwell- ing mammals and birds. For this purpose, we want to draw a distinction between natural old-growth forests of long continuity and a broader class of mature managed stands. Natural old-growth for- ests are nowadays rare, and being an obligatory habitat for many specialist species they always have a high conservation value as such. In spite of that, and regarding the occurrence of some of the declining species in relation to conditions pre- vailing particularly in managed forest landscapes, there is a positive statistical relationship between the fulfilment of habitat requirements and cover- age of mature stands. This division between natu- ral old-growth forests and mature or ‘old’ man- aged forests is partly congruent with the differ- ence in mentioned direct and indirect mechanisms that are behind population declines of forest spe- cies. Threatened and declining species are either tightly dependent on some particular resource combination materialising itself nowadays only in sparse natural forests, or their population de- clines are due to altered inter-specific interactions following structural changes in communities at larger scales (Kurki et al. 1997). Although, this is a rather broad generalisation because both the degree of specialisation and ability to respond to and compensate community changes varies greatly from species to species. It is also probable that most of the population declines result from both direct and indirect reasons. We investigated the occurrence and configu- ration of mature forest (growing stock volume ANN. ZOOL. FENNICI Vol. 37 • Mature forest and scale in NE Finland 81 > 100 m3 ha–1) in relation to spatial scale in a man- aged forest region in north-eastern Finland. The habitat type definition and this approach are based on Finnish and Scandinavian literature, which emphasises the positive relationship between the occurrence of mature forests and several forest dwelling vertebrates (Helle 1985, Rolstad & Wegge 1987a, 1989, Virkkala 1987, 1991, Angel- stam 1992). In this study, we obtained a general picture of the distribution and abundance of ma- ture forest habitat in managed forest landscapes in that region, and we interpret our results in rela- tion to literature-based information concerning home range scales of flying squirrel (Pteromys volans), siberian tit (Parus cinctus), siberian jay (Perisoreus infaustus), three-toed woodpecker (Picoides tridactylus) and capercaillie (Tetrao urogallus) (Haftorn 1973, Kokhanov 1982, Haila 1983a, Rolstad & Wegge 1987a, 1987b, Virkkala & Liehu 1990, Helle et al. 1994, Mönkkönen et al. 1997, Reunanen & Nikula 1998, Hanski 1998). On the individual level the home ranges of all these species contain certain specific resources (Rolstad & Wegge 1987b, Hågvar et al. 1990, Virkkala & Liehu 1990, Virkkala et al. 1991, Stenberg 1996, Hanski 1998, Reunanen & Nikula 1998), which on the one hand are not prevalently present in this artificially pooled > 100 m3 ha–1 habitat, and on the other hand, certainly occur to some extent in the matrix, too. Moreover, as it is evident that in- dividuals do not require area per se, but that area requirements of individuals are a function of the occurrence of relevant resources (Haila et al. 1989), we do not claim that the examined habitat exclusively represents the requirements of the discussed species. Nevertheless, the occurrence of these species can be assumed to be at least con- centrated on the examined habitat class in the study area. Furthermore, irrespective of theoretical prob- lems falling on the term “area requirements”, it is evident that individual level operational scale dif- fers between these species. Our aim is to portray an approach where information on species-spe- cific scales is related to the observed environmen- tal heterogeneity pattern for assessing the poten- tial of actual landscapes to maintain these spe- cies. In the discussion we also compare our re- sults with patterns achieved in randomly gener- ated neutral landscape models. 2. Material and methods 2.1. Study area Our 24 000 km2 study area (Kainuu forestry cen- tre) (Fig. 1) is located in northeastern Finland and lies in the northern boreal zone (Ahti et al. 1968). Of the total area, water covers about 11%, and 5% of the land area is covered by agricultural land and human settlements. The rest, approximately 84% of the total area, is ‘forestry land’, which is further divided, according to the forest manage- ment terminology, into ‘forest land’ (81%), ‘scrub land’ (12%) and ‘waste land (7%). The ‘forestry land’ in the study area is highly dominated by co- niferous forests and bogs. More than two fifths (43.8%) of forest land are privately owned whereas 39.6%, 13.6% and 3.0% are owned by the state of Finland, forest companies, and other owners, re- spectively. Nature reserves compose 3% of land area (Anon. 1996, Anon 1998). Modern forestry, based mainly on clear-cut- ting and artificial regeneration, started more widely after World War II, and thus far at least Fig. 1. Kainuu forestry centre in northeastern Finland. Dots on the map depict the spatial distribution of ran- domly located circular landscapes. Mykrä et al. • ANN. ZOOL. FENNICI Vol. 3782 65% of the ‘forest land’ has been regenerated once. During this period there has been a decreasing trend in the size of regeneration areas. Their typi- cal size has varied from one hectare to some tens of hectares, and nowadays more than 80% of cut- ting areas are smaller than 10 ha. The rotation time in the study area averages 100 years. 2.2. Landscape data and analyses We used classified Landsat images produced by the National Forest Inventory (NFI) (Tomppo 1993, 1996) to examine the occurrence of mature forest. For each pixel on forested land (25 m × 25 m land element) NFI produces an estimate of growing stock volume separately for Scots pine (Pinus sylvestris), Norway spruce (Picea abies), birches (Betula spp.) and other species as a com- bined class. Digital maps of non-forest lands (peat land, water area, agricultural land, roads and set- tlements) are used to separate non-forest areas from forest land (Tomppo 1993, 1996). We reclassified the landscape data by sum- ming the growing stock volume of different tree species for each pixel and defining pixels that had timber volume over 100 m3 ha–1 as ‘mature for- est’. In other words, the data were bisected to ma- ture forest and everything else. The 100 m3 ha–1 division limit refers approximately to class mid- point of the average total volume of young thin- ning forests (58 m3 ha–1) and advanced thinning forests (131 m3 ha–1) in the study area (Tomppo et al. 1998). Furthermore, timber volume correlates well with the age of the forest stand and the crite- ria used in our study is about in the class mid- point of the average timber volumes of age classes 61–80 years (80 m3 ha–1) and 81–100 years (113 m3 ha–1) (Tomppo et al. 1998). After reclassifying the landscape data, we ran- domly selected 120 points (Fig. 1) avoiding only large lakes and towns in the study area, and formed circular landscapes around them using five dif- ferent radii: 250, 500, 1 000, 2 000 and 5 000 metres. Corresponding areas covered were ap- proximately 20, 79, 314, 1 257 and 7 854 hec- tares. In seven cases landscape data did not cover the whole area inside the largest radius. They were therefore omitted, and the data were only used from 113 points. The area of mature forest (CA), number of mature forest patches (NP), mean patch size (MPS; ha), mean nearest neighbour distance (MNN; m) and largest patch index (LPI; (arealargest patch/arealandscape) × 100) were computed for each landscape using FRAGSTATS (McGarigal & Marks 1995). 2.3. Individual level operational scale of spe- cies The information on territory sizes and individual level home range scale of flying squirrel, siberian tit, capercaillie, siberian jay and three-toed wood- pecker were surveyed mainly from Fennoscandian literature, which makes it applicable also in this particular study area. The information on terri- tory size of flying squirrel is based on recent ra- dio-tracking data from southern Finland (Hanski 1998), and from studies conducted in the same region that we used in this study (Mönkkönen et al. 1997, Reunanen & Nikula 1998). According to these studies, flying squirrels require some 0.5– 3 ha of favourable habitat to be able to establish a territory. Total activity areas of an individual, however, may be even 5 to 10 times greater. Ac- cording to Hanski (1998) the average for this is 6.5 ha. The territory size of siberian tit is reported to range from 15 to 20 ha in coniferous forest habitats (Haftorn 1973, Virkkala & Liehu 1990). The home range size of an individual capercaillie varies from some tens of hectares to more than 100 ha, but in the case of this species more im- portant is that a minimum requirement for a vi- able lek is around 300 ha of suitable habitat (see Rolstad & Wegge 1987a, 1987b, Helle et al. 1994). Comparable published information on siberian jay and three-toed woodpecker seems to be surprisingly hard to find. In the case of the former, we rely on Kokhanov (1982). The mean size for five studied territories in the Kola Penin- sula was 52 ha and the range varied from 45 to 57 ha. For the latter species, the only piece of infor- mation concerning the home range scale seems to be the one presented by Haila (1983a) in his study on colonisation of islands by land birds in the lake Inari in northern Finland. He conducted land bird censuses on 41 islands with a size range of 0.5– 885 ha. The three-toed woodpecker occurred only on three islands, the smallest of them being 99 ha. ANN. ZOOL. FENNICI Vol. 37 • Mature forest and scale in NE Finland 83 The other two were 680 and 885 ha. Habitat types comparable with these ‘woodpecker islands’, however, occurred in substantial proportions on many of the smaller sized islands. Also, in the case of other species his results were in accord- ance with the abovementioned studies. The small- est islands from where siberian tit, siberian jay and capercaillie were found, were 17, 67 and 290 ha, respectively. Haila observed individual caper- caillies also on islands sized 22 and 54 ha, but the one with 290 ha area was smallest with more than one (3) individual. This may indicate lekking and/ or nesting. The three smallest landscape extents used in this study (20, 79 and 314 hectares) are almost equivalent with the home range scale hierarchy among these species. In the examination, land- scape extents were associated to species opera- tional scales and the fulfilment of each species’ requirements in terms of mature forest occurrence was estimated on landscape extent that was one level up from its home range scale. The largest scales (1 257 and 7 854 hectares) would be more pertinent in population level considerations, which are beyond the scope of this paper. Only the capercaillie lek is associated to 1 257 ha extent. 3. Results Since we pooled all the other land cover classes (including water and artifactitious classes), the habi- tat proportions are given in relation to total land- scape area on each scale, not relative to ‘forest land’ area. The average proportion of mature forest habi- tat in landscapes was about 15% throughout all examined scales (Table 1). The landscapes around randomly selected points seem to represent aver- age forest and they are congruent with the fact that 65% of the ‘forest land’ have been regenerated since the 1950s; on 31% of ‘forest land’ (21% of total area of forestry centre) forests are older than 81 years, and approximately 35% of ‘forest land’ (24% of total area of forestry centre) belong to develop- mental classes V (advanced thinning stands), VI (mature forests) and to low-yielding forests (Anon. 1996). This last class includes also the majority of natural old-growth forests in Kainuu area. Even if the mean proportion of mature forest habitat did not change with scale, the variation in habitat proportion between landscapes decreased rapidly when landscape radius was increased. The ranges of habitat percentage narrowed from 0.3– 63.3 on the smallest scale of 20 ha to 4.6–31.4 in the largest landscapes of 7 854 ha (Table 1). The decrease in CV% indicates further that among individual landscapes the values concentrated strongly around the average of 15% as the scale increased (Table 1). This can also be seen in the scale-specific distributions of landscapes in rela- tion to the percentage of mature forest (Fig. 2), where the occurrence of mature forests concen- trates on only a few classes on largest scales. In inspecting the relation of home range re- quirements and habitat area we disregarded the fact that habitat patches were relatively small and scattered throughout the landscapes, and assessed the landscape potential only with habitat percent- age (CA). The total area of examined habitat was comparable to home ranges of flying squirrel (2 ha) and siberian jay/female capercaillie (50 ha) on less than half of 20 ha (48%) and 314 ha (42%) landscapes, respectively. One fourth (25%) of 79 ha landscapes met the requirements of siberian tit Table 1. Proportion of mature forest (> 100 m3 ha–1; > 80–100 yr.) in relation to area of circular landscapes around random points (N = 113) in northeastern Finland. ————————————————————————————————————————————————— Radius Area Landscape % of landscapes % of landscapes (m) (ha) ————————————————— with > 30% habitat with > 60% habitat Mean Range CV% ————————————————————————————————————————————————— 250 20 15.1 0.3–63.3 95.4 15.0 0.8 500 79 14.7 0.6–50.1 77.5 11.5 0 1 000 314 14.2 1.4–39.3 63.4 7.1 0 2 000 1 257 14.4 2.6–39.1 53.4 4.4 0 5 000 7 854 14.5 4.6–31.4 38.6 0.8 0 ————————————————————————————————————————————————— Mykrä et al. • ANN. ZOOL. FENNICI Vol. 3784 (20 ha) in this respect, and regarding three-toed woodpecker/capercaillie male home ranges (100 ha) the percentage in 314 ha landscapes was 7%. Habitat area adequate for a viable capercaillie lek (300 ha) was found on 8% of 1 257 ha landscapes. These summed proportions are also illustrated in Fig. 2; arrows on relevant scales (see figure text for further explanation) indicate the subset of dis- tribution, where total habitat area/landscape equals or exceeds the area of home ranges. In terms of comparing our results of landscape metrics’ computations and patterns achieved in a randomly generated neutral landscape model (Andrén 1994), the most interesting relations can be seen in 20 ha landscapes (Fig. 3), and their results are presented and discussed in more de- tail. Number of patches, mean patch size and mean nearest neighbour distance show a clear non-lin- ear relationship with mature forest proportion on the 20 ha scale, and these relationships weaken with diminishing variation as the scale extends, and results of the largest landscapes show as a mere obscuring reflection of those observed on the small scale (Fig. 4). Even on the smallest spatial level with the greatest variation and highest mature forest per- centages the habitat mosaic was very small pat- terned, or physically fragmented on the great majority of landscapes; in 85% of 20 ha landscapes the proportion of mature forest habitat was less than 30% (Table 1), and among them the mean patch size varied between 0.06 and 2.69 ha (Fig. 4). However, one or two pixel patches were rela- tively common and they tend to decrease the mean patch size, even though their ecological signifi- cance is doubtful, as they do not necessarily devi- ate from their surroundings otherwise than hav- ing a standing volume just a bit over 100 m3 ha–1. Furthermore, in the 20 ha landscapes the mean patch size increased only slightly with the habitat percentage increment remaining under 3 ha until the threshold level of 60% in habitat coverage was reached. On that scale the number of patches first increased with increasing habitat in the landscape and peaked at about 25% (Fig. 3). All computed landscape metrics are interrelated and, therefore, it seems also characteristic that junctures of non- linearity coincide; increase and turning point of NP was mirrored by MNN, and MPS increase follows the peak in NP (Fig. 4). Fig. 2. The distribution of randomly located landscapes in relation to the percentage of mature forest at five different landscape extents. Each species is associated with a ‘focal area’ that is one level up from its home range scale. The starting point of each arrow indicates the lowest percentage of mature forest that would approximately respond to the home range requirements of species if the habitat occurred as continuous from the species point of view. Thereby, the summed percentages (y-axis) of classes in the right side of the arrow starting point indicate the ‘maximum’ proportion of suitable landscapes on a given scale. ANN. ZOOL. FENNICI Vol. 37 • Mature forest and scale in NE Finland 85 In order to elucidate further the possible ran- domness/aggregation in the spatial pattern of mature forest, we tested how the distribution of NP deviated from random on different scales. The NP was randomly distributed on 20 ha scale, but not on larger scales (Kolmogorov-Smirnov test with Poisson as a test distribution, on 20 ha scale Z = 1.182, p = 0.122, n = 115; on 79 ha scale Z = 2.717, p < 0.001, n = 115). Finally, we compared the relationship between LPI and habitat propor- tion in 20 ha landscapes with results achieved from a neutral model in Andrén (1994) (Fig. 5). In ran- domly generated neutral landscapes there is a noticeable juncture of non-linearity in LPI be- tween habitat proportions of 50% and 70%, and there was a trace of that also in our data. Com- pared with a pure random pattern, however, the result suggests certain spatial aggregation of ma- ture forest patches. There is, though, a particular disparity between sampling resolution of the land- scape data and the true resolution of environmen- tal heterogeneity (i.e. landscape grain size). Thus, the observed aggregations may in fact be false due to this disparity. This is further discussed in the final section. 4. Discussion 4.1. Environmental heterogeneity The proportion of the > 100 m3 ha–1 stands is, in general, much higher in natural forest landscapes than in managed ones, but in addition to that, also the scale and resolution of heterogeneity are dif- ferent. When interpreting these results about the relationship between spatial scale and variance in habitat proportion, we want to emphasise espe- cially that the scale and pattern of environmental heterogeneity, in particular the patch configura- tion of succesional stages have distinctly altered in the course of forest management in Fenn- oscandian forests. One could argue that this is more of a supposition since studies on the scale and pattern of heterogeneity in past landscapes of Fennoscandia are lacking. However, information derived from research on forests in Russian Carelia support our view (see e.g. Siitonen et al. 1994). It is obvious that the decrease of variation in land- scape composition and configuration as a func- tion of extending scale is a fundamental feature of every heterogeneous environment. We pre- sume, however, that a corresponding reduction in variance occurred on considerably larger scales in natural landscapes than in present managed forests. This change in heterogeneity is a by-prod- uct of small-patterned land ownership, regenera- tion areas small and uniform in size, and economi- cally optimal even age-distribution of forest stands, which has, according to Pukkala (1994), previously been a common target already at the level of individual forest holdings. Particularly in the case of NE Finland, Reunanen and Nikula (1998) propose that selective cutting applied in the first half of this century had a relatively low impact on forest landscape structure, and that this presently evident structural change is due to the altered methodology of commercial forestry Fig. 3. Comparison of number of patches (NP), mean patch size (MPS) and mean nearest neighbour distance (MNN) between our study and randomly generated model landscapes from Andrén (1994). Black dots describe the data of this study (left-hand side y-axis in NP). Open squares (and right-hand side y-axis in NP) show the results from Andrén (1994). In Andrén’s original paper the percentage of examined habitat varied between 0 and 100; due to the smaller variation in our data, only 0%–75% is shown. Contrary to the practice elsewhere in this paper, MPS is measured by the number of pixels and MNN in distance units (one unit = length of a pixel side). Mykrä et al. • ANN. ZOOL. FENNICI Vol. 3786 Fig. 4. Number of patches, mean patch size and mean nearest neighbour distance in relation to the proportion of mature forest in circular landscapes with five different radii. adopted since the 1950s. In our study area, and at the spatial level rel- evant to the individuals of the discussed species, the scale and pattern of environmental heteroge- neity is of artifactitious origin as the patch mosaic results from commercial forestry. Patchiness is an inherent trait of all environments in nature, but since patch structure is hierarchic and patches themselves are created by a given process with a particular frequency (i.e. single patches last for a particular period of time), an apparent graininess of any habi- tat mosaic cannot as such be defined as coarse or ANN. ZOOL. FENNICI Vol. 37 • Mature forest and scale in NE Finland 87 fine. Thereby, the scale dependencies of an organ- ism must be internalised into the grain concept, and only after that can this definition be made. Fur- thermore, in the literature there have been differ- ent connotations for ‘grain’ and even debate over the use of the term (Levins 1968, Wiens 1976, Addicott et al. 1987, Kotliar & Wiens 1990, Norton & Lord 1990, Wiens 1990). Due to the anthropo- genic cause behind patch mosaic in managed for- ests, we then refer to the ‘grain’ definition origi- nally presented by Levins (1968) where mosaic is fine grained for an individual if patches are smaller than the home range size and animals encounter several habitat types in their routine activity. The physical — but not necessarily functional — frag- mentation of mature forest habitat was evident al- ready in smallest landscape class (Figs. 3 and 5). Thus, as already the smallest of the species have their routine cruising distances in the hundreds-of- metres scale, the average patch mosaic in the ex- amined forest landscapes is fine-grained from the point of view of all discussed species. Although we refer to the forest age in our clas- sification, the well-being of old forest species is only rarely directly associated to the age of trees. Rather are they adapted to some combinations of structural elements, micro-habitats and commu- nity structures, which coincide in climax stages. The particular characteristics in habitat classifi- cation as well as the important scale (both the reso- lution and the landscape extent) are different for various species and depend on whether the ques- tion is proposed on the scale of an individual or a population (e.g. Haila 1990, Kotliar & Wiens 1990). Therefore, in the goal setting of biodiversity maintenance in a forest planning area, relevant issues are to acknowledge the important forest characteristics and to optimise the scale-depend- ent variation in their occurrence. Because of our simplified habitat classification relative to species’ needs it is obvious that there are suitable habitats for these species in the < 100 m3 ha–1 class as well as large enough but uninhabited areas in the ma- ture forest class. However, irrespective of this artificial 100 m3 ha–1 border, the observed spatial pattern in environmental heterogeneity suggests that the occurrence of individuals in the study area is restricted to large enough patches few in number, or that the individuals perceive the envi- ronment in fine-grained manner. The negative implication of coarse-grained response on the population density is apparent if large enough patches are a limiting factor. Similar reduction in density results from the fine-grained response too, if the species is territorial, and individuals must establish their territories in qualitatively variable patch mosaic. A fine-grained habitat mosaic may also decrease the fitness of individuals since sub- stantial proportion of their home range is sub-op- timal habitat, although this can be compensated with good moving ability (Tjernberg et al. 1993). All of the surveyed species have a good ability to move across forested landscape and, despite the observed habitat pattern, the individuals may in many cases experience the mature forest habitat as continuous at their own operational scales. In spite of this, however, there seems to be a rela- tively low proportion of landscapes which could respond to the habitat requirements of each spe- cies by containing an adequate total amount of mature forest. This applies particularly to siberian tit, three-toed woodpecker and capercaillie. Fig. 5. Comparison of largest patch index (LPI) be- tween this study and randomly generated model land- scapes from Andrén (1994). LPI is a percentage of largest patch area from the total area of landscape, and it reaches its maximum value if all of the exam- ined habitat is in one patch. Mykrä et al. • ANN. ZOOL. FENNICI Vol. 3788 4.2. Thresholds in non-linear population re- sponses The importance of critical thresholds in the pro- portion of suitable habitat has been emphasised in studies modelling the direct effects of habitat fragmentation on populations (Turner & Gardner 1991, Gustafson & Parker 1992, Andrén 1994). Above the threshold, the effects are caused solely by the decrease in habitat area, as predicted by the random sampling hypothesis (Haila 1983b). At the threshold and below, however, population density or community richness may respond dra- matically to only minor reduction in habitat area due to abrupt decrease in habitat connectivity and increasing significance of fragmentation effects. Later modelling has suggested further that thresh- old proportion is sensitive to habitat specificity and dispersal abilities of the target species (With & Crist 1995, Andrén 1996). The knowledge of non-linearities in relations between species viabil- ity and environmental gradients (measured, for example, by resource occurrence, habitat area, physical landscape structure etc.), and critical threshold stages in them would be useful in im- proving the precision of practical guidelines for biodiversity maintenance in managed forests (Mykrä & Kurki 1998). Haila (1995) assessed the practicability of using natural dynamics as a model for management and he also gave a correspond- ing approach of such threshold values. He em- phasised the identification of critical ‘qualitative junctures’ in natural non-linear processes, i.e. stages in which change is qualitatively important with only a slight change in system parameters, and addressed that an assessment is needed to re- veal whether these junctures occur or are lacking in silvicultural forests. In his review of literature on mammals and birds in mosaic landscapes, Andrén (1994) con- cluded that for these taxa, critical threshold levels are between 10% and 30% of suitable habitat. In spite of a probable species-specificity in threshold values, the fifteen per cent proportion of mature forest habitat in our study area is so low that habi- tat spacing can be argued to be important from an ecological standpoint. In terms of NP, MPS and MNN, the relations between landscape metrics and habitat proportion in the landscapes with a 250 m radius resembled the results achieved from ran- domly generated neutral landscapes (Andrén 1994, With & King 1997) (Fig. 3). This would suggest a random spatial distribution of habitat patches on the smallest examined scale. However, examina- tion of LPI (Fig. 5) reveals that as compared with the random pattern the > 100 m3 ha–1-pixels are indeed aggregated to a certain degree also on the smallest scale, and the values of single landscapes are closer to the LPI maximum. Maximal LPI (solid line in Fig. 5) is the level of utmost aggre- gation, because in that case NP is one. Compari- son between natural landscapes and the modelled random pattern, however, is complicated with the fact that in real landscapes the grain size is deter- mined ‘from the outside’ — random landscapes are a result of single pixels emerging or disappear- ing in a random fashion, but for example in our case the aggregation was partly due to the differ- ence between pixel size (25 m × 25 m) and aver- age area of forest regeneration (1–10 ha). With this resolution a five-hectare patch emerging or disappearing in landscape is innately an 80-pixel aggregation. Comparable to the random spacing pattern of habitat within landscapes was also the fact that NP distributed randomly between circu- lar landscapes on 20 ha scale, but not on larger scales. Furthermore, in the only landscape where the proportion of mature forest habitat exceeded 60%, the habitat formed one percolating, yet con- tinuous patch (NP = 1) (Fig. 3). This is in accord- ance with the percolation theory (e.g. Gardner & O’Neill 1991), which predicts that when the pro- portion of habitat in a randomly generated land- scape decreases below about 60%, one continu- ous habitat cluster breaks into discrete patches and size and connectivity of habitat patches can be- come an important feature of the landscape. 4.3. Conclusions Since the extraction of natural resources by hu- mans is indispensable in Finnish managed forests, the implementation of biodiversity maintenance must begin with precisely targeted area level and regional level management applications, even if many of the components of biodiversity once oc- curred prevalently in the forest landscape mosaic. ANN. ZOOL. FENNICI Vol. 37 • Mature forest and scale in NE Finland 89 Instead of applying moderate and standardised improvements extensively in managed forests, it would be more effective to systematically allo- cate various biodiversity maintenance responsibili- ties between forest planning areas by aggregating relevant resources spatially (Mykrä & Kurki 1998). Similarly, in a recent discussion on critical thresh- olds as management guidelines, caution in apply- ing any generalised threshold values was empha- sised (Andrén 1999, Mönkkönen & Reunanen 1999), since connectivity is a species-specific char- acteristic of a landscape, and average threshold values derived from the responses of a suite of species might be far below the habitat area needs of the most demanding ones. Future strategies and their application must be far better tailored accord- ing to varying objectives, and this also requires knowledge of local particularities at different scales. Guided by this thinking, we presented here a simple tool for assessing the possibility to main- tain sufficient amount of suitable habitat for dif- ferent forest dwelling mammals and birds. The spatial scale of our study covered areas from tens of hectares to several thousands of hec- tares. The planned cut in commercially managed forests is relatively fixed already at those small- est scales due to economical constraints. Even if the most important efforts in conservation biol- ogy are to stop habitat loss and to restore degraded habitats (Fahrig 1997), a really significant increase in the proportions of economically inferior but ecologically beneficial features in managed for- ests seems impossible without a truly profound change from a pure volume based thinking in wood production to strategies that rely more on product quality. Until this kind of change hap- pens, the most important question is: how should this fixed habitat composition of managed forests be spatially configured? ACKNOWLEDGEMENTS: This study was financially sup- ported by the Academy of Finland and the Maj and Tor Nessling and Kone foundations. We thank Erkki Tomppo and the staff of the NFI for providing the inventory data, and Paavo Kumpu for his assistance in the landscape data reclassification. Pekka Helle, Harto Lindén and Otso Suomi- nen have given valuable comments on earlier versions of this manuscript. Final improvements that rendered the pub- lishing of this paper possible are due to comments of two anonymous referees Yrjö Haila and Mikko Mönkkönen. Ms Marja Vieno revised the English of the manuscript. References Addicott, J. F., Aho, J. M., Antolin, M. F., Padilla, D. K., Rickhardson, J. S. & Soluk, D. A. 1987: Ecological neighborhoods: scaling environmental patterns. — Oikos 49: 340–346. Ahti, T., Hämet-Ahti, L. & Jalas, J. 1968: Vegetation zones and their sections in northwestern Europe. — Ann. Bot. Fennici 5: 169–211. Andrén, H. 1994: Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: a review. — Oikos 17: 355–366. Andrén, H. 1996: Population responses to habitat fragmen- tation: statistical power and the random sample hypoth- esis. — Oikos 76: 235–246. Andrén, H. 1999: Habitat fragmentation, the random sam- ple hypothesis and critical thresholds. — Oikos 84: 306– 308. Anon. 1996: Statistical Yearbook of Forestry 1996. — The Finnish Forest Research Institute. 352 pp. Anon. 1998: Finnish Statistical Yearbook of Forestry 1998. — The Finnish Forest Research Institute. 343 pp. Angelstam, P. 1992: Conservation of communities — the importance of edges, surroundings and landscape mo- saic structure. — In: Hansson, L. (ed.), Ecological prin- ciples of nature conservation: 9–70. Elsevier Applied Science, London. Angelstam, P. & Mikusinski, G. 1994: Woodpecker assem- blages in natural and managed boreal and hemiboreal forest — a review. — Ann. Zool. Fennici 31: 157–172. Dunning, J. B., Danielson, B. J. & Pulliam, H. R. 1992: Ecological processes that affect populations in com- plex landscapes. — Oikos 65: 169–175. Edenius, L. & Elmberg, J. 1996: Landscape level effects of modern forestry on bird communities in North Swed- ish boreal forests. — Landscape Ecology 11: 325–338. Esseen, P.-A., Ehnström, B., Ericson, L. & Sjöberg, K. 1992: Boreal forests — the focal habitats of Fennoscandia. — In: Hansson, L. (ed.), Ecological principles of na- ture conservation: 252–325. — Elsevier Applied Sci- ence, London. Esseen, P.-A., Ehnström, B., Ericson, L. & Sjöberg, K. 1997: Boreal forests. — Ecological Bulletins 46: 16–47. Fahrig, L. 1997: Relative effects of habitat loss and frag- mentation on population extinction. — Journal of Wild- life Management 61: 603–610. Franklin, J. F. & Forman, R. T. 1987: Creating landscape patterns by forest cutting: ecological consequences and principles. — Landscape Ecology 1: 5–18. Gardner, R. H. & O’Neill, R. V. 1991: Pattern, process, and predictability: The use of neutral models for landscape analysis. — In: Turner, M. G. & Gardner, R. H. (eds.), Quantitative methods in landscape ecology: 289–307. Springer-Verlag New York, New York, USA. Gustafson, E. J. & Parker, G. R. 1992: Relationships be- tween landcover proportion and indices of landscape spatial pattern. — Landscape Ecology 7: 101–110. Mykrä et al. • ANN. ZOOL. FENNICI Vol. 3790 Haftorn, S. 1973: A study of the siberian tit (Parus cinctus) during the breeding season. — Sterna 12: 91–155. [In Norwegian with English Summary]. Haila, Y. 1983a: Colonization of islands in a north-boreal Finnish lake by land birds. — Ann. Zool. Fennici 20: 179–197. Haila, Y. 1983b: Land birds on northern islands: a sam- pling metaphor for insular colonization. — Oikos 41: 334–351. Haila, Y. 1990: Toward an ecological definition of an is- land: a north-west European perspective. — J. Biogeogr. 17: 561–568. Haila, Y. 1994: Preserving ecological diversity in boreal forests: ecological background, research, and manage- ment. — Ann. Zool. Fennici 31: 203–217. Haila, Y. 1995: Natural dynamics as a model for manage- ment: is the analogue practicable? — Arctic Centre Pub- lications 7: 9–26. Haila, Y., Hanski, I. K. & Raivio, S. 1989: Methodology for studying the minimum habitat requirements of for- est birds. — Ann. Zool. Fennici 26: 173–180. Hanski, I. K. 1998: Home ranges and habitat use in the de- clining flying squirrel (Pteromys volans) in managed forests. — Wildlife Biology 4: 33–46. Hansson, L. 1992: Landscape ecology of boreal forests. — TREE 7: 299–302. Helle, P. 1985: Effects of forest fragmentation on bird den- sities in northern boreal forests. — Ornis Fennica 62: 35–41. Helle, P., Helle, T. & Lindén, H. 1994: Capercaillie (Tetrao urogallus) lekking sites in fragmented Finnish forest landscape. — Scand. J. For. Res. 9: 386–396. Helle, P. & Järvinen, O. 1986: Population trends of North Finnish land birds in relation to their habitat selection and changes in forest structure. — Oikos 46: 107–115. Hågvar, S., Hågvar, G. & Mønness, E. 1990: Nest site se- lection in Norwegian woodpeckers. — Holarctic Ecol- ogy 13: 156–165. Kokhanov, V. D. [Kohanov, V. D.] 1982: [On the ecology of siberian Jay in the Murmansk region]. — In: Zabrodin, V. A. [Zabrodin, V. A.] (ed.), [Ecology and morphology of birds in far north-east]: 124–137. Mos- cow. [In Russian]. Kotliar, N. B. & Wiens, J. A. 1990: Multiple scales of patchi- ness and patch structure: a hierarchical framework for the study of heterogeneity. — Oikos 59: 253–260. Kurki, S., Helle, P., Lindén, H. & Nikula, A. 1997: Breed- ing success of black grouse and capercaillie in relation to mammalian predator densities on two spatial scales. — Oikos 79: 301–310. Kurki, S., Nikula, A., Helle, P. & Lindén, H. 1998: Abun- dances of red fox and pine marten in relation to the composition of boreal forest landscapes. — Journal of Animal Ecology 67: 874–886. Kurki, S., Nikula, A., Helle, P. & Lindén, H. 2000: Effects of forest fragmentation and composition on breeding success of grouse. — Ecology 81. [In press.] Levins, R. 1968: Evolution in changing Environments. — Princeton University Press, Princeton, New Jersey. Li, H., Franklin J. F., Swanson, F. J. & Spies, T. A. 1993: Developing alternative forest cutting patterns: A simu- lation approach. — Landscape Ecology 8: 63–75. McGarigal, K. & Marks, B. J. 1995: FRAGSTATS: spatial pattern analysis program for quantifying landscape structure. — U.S. Forest Service General Technical Report PNV 351. Mykrä, S. & Kurki, S. 1998: ESC-Strategy for Rational Operationalization of Forest Biodiversity Maintenance in Finland. — Silva Fennica 32: 389–399. Mönkkönen, M., Reunanen, P., Nikula, A., Inkeröinen, J. & Forsman, J. 1997: Landscape characteristics associ- ated with the occurrence of the flying squirrel (Pteromys volans) in old-growth forests of northern Finland. — Ecography 20: 634–642. Mönkkönen, M. & Reunanen, P. 1999: On critical thresh- olds in landscape connectivity: a management perspec- tive. — Oikos 84: 302–305. Norton, D. A. & Lord, J. M. 1990: On the use of ‘grain- size’ in ecology. — Functional Ecology 4: 719–720. Östlund, L., Zackrisson, O. & Axelsson, A.-L. 1997: The history and transformation of a Scandinavian boreal forest landscape since the 19th century. — Canadian Journal of Forest Research 27: 1198–1206. Petterson, R. B. 1997: Lichens, invertebrates and birds in spruce canopies: impacts of forestry. — Doctoral the- sis, University of Umeå, Sweden. Pukkala, T. 1994: Metsäsuunnittelun perusteet. — Gummerus kirjapaino Oy. Jyväskylä. Raivio, S. 1992: Bird communities in fragmented conifer- ous forests: the importance quantitative data and ad- equate scaling. — Ph.D. thesis. University of Helsinki, Finland. Rassi, P., Kaipiainen, H., Mannerkoski, I. & Ståhls, G. 1992: Uhanalaisten eläinten ja kasvien seurantatoimikunnan mietintö [Report on monitoring of threatened animals and plants in Finland]. — Committee report, Ministry of Environment, 328 pp. [In Finnish with English Sum- mary]. Reunanen, P. & Nikula, A. 1998: Liito-oravan elinympä- ristöt ja maiseman rakenne Koillismaalla. — Metsä- hallituksen luonnonsuojelujulkaisuja, Sarja A, No. 93. Rolstad, J. & Wegge, P. 1987a: Distribution and size of Capercaillie leks in relation to old forest fragmenta- tion. — Oecologia 72: 389–394. Rolstad, J. & Wegge, P. 1987b: Capercaillie habitat: a criti- cal assessment of the role of old forests. — Proc. Int. Grouse Symp. No 4: 235–250. Rolstad, J. & Wegge, P. 1989: Capercaillie (Tetrao urogallus L.) populations and modern forestry — a case for land- scape ecological studies. — Finnish Game Research 46: 43–52. Siitonen, J. & Martikainen, P. 1994: Occurrence of rare and threatened insects living on decaying Populus tremula: a comparison between Finnish and Russian Karelia. — Scand. J. For. Res. 9: 185–191. Siitonen, J. & Martikainen, P., Kaila, L., Nikula, A. & ANN. ZOOL. FENNICI Vol. 37 • Mature forest and scale in NE Finland 91 Punttila, P. 1995: Kovakuoriaislajiston monimuotoisuus eri tavoin käsitellyillä metsäalueilla Suomessa ja Karjalan tasavallassa. — In: Niemelä, P. & Hannelius, S. (eds.), Monimuotoisuus metsien hoidossa: 43–63. Metsäntutkimuslaitoksen tiedonantoja. Vantaa. Stenberg, I. 1996: Nest site selection in six woodpecker spe- cies. — Fauna Norvegica, Ser. C, Cinclus 19: 21–38. Tjernberg, M., Johnsson, K. & Nilsson, S. G. 1993: Den- sity variation and breeding success of the Black Wood- pecker (Dryocopus martius) in relation to forest frag- mentation. — Ornis Fennica 70: 155–162. Tomppo, E. 1993. Multi-source National Forest Inventory of Finland. — In: Nyyssönen, A., Poso, S. & Rautala, J. (eds.), Proceedings of Ilvessalo Symposium on Na- tional Forest Inventories, Finland 17–21 August 1992: 52–59. The Finnish Forest Research Institute. Research Papers 444. Tomppo, E. 1996: Multi-source National Forest Inventory of Finland. — In: Päivinen, R., Vanclay, J. & Miina, S. (eds.), New thrusts in forest inventory. Proceedings of the Subject Group S4.02-00 ‘Forest Resource Inven- tory and Monitoring’ and Subject Group S4.12-00 ‘Re- mote Sensing Technology’. Volume I. IUFRO XX World Congress, 6–12 August 1995, Tampere, Finland. EFI, EFI Proceedings 7: 27–41. Tomppo, E., Katila, M., Moilanen, J., Mäkelä, H. & Peräsaari, J. 1998: Kunnittaiset metsävaratiedot 1990– 1994. Metsätieteen aikakauskirja. Folia Forestalia 4B/ 1998: 619–839. Turner, M. G. & Gardner, R. H. 1991: Quantitative meth- ods in landscape ecology: an introduction. — In: Turner, M. G. & Gardner, R. H. (eds.), Quantitative methods in landscape ecology: 3–14. Springer-Verlag New York, New York, USA. Virkkala, R. 1987: Effects of forest management on birds breeding in northern Finland. — Ann. Zool. Fennici 24: 281–294. Virkkala, R. & Liehu, H, 1990: Habitat selection by the siberian tit (Parus cinctus) in virgin and managed for- ests in northern Finland. — Ornis Fennica 67: 1–12. Virkkala, R. 1991: Population trends of forest birds in Finn- ish Lapland in a landscape of large habitat blocks: con- sequences of stochastic environmental variation or re- gional habitat alteration? — Biol. Conservation 56: 223–240. Virkkala, R., Heinonen, M. & Routasuo, P. 1991: The re- sponse of northern taiga birds to storm disturbance in the Koilliskaira National Park, Finnish Lapland. — Ornis Fennica 68: 123–126. Wiens, J. A. 1976: Population responses to patchy environ- ments. — Annual Review of Ecology and Systematics 7: 81–120. Wiens, J. A. 1990: On the use of ‘grain’ and ‘grain size’ in ecology. — Functional Ecology 4: 720. Wiens, J. A., Stenseth, N. Chr., van Horne, B. & Ims, R. A. 1993: Ecological mechanisms and landscape ecology. — Oikos 66: 369–380. Wikars, L.-O. 1997: Effects of forest fire and the ecology of fire-adapted insects. — Ph.D. thesis. University of Uppsala. Sweden. With, K. A. & Crist, T. O. 1995: Critical thresholds in spe- cies’ responses to landscape structure. — Ecology 76: 2446–2439. With, K. A. & King A. W. 1997: The use and misuse of neutral landscape models in ecology. Oikos 79: 219– 229.