Beringia diversity and P(fire)
South (no permafrost)
North America i r diversity and P(fire)
South (no permafrost)
diversity and P(fire)
Figure 3.1 Latitudinal and longitudinal biogeographic patterns of species diversity and fire frequency, and latitudinal gradients of environmental gradients and selected adaptive traits important to ecosystem properties in circumpolar boreal regions, based in part on Danell et al. (1996)
further suggests that gradients of plant chemical defense in boreal forests may be partly responsible for gradients of mammalian species richness.
Regionally, the number of bird specics of predominantly boreal coniferous forest pools varies between 49 and 63 across the Holarctic (Haila and Jarvincn 1990). As a consequence of the pronounced seasonality of the environment, a large proportion of the species are migrants; this varies from 46 to 54% of the regional pools of conifer birds. The number of tropical migrants is somewhat higher in the Nearctic than in the Palearctic (18-22 species, 35-37% of the pool, vs. 12 16 species, 19-33% of the pool, respectively, Haila and Jarvinen 1990). For reasons of geologic history, tropical migrants in the Nearctic have been evolutionarily in closer contact with tropical regions than those in the Palearctic (Mayr 1976; Monkkonen and Welsh 1994).
In contrast to neo-tropical migrants, the major groups of winter residents, with the exception of tetraonid grouse, have cogencric species on the two continents; some 10% of the conifer-specialist, winter-resident bird species are shared on both continents. Winter conditions form a bottle-neck in the population dynamics of the winter residents (Jarvinene et al 1977; Virkkala 1987, 1991).
The distribution of bird species across boreal landscapes corresponds to their feeding guilds and conditionally upon habitat, in particular canopy height, stand density and species composition. Stegman (1930) originally suggested that the regional bird fauna in the Eurasian continent might have evolved in concert with major vegetation types (see also Brunov 1980). Both numbers of birds and species richness increases with successional age in Nearctic and Palearctic boreal forests, but the proportion of neotropical migrants in the species poo! increases with successional age in the Nearctic but declines with succcssional age in the Palearctic (Helle and Nicmi 1996). Niemi (1985) suggests that the physical features of peatland habitats are selection pressures for morphological divergences among guilds in both Minnesota and Finland. Bird community composition is therefore related to the diversity of functional characteristics of the tree species, particularly with regard to growth form as a physical habitat element, seed production as a food source for some guilds, and tissue chemistry as it determines availability of insect prey for other guilds (Table 3.1). Some qualitative generalizations concerning the main guilds are warranted.
1. Foliage insectivores (and omnivores) such as warblers and finches include the most abundant forest birds in regional species pools (Haila et al. 1994). Foliage gleaners are a much more pronounced component in conifer bird assemblages in the Neartic than in the Palearctic (17-20 vs. 9-12 species, respectively, in regional species pools across the continents; Haila and Jarvinen 1990). Foliage gleaners in the Palearctic are mostly represented by the family Sylviidae and in the Nearctic by the family Parulidae. Significantly, the arctic warbler (Phytloscopus borealis) is the only sylviid warbler occurring in far northwestern North America in a restricted area around the Bering Strait and Interior Alaska, but has a distribution stretching across the northern Palearctic. The association of parulid warblers with conifer assemblages in the Nearctic is far stronger than that of sylviid warblers in the Palearctic (Helte 1988; Helle and Monkkonen 1990).
2. A few species of flycatchers are ubiquitous, but they occur mainly in mature forests that contain sufficient openings due to tree-fall gaps or open understories for their feeding habit to be possible.
3. The diversity of ground-feeding insectivores, mainly thrushes, in both the Palearctic and the Nearctic may depend in part on the heterogeneous habitat provided for insects by fallen logs and litter, and therefore on the maintainence of particular litterfail and decomposition regimes.
4. fn addition, several more specialized guilds are included in the conifer bird faunas, such as woodpeckers, tetraonids (herbivores) and birds of prey. Woodpeckers are a more species-rich group in the Nearctic than in the Palearctic. but in contrast to conifer-preferring Palearctic woodpeckers, Nearctic woodpeckers prefer deciduous trees. Thus, a change in the relative representation of conifer and deciduous tree species entails changes not only in the rates of nutrients cycling through litterfail (Tabic 3.1 and below), but also in the diversity of the assemblages of several avian guilds.
The guild structure of birds is quite diversified, and each guild includes several specialized feeders that can exert considerable influence on their prey (Wiens 1989). For example, in North America several warbler species track and partially control spruce budworm outbreaks (McNamee el al. 1981: Crawford et al. 1983; Holling 1992). In northern Europe the willow warbler (Phylloscupus troehilus, the dominant warbler) shows a functional response to Epirritu autumnata outbreaks (Enemar et al. 1984).
Most boreal birds influence ecosystem functions in ways that are fairly stable on average but flexible in detail (Haila and Jarvinen 1990). First, forest birds are spatially well distributed in that a large proportion of species in the regional species pool arc regularly present locally (Haila et al. 1994). Second, individuals shift feeding and breeding locations in response to seasonal and annual changes in resource availability and habitat. Third, the location of breeding pairs of a particular species varies considerably, and perhaps stochastically, from year to year in the forest landscape, and strictly deterministic models do not adequately predict the annual variability in the location of territories across years (Haila, Nicholls, Hanski and Raivio, manuscript). Finally, breeding birds use habitat elements in a spatially heterogeneous manner by establishing their home ranges from elements of different environmental types (Wiens 1989). Thus, stand and landscape heterogeneity as well as flexible and stochastic elements of habitat selection cause the "pressure" of birds on the rest of the forest system to be spread out.
The role of insects in boreal ecosystems has recently been reviewed by Holling (1992). Phytophagous insects, particularly the bark beetles and defoliators that attack conifers, are among the major sources of tree mortality in boreal regions. These include spruce budworm (Choristoneura fumiferana), sawfly (Neodiprion seriifer) and bark beetles (Dendroctonus ponderosae, Ips typographic). They typically show large oscillations in population levels, with spreading-contagion outbreaks occurring approximately ever few decades. Because the intervals of these outbreaks are usually less than the lifetimes of most major tree species, defoliators and bark beetles can prevent the attainment of stable equilibrium communities and are thus major agents responsible for maintaining variability and diversity in boreal regions, from stand to landscape levels (Holling 1992).
A hypothesis of insect outbreak dynamics in boreal regions has been developed by McNamee el ai. (1981), Isaev et al. (1984) and Berryman el a!. (1987). This hypothesis emphasizes the interactions of population, ecosystem and climatic processes at multiple scales that are responsible for outbreaks. Insect populations are sustained at a low, unstable equilibrium by the resistance of individual trees, the weather and avian predators until a number of factors coincide to allow an outbreak to be released. These factors include critical densities of target tree species in a given stand as well as the age structures of the tree population as determined by successiona! status (Morris 1963; Clark et ai 1979), the weather, either acting directly on the insect population (Holling 1992) or indirectly on the food quality of the host plant (Campbell 1989), and ontogenetic changes in foliage chemistry, which in turn partly depend on soil chemistry (Kemp and Moody 1984).
Once an outbreak is released in a given stand, it can spread to neighboring stands across the landscape; at this point, the spatial arrangement of susceptible stands within dispersal distance of the insect becomes a significant factor in outbreak dynamics (Holling 1992). Once an outbreak achieves spreading contagion proportions, the original factors that triggered the outbreak in the first stand (weather, etc.) are overridden by how landscape structure determines the distribution of potential host species. This release from control by the local conditions that initiated the outbreak is the reason why such insects cause significant tree mortality across the landscape. Finally, the population dynamics of avian predators determine how synchronous the outbreaks are across the landscape (Dowden et at. 1953; Crawford et at.; Holling 1992); these in turn depend on the availability of suitable forest structures for breeding habitat, as noted above. The interactions of multiple factors at multiple scales produces a variety of temporal dynamics, from asymptotic stability to multiple oscillations to chaos (McNamee et al. 1981).
Outbreaks of phytophagous insects can be significant regulators of forest primary production, particularly by releasing understory trees of species or age classes not susceptible to outbreaks, and by returning nutrients to the forest floor in easily decomposible frass rather than recalcitrant coniferous needle litter (Mattson and Addy 1975).
Ants are an important component in forest ecosystems, particularly in the Palcarctic. Competitively dominant mound-building ants of the genus Formica form colonies that may cover tens of hectares of forest and have a strong impact on the occurence of other ants, other ground-living arthropods, and soil nutrient and water relationships (Rosengren et al 1979; Reznikova 1983; Savolainen et al. 1989; Punttila et al. 1994). Disturbances such as wildfires usually destroy the colonies of Formica ants, and the ensuing succession of local ant communities is influenced by species interactions and habitat heterogeneity (Punttila et al. 1994). Formica colonies, and even single nests, can reach several decades in age and have a strong influence on nutrient cycling within the foraging area, partly through the direct effects of the colony on soil structure and partly by tending aphid colonies that collect sap from nearby plants (Rosengren et al. 1979).
An interesting group of organisms in post-wildfire forests are insects adapted to track forest fires; these number tens of species. Some of them are obligately "pyrophilous", and are endangered in regions where fire suppression is efficient (Wikars 1992; Muona and Rutanen 1994). Some of the species seem to be primarily attracted by the warm microclimate of open burned areas and clearcuts (Ahnlund and Lindhe 1992). Numbers of ground-living arthropods seem generally to be higher in heterogeneous successional forest stages than in mature forests and old growth (Niemela et al. 1988; Punttila et al. 1994), and within-stand heterogeneity of litter types is also important to the local diversity of these arthropods (Niemela et al. 1992; Niemela et al. 1994b).
The extreme fluctuations of animal populations are among the more striking features of the boreal forest. Such fluctuations are a temporally dynamic aspect of biodiversity. Ten years appears to be the dominant cycle length for a wide variety of mammals and birds in North America (Keith 1963; Finerty 1980; Hrlien and Tester 1984), with a 4-year cycle being dominant in Fennoscandinavia (Hansson 1979), leading one to suspect that they are a general feature of boreal forests. The cycles are noted for their extreme amplitude, leading to apparent "boom-and-bust" fluctuations in local populations, and even chaotic behavior (Schaffer 1984; Hanski et al. 1993). Cycles of herbivores may result in differential survival of their preferred food species, such as balsam fir, aspen and birch (Stenseth 1977; Hansson 1979; Haukioja el al. 1983; Bryant and C'hapin 1986; Mclnnes el al. 1992), as well as their predators, such as warblers that prey upon budworm, or the Canada lynx (Lynx canadensis) that preys upon small mammals (Keith 1963). When the cycles occur in migrating populations such as caribou (Rangifer larandus), they have a spatial as well as a temporal aspect.
Although climatic cycles may trigger or even stabilize and synchronize population cycles (Sinclair et al. 1988, 1993), there is some evidence that the cycles are also caused by time delays in the interactions of several trophic levels, particularly as these delays affect the flow of limiting energy and nutrients between trophic levels (Keith 1963; Keith and Windberg 1978; Peterson et al. 1984; Krebs et al. 1986; Pastor and Naiman 1992). The consequences of extreme population cycles arc thereby ramified throughout the ecosystem.
Landscape-scale diversity in the boreal forest derives from the characteristic natural heterogeneity in forest patch structure, especially the contrast in age-class distribution and species composition between adjacent patches. This landscape pattern is driven by biotic and ecosystem responses to disturbance regimes operating on the landform and soil template.
Topography is the major environmental constraint on which smaller-scale features are superimposed. This is reflected in the biogeographic pattern of plant and animal assemblages: typical species of the steppe zone of inner Asia and northern handwood stands in eastern North American occur on favorable, southern slopes and intersperse boreal habitats (Chernov 1975; Pastor and MladenotT 1992). Further north, aspect also influences the distribution of permafrost, which in turn determines tree species distribution, nutrient cycling, and productivity (Viereck et al. 1983).
Natural disturbances are major processes maintaining the coexistence of species and diversity of the boreal zone. These range from stand level tree-by-tree replacement to landscape-scale disturbances such as fire. The former process occurs everywhere in mature or maturing stands, and in some environments is the predominant mechanism of forest regeneration (Hofgaard
1993; Kuuluvainen 1994; Mladenoff 1987}. In some parts of the boreal zone, heavy winds and river meandering also predominate as disturbance factors (Jeffrey 1961; Gill 1973; Sprugel 1976; Syrjanen el al. 1994).
Fire is the dominant form of disturbance in the global boreal forest (Payette 1992). Characteristic disturbance return intervals, size distributions and intensity produce spatially and temporally varying patterns (Heinselman 1973, 1981; Johnson 1992). Single fires may be very large (1000 10 000 ha), but spatial variation in intensity, sites and regeneration responses also produce patch-responses at smaller scales of 1-1000 ha nested within the original disturbance (Baker 1989; Payette et al. 1985, 1989; Johnson 1992; Payette 1992; Mladenoff et al. 1993). Temporarily, medium-scale (10-100 years) climatic variation that produces droughts also maintains fire regimes (Heinselman 1981; Clark 1989). Consequently, in the southern North American boreal forest, a steady-state patch mosaic does not exist (Heinselman 1981; Baker 1989).
Another important agent of disturbance with particular interest for the relation of diversity and ecosystem processes is the beaver (Castor canadensis). By building dams, beavers create wetland complexes that alter hydrologic regimes, energy flow and water chemistry (Hodkinson 1975; Naiman et al. 1986). Beaver ponds and associated wetlands can cover at least 13% of the land area in boreal landscapes (Johnston and Naiman 1990a). Browsing by beaver along the riparian corridor of their ponds also alters forest composition and successional pathways (Johnston and Naiman 1990b). With pond occupation and abandonment, the landscape influenced by beaver is temporarily dynamic and, as with fire, a steady-state of wetland distribution is not achieved (Gill 1972; Pastor et al. 1993b).
Tree species diversity with respect to distance and mode of seed dispersal and reproduction (Table 3.1) is important in forest regeneration and the development of landscape diversity following disturbance. The vegetation development following a particular "disturbance type" varies according to the intensity and scale of the disturbance, previous site history, tree species composition (proportions of deciduous and coniferous trees), dispersal and germination traits of different species (Pastor and Mladenoff 1992; Zasada et al. 1992; Table 3.1),and forest age. Several characteristics of boreal forests change in a fairly regular fashion with age, but succession is characterized by cyclic and chaotic as well as linear patterns (Gill 1972; Wein and El-Bayoumi 1983; Pastor and Mladenoff 1992; Mladenoff and Pastor 1993). The understory vegetation is dominated by grasses and herbs in young forest stages, but mosses and dwarf shrubs typical of boreal forests take over following canopy closure (Esseen et al. 1992), and species richness declines, particularly on poor sites (Tonteri 1994). The forest stand becomes phvsiognomically more monotonous, dominated by a few species of trees and ground layer plants, but at the same time small-scale ecosystem hetcro-
geneity increases, such as the distribution of tree-falls and coarse woody debris pools. These arc important for a suite of specialized fungi, plants and animals.
The continuous production and maintenance of habitat heterogeneity on many scales, including a variety of age classes with their characteristic species (Solbrig 1993), seems to be important here. First, rare animals are often associated with specific microecosystem processes in, for instance, decaying wood (Kaila et al. 1994). They thus depend on a continuous production and reproduction of suitable microsites within the dispersal radius of new generations of individuals, and require habitat continuity on a species-specific scale (Siitonen and Martikainen 1994). Second, some lichens and bryophytes seem to have narrow environmental niches and long regeneration times, and they thus require stable conditions within old-growth stands for several decades (Esseen et al. 1992). For birds and mammals, habitat heterogeneity may be fine-grained on the individual scale: they continuously utilize several types of habitats within their home ranges (Hansson 1979; Hanski and Haila 1988; Haila et al. 1994). Dynamically, these specific requirements for habitat heterogeneity at different scales may lead to two alternative situations, either strict metapopulations or source ■ sink population dynamics. The second alternative seems more probable because most forest species seem to be widely distributed, albeit in varying numbers, across forest habitat and age classes. Heterogeneity in ecosystem dynamics at multiple scales therefore both depends upon and maintains the diversity of the forest biota.
Was this article helpful?