Patterns in species richness, food web structure and functional organization have been ob- served among biomes and across landscapes. To some extent, patterns may reflect varia- tion in occurrence or dominance of certain taxa in different biomes. Regional species pools may obscure effects of local habitat conditions on species richness (Kozár 1992a), especially in temperate ecosystems (Basset 1996), but few ecologists have addressed the extent to which the regional species pool may influence local species richness. Gering et al. (2003), Kitching et al. (1993) and Progar and Schowalter (2002) distinguished arthro- pod assemblages among sites within biomes that reflected regional gradients in environ- mental conditions. Various hypotheses have been proposed to account for apparent meta- community patterns at the biome and landscape level (e.g., Leibold and Mikkelson 2002, Price 1997, Tilman and Pacala 1993). Patterns include nested subsets (Summerville et al.
2002), checkerboards and various types of gradients (Leibold and Mikkelson 2002). Lei- bold and Mikkelson (2002) proposed a set of criteria to distinguish which pattern charac- terizes a given landscape. Coherence is the degree to which a pattern can be represented by a single dimension, species turnover is the number of species replacements along this dimension, and boundary clumping describes how the edges of species ranges are distrib- uted along this dimension. Presley et al. (2010) described additional patterns and demon- strated that combinations of patterns at fine spatial scales can aggregate to form different patterns at larger scales.
General functional groups are common to all terrestrial and aquatic biomes, e.g., graz- ing herbivores (depending on degree of autochthonous primary production in streams), predators, parasites, and detritivores, whereas other functional groups depend on par- ticular resources being present, e.g., sap-suckers require vascular plants, and wood bor- ers require wood resources. The proportions of the fauna that represent the different functional groups vary among biomes. Low order streams have primarily detrital-based
279 II. PATTErnS of CoMMUnITy STrUCTUrE
resources, and their communities are dominated by detritivores and associated predators and parasites. Other communities represent various proportions of autotroph functional groups (e.g., chemoautotrophs, ruderal, competitive, and stress-tolerant vascular vs. non- vascular plants) and heterotroph functional groups (herbivores, predators, detritivores) (see Chapter 11).
Different species compose these functional groups in different biomes. For example, the insect grazer functional group is composed primarily of moths, beetles, and tree crick- ets in broadleaved forests, moths and sawflies in coniferous forests (Schowalter 1995, Schowalter and Ganio 1999, Schowalter et al. 1981c), grasshoppers in grasslands and shru- blands (Curry 1994), and caddisflies and flies in aquatic communities (e.g., Hart 1992).
The predator functional group in terrestrial arthropod communities is dominated by a variety of arachnids, beetles, flies and wasps, whereas in aquatic arthropod communities this functional group is dominated by dragonflies, true bugs and beetles.
Among terrestrial biomes, species richness generally is assumed to increase from harsh biomes (e.g., tundra and desert) to grassland to forest, again reflecting differences in physical complexity, suitability and stability of the habitat (Bazzaz 1975, Tilman and Pa- cala 1993). However, this trend is not apparent for arthropods among communities where extensive species inventories are available (e.g., Table 9.1). Species richness is not always linearly related to primary productivity, and patterns likely depend on scale (Rosenzweig and Abramsky 1993, Tilman and Pacala 1993, Waide et al. 1999). Species richness of- ten declines above intermediate levels of productivity, perhaps because more productive communities are dominated by larger individuals that reduce habitat heterogeneity, or because more productive and stable communities favor competitive exclusion of some species by the best adapted species (Tilman and Pacala 1993). For example, continuous fertilization of permanent pasture at Rothamsted, U.K. since 1856 has resulted in changes in the species rank-abundance pattern from a log normal curve in 1856 to progressively more geometric curves by 1949 (Fig. 9.3) (Kempton 1979).
Functional group composition has not shown consistent differences among biomes (C. Hawkins and MacMahon 1989, Stork 1987). Detritivores generally represent a great- er proportion of the community in boreal forests, headwater streams, and other biomes characterized by accumulated organic material and a lower proportion in tropical forests, deserts and other biomes with little organic matter accumulation (Haggerty et al. 2002, Seastedt 1984). Wood borers occur only in forest or shrub ecosystems with abundant wood resources. Pollinators are more diverse in tropical forests and deserts, where plant diversity and isolation have led to greater reliance on insect and vertebrate pollinators, compared to temperate grassland and forest, and arctic biomes. Proportional representation of species and individuals among functional groups varies widely among canopy arthropod commu- nities in temperate and tropical forests, depending on tree species composition (Fig. 9.9) (V. Moran and Southwood 1982, Schowalter and Ganio 1998, 1999, Stork 1987).
At the landscape or drainage basin scale, patterns in species richness and functional group organization can be related to local variations in physical conditions. The history and geographic pattern of disturbance may be particularly important factors for deter- mining the variation in community structure. Polis et al. (1997a) concluded that the move- ment of organisms and resources among the interconnected community types comprising a landscape can contribute to the organization of the broader landscape community by subsidizing more resource-limited local communities. However, Basset (1996) found that the diversity in the trees of the tropical rain forest was related to five factors: numbers of young leaves available throughout the year, ant abundance, leaf palatability, leaf water
content, and altitudinal range. These data suggest that local factors may be more impor- tant determinants of local species diversity and community structure in complex ecosys- tems, such as tropical forests, than in less complex ecosystems, such as temperate forests.
Diversity of stream insects varies among riffle and pool habitats and substrate condi- tions (Ward 1992). Diversity generally is higher in running water with cobble substrates, with high oxygen supply and heterogenous structure, than in standing water with mud, sand or gravel substrates. Vinson and Hawkins (1998) reviewed six studies that compared species richness of stream insects over drainage basins. Species diversity varied with el- evation, which co-varied with a number of important factors, such as stream morphology, flow rate and volume, riparian cover, and agricultural or urban land use. In one study J. Carter et al. (1996) used multivariate analysis (TWINSPAN) to compare species com- position among 60 sites representing first-order (characterized by narrow V-shaped chan- nel, steep gradient, nearly complete canopy cover) to sixth-order (characterized by wide channel, low gradient, little canopy cover) streams over a 15,540 km2 drainage basin. They identified five communities that were distinguished largely by elevation. The highest spe- cies richness occurred in mid-order, mid-elevation streams that included species groups characterizing both higher- and lower-order streams.
Transition zones (ecotones) between community types typically have higher species richness, because they represent habitat variables and include species from each of the neighboring communities (e.g., Muff et al. 2009). Zhong et al. (2003) reported that the diversity of adult mosquito species was higher at sites that were surrounded by freshwater and salt-marsh than at those surrounded by either freshwater or salt-marsh alone. How- ever, Sabo et al. (2005) reported that riparian zones represent unique habitats that sup- port species not represented in the neighboring communities. Ecotones can move across the landscape as environmental conditions change. For example, the northern edge of fIG. 9.9 Functional group organization of arthropod communities in canopies of four old-growth conifer species at the Wind River Canopy Crane Research Facility in southwestern Washington. Data from Schowalter and Ganio (1998).
281 II. PATTErnS of CoMMUnITy STrUCTUrE
Scots pine, Pinus sylvestris, forest in Scotland moved rapidly 70–80 km northward about 4000 yrs ago then retreated southward again about 400 yrs later (Gear and Huntley 1991).
Sharp edges between community types, such as those that result from land use practices, reduce the value of this ecotone as a transition zone.
Patches at different stages of post-disturbance recovery show distinct patterns of spe- cies richness, food web structure, and functional group organization (see Chapter 10).
Species richness typically increases during community development up to an equilibrium, perhaps declining somewhat prior to reaching equilibrium (e.g., MacArthur and Wilson 1967, E. Wilson 1969). As the number of species increases, the number of species inter- actions increases. Food chains that characterize simpler communities develop into more complex food webs (E. Wilson 1969). Schowalter (1995), Schowalter and Ganio (1999) and Schowalter et al. (1981c) found that patches of recently disturbed temperate and tropical forests were characterized by higher sap-sucker/folivore ratios than were patches of undisturbed forests, even when the data were reported as biomass.
Shure and Phillips (1991) found that species richness and functional group composition were modified by manipulated patch size (Fig. 6.6). Species richness was lowest in mid- sized canopy openings (0.08–0.4 ha). Herbivore guilds generally had the lowest biomass in mid-sized canopy openings; omnivore biomass peaked in the smallest openings (0.016 ha) and then declined as opening size increased; predator biomass was highest in the control forest and smallest openings, and lowest in the mid-sized openings; and detritivore bio- mass was similar among most openings, but much lower in the largest openings (10 ha).
This pattern may indicate the scale that distinguishes communities characterizing closed- canopy and open-canopy forest. Smaller openings were influenced by surrounding forest, whereas larger openings favored species that were tolerant of solar exposure and altered plant conditions, e.g., early successional species and higher phenolic concentrations (Dudt and Shure 1994, Shure and Wilson 1993). Openings of intermediate size may be too ex- posed for forest species, but insufficiently exposed for earlier successional species. How- ever, species richness generally increases with habitat area (Fig. 9.10) (M.P. Johnson and Simberloff 1974, MacArthur and Wilson 1967), for reasons discussed below.
fIG. 9.10 Relationship between species richness and geographic area.
III. deteRmInAntS of communIty StRuctuRe
A number of factors affect community structure (e.g. Price 1997). Factors associated with habitat area, stability, habitat or resource conditions, and species interactions appear to have the greatest influence.