Coastal systems can be defined as the general marine region extending from the shore out across the continental shelf, slope and rise (Brink 1993). Other chapters in this volume focus on various portions of the coastal system (Twilley et al., Chapter 13; Done et at., Chapter 15). Our review of impacts on biodiversity considers coastal regions generally; the latter portion of the chapter focuses more specifically on rocky iniertida] and shallow subtidal communities where more complete information exists about the likely consequences of changes.
Coastal systems are the source of over 75% of commercial seafood landings (FAO 1991). The capture or collection of selected species for food and other uses is the most obvious human impact in marine habitats. Besides causing severe reductions in commercial landings (NOAA 1993), overexploitation of species often leads to (1) drastic changes in the species composition of communities (e.g. Beatley 1991; Castilla 1993; Castilla et al. 1994), (2) serial overexploitation, whereby new species are added to a "fishery" as the abundance of fished species is reduced below commercial viability (Dugan and Davis 1993; Weber and Gradwohl 1995), (3) extensive "by-catch" mortality, which is caused by the unintentional capture of species that associate with target species (Andrew and Pepperell 1992), and (4) dramatic shifts towards the dominance of smaller individuals in the population size structure of the remaining target populations. The indirect impacts of such wide-spread manipulations are rarely studied for most commercially exploited ecosystems, but where they arc well studied, the effects are shown to be extensive. For example, by excluding humans from a stretch of coast in central Chile, Castilla and Duran (1985; see also Duran and Castilla 1989) showed that human harvesting of Conchoiepas concho-lepas, a predatory gastropod, had substantial direct and indirect effects on the community structure of rocky intertidal shores.
Species introductions, another potentially serious threat to biodiversity, are most obvious in estuarine habitats where transport of propagules of non-native species in the ballast water of ships appears to be a common means of introduction (Carlton and Geller 1993). Documentation of species introductions on open coasts has been less frequent, but the distribution of many cosmopolitan species implies that the species composition of these open coastlines has changed considerably since humans have been traveling among the continents (Carlton 1989; Carlton and Hodder 1995).
Habitat loss or damage is a common human impact in coastal systems. Dredging and trawling can disrupt soft bottom communities by (1) destroying sediment structure, (2) increasing particulates in the water column which can clog the feeding apparatus of filter-feeding invertebrates, and (3) increasing the turbidity of the water, thereby reducing the amount of light reaching plants (Riemann and Hoffmann 1991; Norse 1993). Use of dynamite in shallow-water habitats is still an accepted practice for the collection of some bivalves (Lithophaga lithophaga), with devastating consequences for the entire community (Fanelli et al. 1994). Human trampling in highly used intertidal areas can alter species composition (Povey and Keough 1991; Brosnan and Crumrine 1994). While the destruction or modification of marine habitats is typically local in scale, it can have much larger-scale consequences. For example, if the affected area is occupied by a local population that serves as a net source for propagules in a region, an apparently local, small-scale disturbance could have devastating consequences for species abundance at a much large scale. While several studies show that specific areas can provide critical habitat such as nurseries, spawning grounds, or specialized habitats (e.g. Johannes 1979; Cowen 1985; Chambers 1991; Clark 1994), for the vast majority of coastal species we do not know if or where such areas exist.
Reduction or modification of habitat quality from pollution is another influence which is likely to grow as human population size increases in coastal areas. Chemical pollution can take the forms of either acute levels of nutrient input or of chronic inputs of low-level toxins such as pesticides and heavy metals (GESAMP 199Í). Other forms include thermal pollution (e.g. outfall of power plants), terrestrial sediment runoff, and possibly even noise pollution.
Human-induced climate change will most likely influence habitat quality as water temperatures, upwelling regimes and storm regimes change (Bakun 1990; Ray et al. Î992; Castilla et al. 1993; Lubchenco et al. 1993). Furthermore, the depletion of atmospheric ozone is likely to have detrimental effects on phytoplankton production in some regions (Smith et al. 1992) and may also influence larval viability. While some studies link modifications of the abundance and distribution of some coastai spccies to recent temperature increases (Barry et al. 1995; Roemmich and McGowan 1995a, b), the immediate, short-term impact of climate change on coastal systems is still largely unknown (Navarrete et al. 1993; Paine 1993).
Most of the consequences of changes in diversity resulting from the impacts summarized above are not well known. However, many involve loss or addition of species, or changes in relative abundance or distribution of species. The consequences of such changes to community structure in some of these communities are relatively well understood and provide insight into possible general effects. In the next section, we summarize some relevant studies.
14.3 SHALLOW-WATER HARD-BOTTOM COMMUNITIES: MODEL SYSTEMS FOR EVALUATING CONSEQUENCES OF DIVERSITY LOSS
Experimental work in rocky intertidal and shallow subtidal ecosystems has proven particularly fruitful due to the small scale at which interactions occur, the ease of species manipulations, and the rapid turnover time (Connell 1974; Paine 1977). Because comparable experiments have been performed in these habitats on rocky shores of most continents (e.g. Paine 1994; Menge 1995), insights about the extent of generality are possible. Our survey focuses on studies in which species manipulations could be used to develop an understanding of the relationship between species diversity and other ecosystem properties. For broader surveys of shallow hard-bottom studies of community dynamics, see VanBlaricom and Estes (1988), Hairston (1989), Menge and Farrell (1989), Menge el al. (1994), Paine (1994), Estes and Duggins (1995), Menge (1995) and Underwood and Chapman (1995).
At present, the link between the role a species or group of species plays and ultimate ecosystem-level processes is generally unknown. In coastal systems, the relative continuity and huge geographic range of water masses, and the open nature of most populations, makes it difficult to delineate the boundaries of ecosystems and to judge the consequences of local changes in diversity. However, changes in ecosystem properties such as biogeochemical cycling, productivity and energy flow are most likely linked to changes in diversity through the structure and regulation of communities. Species that exert a strong influence on the regulation of the density, abundance and diversity of other organisms in the community are more likely to have an impact on ecosystem properties than those species having a weak influence on overall community structure. Furthermore, we would expect that some community components will have a stronger influence on ecosystem proper ties than others (e.g. changes in kelp abundance should influence ccosystem productivity more than changes in abundance of algal turfs). Both connections (functional roles of species to community structure, and community structure to ecosystem properties) are very important to the evaluation of how threats to biodiversity influence ecosystem properties.
Several components of diversity underlie the relationship between diversity and the functioning of communities and ecosystems. These include the number and relative abundance of species, the specific kinds or identities (and thus characteristics) of individual species, and the similarity among species (Chapin et al. 1995). Although the number of species may directly affect ecosystem properties (e.g. Naeem et al. 1994; Tiiman and Downing 1994), experimental data are currently inadequate to address how species number influences community regulation in marine systems. In this analysis, we consider two components of diversity: the overall impact of a functional group on community and ecosystem processes, and the relative contribution of individual species to that overall effect.
To understand the full implications of changes in diversity, it is necessary to consider both individual species and groupings of ecologically similar species, here termed a "functional group" (Chapin et al. 1995). Ecologically similar species are those occupying a comparable "ecological role" in a community or ecosystem (e.g. Paine 1980). This definition includes classification by resource use (i.e. guilds, Root 1967), but also other potentially important roles because an assessment of the consequences of a species loss must examine the total influence of a species within a community. Classes of potentially important functional roles include physical structuring roles (e.g. habitat-formers, space-occupiers), trophic roles (e.g. predators, herbivores, filter-feeders, primary producers, detritivores), chemical roles (e.g. nitrogen-fixers) and community dynamic roles (e.g. early succession colonizers, disturbance-regime modifiers) (Chapin et al. 1995). Therefore, every species will probably occupy multiple roles within a community and ecosystem. Two examples of species with multiple roles are kelp, which are important primary producers who also create habitat for a large number of species, and oysters, which are important filter-feeders as well as space-occupiers and reef-formers.
In the following summary, we highlight a few examples of the consequences of reductions in diversity at higher trophic levels. This top-down perspective (e.g. Menge 1992; Power 1992; Strong 1992) does not preclude the possibly great importance of "bottom-up" effects (e.g. Carpenter 1988; McQueen et al. 1989; Hunter and Price 1992; Tiiman and Downing 1994; Menge et al. 1995), but instead reflects the limitations of our knowledge. Although consumers sometimes have unique influences in communities, certain implications can be generalized to other trophic levels or functional groups (see Section 14.4)
14.3.3 Consequences of consumer loss: selected examples
We separate our examples by two important distinctions: the strength of the overall effect of the functional group on the system and, where that overall effect is strong, the relative influence of each specics performing the role. Within strong functional groups, we use the term "diffuse" to characterize the condition of many influential consumer species because the major effects are spread among a number of species, in contrast to functional groups in which a single influential, or "keystone", species determines the major patterns (Menge et al. 1994; Power et at. 1996).
Strong predation: Keystone effects When a species has a strong influence on community structure, one which is out of proportion to its relative abundance, it is said to be a "keystone species" (Power et al. 1996). In a classic study (Paine 1966, 1974, 1994) in the rocky intertidal landscape of the northeast Pacific, the removal of the predatory seastar Pisaster ochraceus had dramatic effects on community structure. Under "normal" (control) conditions, approximately 16 benthic macroscopic species of seaweeds and invertebrates inhabited primary rock space, but in the prolonged absence of Pisaster, ultimately only one species, the mussel Mytilus californianus, occupied primary space. In other words, the loss of this important predator, later termed a "keystone" (Paine 1969), had dramatic consequences for the community. Although a diverse assemblage of species was associated with mussel dominance (Suchanek ¡992; Lohse 1993), the ¿ntertidaí landscape in the altered system was substantially different from that in the base-line state. Furthermore, the elimination of most algal species (and their biomass) by mussels probably reduced local rates of primary productivity and, because the food web was altered, patterns of energy flow and nutrient cycling were probably altered as well. Thus, in this case, the loss of a single species led to substantial changes in community structure and probably ecosystem-level processes.
Recent studies have greatly expanded our understanding of the dynamics of this system over a wider range of environmental conditions and somewhat larger spatial scales. Paine's (1966) study was conducted at two wave-exposed headlands on the outer coast of Washington State. Two subsequent studies (Dayton 1971; Menge et al. 1994, 1995) indicate that keystone predation is "context-dependent". That is, if a broader range of rocky shore habitats is considered, the community impact of this particular seastar is strong under certain conditions and weak or absent under others. In this system, the "context" or variation in habitat includes a range of wave exposures, types of environmental stress, and levels of nearshore phytoplankton productivity. Keystone predation was consistent in productive habitats of high wave turbulence and low sand movement, but weaker or absent in less productive habitats of low wave turbulence or high sand movement. Furthermore, while other predators (e.g. whelks, crabs, birds) had only minor effects when Pisaster dominated, at least one of these (whelks) increased their predatory impact when Pisaster was removed (Dayton 1971; Menge ef al. 1994; Navarrete and Menge, 1996). Thus, while Pisaster maintained the structure of major portions of rocky intertidal habitats, some community components (whelks) were capable of partially compensating for reductions of Pisaster under some conditions. Furthermore, the importance of Pisaster has been linked to variations in bottom-up processes such as primary and secondary production and prey recruitment (Menge 1992; Menge et al. 1994, 1995). Thus, community dynamics and ecosystem-level processes in this system depend on a set of ecological processes, with keystone predation being central.
Comparable results have been obtained in other rocky intertidal studies. For example, in the northwest Atlantic, the potential dominance of another mussel, Mytilus edulis, was constrained by a similar array of physical and biotie processes which varied along environmental gradients (aerial exposure at different tidal elevations on shore; wave impact along horizontal wave exposure gradients; Menge 1976, 1978a, 1983; Lubchenco 1978, 1980, 1983, 1986; Lubchenco and Menge 1978). Keystone predation, however, was more spatially restricted. The whelk Nucella lapillus maintained a fucoid-dominated community in the mid-intertidal zone of wave-sheltered shores by controlling the abundance of the competitively dominant mussels in these areas. At wave-exposed sites, other factors (disturbance and competition) were the dominant forces structuring communities. Whelks maintained a system of slightly higher diversity, but the difference was not as striking as in the P/mv/er-dominated system. However, as in the Washington and Oregon communities, ecosystem attributes such as local primary productivity, nutrient cycling and energy flow patterns probably differed in the presence and absence of whelks.
Other examples in different geographic locations have also suggested strong predation by single species of whelks, but with some variability in the impact of predation. The very strong effects of the keystone whelk Concholepas concholepas on Chilean rocky shores was mentioned above (Castilla and Duran 1985; Moreno et al. 1986; Duran and Castilla 1989). On some warm-temperate shores in eastern Australia, the whelk Morula marginalba has been shown to have major effects on prey (barnacles, limpets) abundance, thereby having important direct and indirect effects on community structure (Fairweather and Underwood 1991). These effects varied in both space and time, however, and the overall impact of whelks seemed weaker than in New England or Chile.
In a shallow subtidal example, the sea otter Enhydra iutris plays a strong community- and ecosystem-regulating role in the Western Aleutian Islands of Alaska (Estes and Palmisano 1974; Estes et al. 1978; Duggins et at. 1989; Estes and Duggins 1995). By the early 1900s, overexploitation of otters had eliminated them from all portions of their geographic range except a few remnant populations in Alaska and central California. Where present, otters controlled abundance of their preferred prey, herbivorous sea urchins (Strongyloceniroius spp.), and as a consequence dense, persistent kelp beds developed in the otter's foraging depth range (Estes et al. 1978). Where otters were absent, urchins were large and abundant, and they virtually eliminated kelps, producing "urchin barrens", or areas devoid of algae other than a pavement of encrusting coralline algae. The scarcity of algal food for the urchins caused them to forage actively over the barrens, preventing new kelp recruitment. In recent years in the Western Aleutians, as the otters have become re-established around islands that were previously dominated by urchin barrens, kelp forests have been developing (Estes and Duggins 1995). This keystone effect is also context-dependent: the strength of the otter effect varies spatially, partly as a consequence of variation in prey recruitment (Estes and Duggins 1995). Furthermore, dense kelp beds exist beyond the range of sea otters, so clearly other mechanisms can maintain kelp beds (Estes and Harrold 1988; Foster and Schiel 1988).
Community and ecosystem consequences of the loss of otters are dramatic. Kelp forests typically have a high diversity of organisms associated with them, including invertebrates and fishes, while urchin barrens have low diversity. The kelp itself provides a substantial fraction of the food that enters other, adjacent food webs (Duggins et al. 1989), perhaps even reaching some terrestrial communities (Estes et al. 1978). Removal of otters would significantly reduce kelp bed productivity and eliminate a host of other species, including many of commercial interest (e.g. fishes). The mechanisms of further spccies loss are likely to include loss of food, loss of habitat/shelter, and loss of nursery grounds.
The phenomenon of keystone predation is common (Menge et al. 1994; Power et al. 1996) and community dynamics and often ecosystem-level processes appear to differ greatly as a consequence of the loss of a single influential consumer. Other examples include the seastar Stichaster australis on the west coast of the North Island of New Zealand (Paine 1971), which performs a role similar to that of Pisaster. On Santa Catalina Island, off southern California, the lobster Panuliris interruptus maintains an algal turf community by eliminating mussels from wave-exposed shores (Robles 1987; Robles and Robb 1993). However, little is generally known about the magnitude and extent of the influence of subordinate consumers. The limited evidence available from systems in which studies have been done over a broad range of environmental contexts suggests that species that are subor dinate under some conditions can be important under others (e.g. Dayton 1971; Navarrete and Menge 1996).
Strong predation: Diffuse effects Where strong ctfccts are not dominated by a single influential species, but rather by a suite of strong interactors, the effects are "diffuse" (Menge et al. 1994). Such effects are apparent in a few studies. For example, experiments on the rocky shores of the Pacific coast of Panama indicated that the overall effects of consumers on benthic species were strong (Menge and Lubchenco 1981; Lubchenco et at. 1984; Menge el al. 1985, 1986a,b; Menge 1991; see Hairston 1989 for a review), and were even stronger than for temperate systems (see above). The assemblage of consumers was diverse, and included multispecific groups of limpets, chitons, predaceous gastropods, herbivorous gastropods, crabs, and, notably. 22 species of fishes, all of which fed on intertidal prey. Consumers kept all prey in check including the potentially dominant bivalves. Community structure in the presence and absence of al! consumers differed strikingly. However, a major deviation from keystone-type dynamics occurred tu the Panama system. The experiments indicated that no single consumer group served a keystone role. Instead, each group seemed to have roughly equivalent effects, and the strong predation that characterized this community only became apparent after most or all consumer groups were excluded (Menge et at. 1986a).
Thus, this system was characterized by "diffuse" but strong predation. Importantly, predator groups (and by implication, species) exhibited strong compensatory responses; in the event of species deletions, little community change occurred until a iarge fraction of al! species was eliminated. Species compensated for one another in their functions. The effects of consumer species loss on ecosystem-level processes was not studied, however. Primary producer biomass was always low, although primary productivity may have been high. Energy flow patterns would also be Hkely to change in the absence of consumers.
Diffuse-type community dynamics have been demonstrated in other systems. First, the low intertidal zone of the northwest Atlantic displayed mussel/barnaclc/algal interactions that were similar to those observed in the mid-zone, where Nuceila lapillus was a keystone predator (Menge 1976; Lubchenco and Menge 1978). In the low zone, however, predation appeared diffuse (Menge 1983; B. Menge, unpublished data, 1975), usually with 2 3 of a total of five predator species involved in controlling mussels and maintaining a zone of the red alga Chondrus crispus. The predators included the same whelk from the mid-zone, two crab species and two seastar species. In this example, community dynamics in the absence of predators would again be dramatically different from dynamics in their presence, and ecosystem-level processes would most likely vary in ways suggested earlier for the mid-zone. The effects of different predators were not separated experimentally, however, so the existence of compensatory predator responses could only be suggested (Menge 1983). Second, Robles and Robb (1993) suggested that lobsters and whelks jointly structured low intertidal rocky shore communities at wave-protected areas on Catalina Island in California. Finally, the experimental results of Kitching et al. (1959) strongly imply that diffuse-type dynamics occur on shores of intermediate wave exposure in Lough Ine. The system was very similar to that in New England, with many of the same species. Thus, while keystone effects were common in functional groups that had a strong influence on their community, they were not universal. Functional groups with diffuse effects strongly influenced community structure, but compensatory responses prevented change in the system even if some influential species were removed.
Weak predation In some communities, predation has only minor effects on community patterns such as distribution, abundance, size structure or diversity. Relatively few examples of these weak effects have been published (perhaps because the results are construed as "negative", Connell 1983; Sih et al. 1985). Three examples demonstrate weak predation in rocky intertidal communities. (1) On wave-exposed headlands in New England, predator effects were not detected in predator-exclusion experiments despite the presence of predator species and densities of consumers that were comparable to those in more sheltered habitats where predation was strong (Menge 1976). Predators, mostly whelks, rarely foraged actively and their foraging excursions were spatially restricted to the immediate vicinity (within centimetres) of habitat discontinuities offering shelter from severe wave turbulence (Menge 1978a). Biomechanical studies have subsequently confirmed the likelihood of the suggested mechanism: with increasing turbulence, whelks experience increasingly high risk of mortality due to dislodgement from the substratum, especially when active (Denny 1988). (2) In an example similar to (1), in Costa Rica, the whelk Acanthina brevidentata had little effect on prey over most of the shore (except within a few centimetres on some crevices), and thus had little or no impact on community structure (Sutherland 1990). (3) In Oregon, in sheltered rocky intertidal habitats that are frequently buried by sand, predation was a minor source of prey mortality (Menge et al. 1994). Instead, sand burial appeared to be the primary determinant of prey mortality in the low zone. (4) In wave-exposed areas of southwestern South Africa, predation effects were not detected in field experiments (G. Branch and R. Bustamante, unpublished data, 1994), despite the presence of predator species (crabs and whelks). Interactions between mussels, macrophytes and a guild of unusually large limpets determined patterns of space occupancy (Bustamante et al. 1995).
14.3.4 Conclusions from the survey
(1) Two distinct forms of strong predation, keystone and diffuse, are evident in the systems for which we have experimental information. Variability between keystone and diffuse predation is observed among communities with different assemblages of species as well as within communities along environmental gradients (e.g. wave action, productivity). The specific predation regime is thus context-dependent. Keystone-type dynamics are not universal in natural ecosystems.
(2) In a system dominated by keystone predation, some subordinate predators may play major roles following the loss of the keystone. Current information is inadequate to determine if this is a general characteristic of keystone-dominated functional groups. Although keystone predation varies along environmental gradients, where environmental factors are consistent (e.g. at wave-exposed headlands on northeast Pacific shores), the keystone role appears remarkably consistent over a large geographic range. Our ability to predict the community consequences following the loss of the keystone is high under these conditions.
(3) Compensation characterizes diffuse predation systems, whereby the loss of one predator is "compensated" for by others, with little if any community change. In these systems predation is still a major structuring process, but demonstration of its effects requires the removal of most predator specics.
(4) In some communities, predation is not a significant structuring process, despite the presence of predator species. In such cases, removal of predators has little effect.
As documented by Menge et ul. (1994), specics or habitat characteristics that indicate whether a system is dominated by keystone, diffuse, or weak predation are not yet obvious. Despite this lack of specific predictors, these studies of community dynamics in marine systems may provide valuable guidelines in identifying potential keystone systems. We suggest that in the context of this body of knowledge, and with a modest amount of natural history investigation, pattern quantification and short-term experimentation, approximate community dynamics can be predicted.
The overall influence of a functional group (for example, the strength of the predation effect) and the relative contribution of cach species within the group can be represented graphically (Figure 14.1). We highlight three major categories of systems and focus on the effects of consumers in determining high i
Was this article helpful?