Fragmentation is the conversion of contiguous habitat into patches of different habitats or land uses. Habitat fragmentation is especially deleterious to species adapted to relatively stable ecosystems (e.g., Samways 1995) and to rare species (Summerville and Crist 2001). Such species usually are less adapted to rapid or long-distance dispersal and may be less able to recolonize vacant or new habitats (resulting from disturbance or climate change) across inhospitable patches, compared to ruderal species adapted to long-distance colonization of disturbed habitats (St. Pierre and Hendrix 2003, Powell and Powell 1987; see Chapter 5). Furthermore, insects will not be able to colonize new habitat patches successfully until their hosts are established.
Old-growth (500-1000-year-old) conifer forests in Pacific Northwestern North America were substantially fragmented by clearcut harvesting over a 50-year period (1940-1990). The forest landscape changed from about 75% old-growth to about 75% stands <50 years old. A significant proportion of species associated with old-growth forest now exist as relatively small, isolated, and declining populations in a matrix of apparently inhospitable young forest (N. Christensen et al. 2000). Schowalter (1995) found that 70% of arboreal arthropod species in old-growth conifer forests in western Oregon were not present in adjacent young (20-year-old) conifer plantations. Predators and detritivores were particularly affected. Similarly, Powell and Powell (1987) found that flower visitation by male euglossine bees declined following forest fragmentation, even in the 100-ha fragment size, and was proportional to fragment size, indicating that very large areas of forest are necessary to maintain viable population sizes for some species.
Whereas fragment size affects persistence of demes, the degree of fragment isolation affects colonization. Steffan-Dewenter and Tscharntke (1999) demonstrated that abundance of pollinating bees and seed production declined with increasing isolation (distance) of experimental mustard, Sinapis arvensis, and radish, Raphanus sativus, plants from intact grassland in Germany.
Krawchuk and Taylor (2003) studied patterns of abundance of three dipter-ans, Wyeomyia smithii (Culicidae), Metriocnemus knabi (Chironimidae), and Fletcherimyia fletcheri (Sarcophagidae) inhabiting pitcher plants, Sarracenia purpurea, in western Newfoundland, Canada. For all three insect species, habitat configuration (patch size and isolation) was more important than the total area of habitat, but the relative importances of patch size versus isolation changed with spatial scale. Patch size was more important at the scale of movement and survival of individuals, whereas patch isolation was more important at the scale of matrix configuration and metapopulation dynamics.
Edges between patches are particularly pronounced in anthropogenic landscapes and affect dispersal of many species. Natural gradients of climate and geology interacting with disturbances produce relatively large patches, with broad transition zones (ecotones) between patches that dampen interference by one patch on environmental conditions of another. By contrast, human land use practices tend to produce smaller patches with abrupt edges (e.g., distinct agricultural monocultures within fenced boundaries, plowed edges against grasslands, harvested and regenerating plantations against mature forests, and greater edge density measured as edge perimeter (m) per ha) (e.g., Radeloff et al. 2000).These distinct edges substantially influence environmental conditions of the adjacent patches. For example, an edge of tall trees along an abrupt boundary with an adjacent plantation of short trees is exposed to much greater insolation and airflow, depending on edge orientation, leading to higher temperatures, lower humidities, and greater vulnerability to windthrow than prevailed when the edge was buffered by forest. J. Chen et al. (1995) discovered that microclimatic gradients extended 180-480 m into old-growth Douglas fir, Pseudotsuga menziesii, forests from clearcut edges, affecting habitat conditions for associated organisms. They concluded that forest patches <64 ha would be completely compromised by external environmental conditions (i.e., would be characterized entirely as edge habitat rather than as interior forest habitat). Similarly, grasslands overgrazed by livestock within fenced boundaries expose soil to desiccation, leading to death of surrounding vegetation and an increasing area of desertification (e.g., Schlesinger et al. 1990, see Fig. 2.8f).
Insects are sensitive to these edge effects. Roland and Kaupp (1995) found that transmission of nuclear polyhedrosis virus was reduced along forest edges, prolonging outbreaks of the forest tent caterpillar, Malacosoma disstria. Ozanne et al. (1997) documented lower abundances of Psocoptera, Lepidoptera, Coleoptera, Hymenoptera, Collembola, and Araneae and higher abundances of Homoptera and Thysanoptera at forest edges compared to interior forest habitats. Schowalter (1994, 1995) reported that these two groups of taxa generally characterized undisturbed and disturbed forests, respectively. Haynes and Cronin (2003) found that planthoppers, P. crocea, accumulated along edges, compared to the interior, of prairie cordgrass patches adjacent to mudflat but not patches adjacent to nonhost grasses, reflecting lower rates of dispersal across inhospitable mudflats (see Fig. 7.3). Similar results were found for understory insectivorous birds in tropical forest, suggesting that outbreaks of some insects could be more likely in fragments from which predators have disappeared (Sekereioglu et al. 2002). Remnant patches of natural habitat also are highly vulnerable to influx of nonindigenous species, from neighboring patches, that may compete with, or prey upon, indigenous species (Punttila et al. 1994).
Effects of edge density on the landscape can change during the course of population growth and decline. Radeloff et al. (2000) found that correlations between landscape patterns and jack pine budworm, Choristoneura pinus, population size varied over time, with proportion of jack pine, Pinus banksiana, and edge density (sum of perimeter length for land use classes per ha) positively correlated up to the peak of the outbreak, but edge density negatively correlated during population decline. These results probably reflect the more suitable resources represented by pollen cones that are more abundant on edge trees and the greater abundance of avian predators and the primary wasp parasitoid, Itoplectis con-quisitor, along edges.
Fragmentation does not affect all species equally or all negatively. Tscharntke (1992) reviewed studies that examined responses of several insect species to differences in reed, Phragmites australis, quality in fragmented (agricultural) and unfragmented (nature reserve) wetlands. Reeds in small patches had thinner shoots but more leaves than did reeds in large patches. Two chloropid flies, Lipara spp., that depend on thin shoots survived only in small patches or in the unmown edges of large patches. However, the stem-boring noctuid moth, Archanara ger-minipuncta, that depends on thick shoots persisted only in large patches. Shoot damage caused by this moth created necessary habitat for >20 other herbivores, saprovores, and parasitoids. For example, the gall midge, Lasioptera arundinis, survived only in the side shoots induced by A. germinipucta damage, making this midge equally dependent on large patches. Tscharntke (1992) calculated that survival of local populations of A. germinipuncta requires at least 180,000 individuals or at least a 2-ha area.
Fragmentation of natural ecosystems usually is associated with homogeniza-tion of vegetation patterns. Widespread planting of commercial crops and suppression of natural disturbances have eliminated much of the diversity of vegetation patches characterizing natural landscapes. In a diverse landscape, outbreaks of particular demes most often would be confined to patches of susceptible vegetation. Agricultural and forested landscapes have become more conducive to expansion and regionwide outbreaks of adapted species (Schowalter and Turchin 1993).
Was this article helpful?