Extinction and Ecosystem Function in the Marine Benthos

See allHide authors and affiliations

Science  12 Nov 2004:
Vol. 306, Issue 5699, pp. 1177-1180
DOI: 10.1126/science.1103960


Rapid changes in biodiversity are occurring globally, yet the ecological impacts of diversity loss are poorly understood. Here we use data from marine invertebrate communities to parameterize models that predict how extinctions will affect sediment bioturbation, a process vital to the persistence of aquatic communities. We show that species extinction is generally expected to reduce bioturbation, but the magnitude of reduction depends on how the functional traits of individual species covary with their risk of extinction. As a result, the particular cause of extinction and the order in which species are lost ultimately govern the ecosystem-level consequences of biodiversity loss.

Marine coastal ecosystems are among the most productive and diverse communities on Earth (1) and are of global importance to climate, nutrient budgets, and primary productivity (2). Yet, the contributions that coastal ecosystems make to these ecological processes are compromised by human-induced stresses, including overfishing, habitat destruction, and pollution (35). These stressors particularly impact benthic (bottom-living) invertebrate communities because many species are sedentary and cannot avoid disturbance. Thus, marine coastal ecosystems are likely to experience a large proportional change in biodiversity should present trends in human activity continue (68).

Given these prospects, researchers have recently asked how the loss of biodiversity might alter the functioning of marine coastal ecosystems. Like most studies to date, these experiments have manipulated diversity by assembling random subsets of species drawn from a common pool of taxa (911). This approach (12, 13) may be useful for understanding the theoretical consequences of diversity loss but is unrealistic in the sense that it assumes species can go extinct in any order. Extinction, however, is generally a nonrandom process (14) with risk determined by life-history traits such as rarity, body size, and sensitivity to environmental stressors like pollution (1518). Interspecific differences in extinction risk have implications for the ensuing changes in trophic interactions and community structure (18, 19), such that the ecosystem-level consequences of random versus ordered extinctions are likely to be fundamentally different (14, 2022).

Here we explore how various scenarios of extinction for marine benthic invertebrates are likely to influence bioturbation (the biogenic mixing of sediment)—a primary determinant of sediment oxygen concentrations which, in turn, influences the biomass of organisms, the rate of organic matter decomposition, and the regeneration of nutrients vital for primary productivity (23, 24).

Using a comprehensive study of 139 benthic invertebrate species that inhabit Inner Galway Bay, Ireland (25), we parameterized models that predict how species extinction is likely to affect the biogenic mixing depth (BMD), an indicator of bioturbation that can be measured from sediment profile images (Fig. 1). To estimate species contributions to the BMD, we used an index of bioturbation potential (BPi, Equation S1) that accounts for each species' body size, abundance, mobility, and mode of sediment mixing. We used data from monthly samples (over 1 year) of the benthic community to empirically derive a relation (Equation S2) between the BMD and the bioturbation potential of the community (BPc). Using this relation, we performed numerical simulations to explore how the BMD is expected to change when species go extinct at random versus ordered by their sensitivities to environmental stress, body size, or population size (25). As the functional consequences of extinction are known to depend on the response of surviving species (19, 20, 26), we simulated two different types of community interactions (8). First, we used a model in which species do not interact with one another; thus, surviving species do not exhibit compensatory responses (changes in population size) after extinction. This scenario leads to complete loss of bioturbation performed by an extinct species and represents a “worst-case” scenario. Second, we used an interactive model of community assembly in which species' abundances are limited by competition with other members of their functional guild (i.e., species with similar bioturbation modes but not necessarily similar extinction risks). This represents a “best-case” scenario that assumes compensation is additive and substitutions of abundance maintain total community density [i.e., full numerical compensation (25)].

Fig. 1.

The biogenic mixing depth (BMD, white arrows) of sediments [(A), site 1; (B), site 2] in Inner Galway Bay, Ireland. BMD was related to the bioturbation potential of a community (BPc), an index that accounts for each species' population size and life-history traits (body size, mobility, mode of bioturbation) to estimate the capacity of a community to mix sediments (25).

Our models predict that loss of species diversity leads to a decline in mean BMD, regardless of extinction scenario (Fig. 2). Note, however, that Fig. 2, A to H, depict strikingly different patterns, suggesting that changes in the BMD depend on extinction scenario. Indeed, the rate of change, the species richness at which the BMD first declines, the variance surrounding the relation (i.e., predictability of change), and the range of potential values all depend on how species go extinct (Table 1). These divergent patterns are best explained by examining the covariance between each species extinction risk and the biological traits that influence bioturbation (Fig. 3). To illustrate these patterns, we first focus on scenarios of extinction that involve no compensatory responses (i.e., the noninteractive model; Fig. 2, A, B, C, and D). Random extinction (Fig. 2A) produces a clear bifurcation of the BMD, with values determined by the presence (>4.0 cm) versus absence (<4.0 cm) of a single species—the burrowing brittlestar, Amphiura filiformis. The strong impact of A. filiformis on bioturbation is well documented (27). In this study, A. filiformis has a disproportionate impact (Fig. 3A) on bioturbation because it is consistently one of the most abundant species in Galway Bay (Fig. 3B) and has a high per capita effect that results from it being a large (Fig. 3C), highly mobile species. Consequently, changes in the BMD following extinction largely depend on whether A. filiformis is among the survivors.

Fig. 2.

Predicted changes in the BMD following benthic invertebrate extinctions. Each panel shows the results of 20 simulations per level of species richness, constrained by a probabilistic order of species extinction (indicated on the right). Simulations (A), (B), (C), and (D) are for a noninteractive model of community assembly assuming no numerical compensation by surviving species. Simulations (E), (F), (G), and (H) are for an interactive model that assumes full numerical compensation following extinction of competitors.

Fig. 3.

The relation between per capita bioturbation, BPi, and mean species abundance (A) reveals that at the population level (diagonal dashed lines, each an order of magnitude difference in bioturbation), most species contribute little to bioturbation (left of short-dashed line). Bioturbation is disproportionately affected by one large and highly active species, Amphiura filiformis (brittlestar, open circle). Population level bioturbation, BPp, is proportional to species abundance (B) (r = 0.83, P < 0.001), body size (C) (r = 0.39, P < 0.001), and sensitivity to stress (D) (r = –0.2, P < 0.05). Arrows indicate order of extinctions.

Table 1.

Comparisons of how bioturbation changes with species loss for each extinction scenario (stress, size, rarity) relative to a random model of extinction, and between the interactive and noninteractive models of community assembly. The asterisk (*) denotes significant differences, P < 0.0125 [set conservatively to correct for the number of comparisons (25)]. CV, coefficient of variation.

Mean CV Minimum Maximum
Comparison of random extinction to extinctions ordered by...
Sensitivity to stress χ24 = 0.73 F4, 1094 = 3.38* χ24 = 1.63 χ24 = 0.23
Body size χ24 = 53.8* F4, 1094 = 42.8* χ24 = 15.1* χ24 = 15.1*
Rarity χ24 = 28.2* F4, 1094 = 250* χ24 = 97.6* χ24 = 3.8
Comparison of interactive to noninteractive model for extinctions that are...
Random χ22 = 35.07* F2, 274 = 629* χ22 = 30.94* χ22 = 10.37*
Ordered by sensitivity to stress χ22 = 25.76* F2, 274 = 307* χ22 = 20.94* χ22 = 10.19*
Ordered by body size χ22 = 7.42 F2, 274 = 166* χ22 = 10.71* χ22 = 5.56
Ordered by rarity χ22 = 1.38 F2, 274 = 13.9* χ22 = 0.69 χ22 = 0.50

When extinctions are ordered by species sensitivity to stress (Fig. 2B), estimated as the relative change in the abundance of species along a gradient of disturbance (25), the risk of extinction among species varies by a factor of 215; yet, stress sensitivity for A. filiformis (–0.99, Fig. 3D) is near the median value for the community as a whole (–0.98), which explains why changes in the BMD are comparable to the scenario of random extinction (compare Fig. 2, A and B). This conclusion is confirmed by statistical comparisons of the mean and range of values (minimum and maximum) of the BMD, which show an identical change with species loss for both scenarios; and a comparison of the variability in BMD, which reveals only a marginal difference between scenarios (α = 0.0125; P = 0.01, Table 1).

For extinctions ordered by body size (Fig. 2C), probabilities of extinction were assumed to be proportional to mean species biomass to mimic the higher extinction risk generally faced by large-bodied organisms that often have small population sizes, have longer generation times, or are found at higher trophic levels (17, 28). Body size varied by a factor of >500,000 among species and was positively correlated with per capita effects on bioturbation (r = 0.98, P < 0.01) but not abundance (r = –0.05, P = 0.56, even excluding A. filiformis, r = –0.08, P = 0.33). In this scenario, larger species (high per capita effects) tended to be lost before smaller species (low per capita effects), leading to a faster decline in the mean BMD compared with random extinction (compare Fig. 2, A and C; Table 1). The range of values of the BMD (minimum and maximum) and total variation (CV) also changed with species richness more quickly than for random extinctions (Table 1). This was not due to the loss of entire functional guilds composed of large species because there was considerable overlap in species body size, and thus extinction risk, among functional guilds (25). Rather, patterns were generally a consequence of the early extinction of A. filiformis, the 19th largest species, which produced a step change in the BMD at a species richness of ≈100.

Extinction risk is typically high for rare species, defined here as those with low local abundances, because small populations are more vulnerable to environmental and demographic stochasticity (17, 28). They also often have narrow geographic ranges and/or high specialization, further compounding extinction risk (2830). When we assumed extinction probability was inversely proportional to species density, rare species were >6000 times more likely to be lost than the most common species. Yet, because small populations typically contribute little to bioturbation (Fig. 3B), extinctions of rare species had little impact on the BMD, and ecosystem functioning was maintained until the loss of more abundant species, such as A. filiformis (lower bifurcation, Fig. 2D). Hence, some scenarios of extinction do not lead to appreciable loss of ecological function until a large proportion of species are lost.

Many studies suggest that when species go extinct from communities characterized by strong interactions, increases in the population size of species released from competition can compensate for loss of ecological function (20, 31, 32). Our models suggest that this is only true when the risk of extinction is not correlated with species functional traits. This is evident because compensatory responses only changed the probabilistic distribution of the BMD when species were lost at random (Fig. 2E) or in order of their sensitivity to stress (compare Fig. 2, A and E, and Fig. 2, B and F) (Table 1). However, when a species' risk of extinction covaried with its body size or abundance, compensatory responses did not alter the consequences of diversity loss (compare Fig. 2, C and G and Fig. 2, D and H) (Table 1). This is because when loss is ordered by body size, small species have little impact on bioturbation and cannot offset functions performed by larger species. When species are lost in order of rarity, even full compensation has no notable effect on the BMD because the proportional change in bioturbation is small. Thus, compensatory responses of surviving species do not necessarily stabilize ecological processes when the traits required for maintaining function simultaneously increase extinction risk.

We have used numerical models parameterized by data from a marine benthic community to show that species extinction is generally expected to reduce the depth of bioturbated sediments. Such changes might be expected to alter the fluxes of energy and matter that are vital to the global persistence of marine communities (23), a conclusion that corresponds to evolutionary patterns in the fossil record showing a close association between the frequency of anoxia and the diversification of marine soft-bottom communities (33). We have also shown that crucial details (mean, range, and predictability of change) of how bioturbation changes following extinction depend on the order in which species are lost, because extinction risk is frequently correlated with life-history traits that determine the intensity of bioturbation. This finding is important because it argues that the particular cause of extinction ultimately governs the ecosystem-level consequences of biodiversity loss. Therefore, if we are to predict the ecological impacts of extinction and if we hope to protect coastal environments from human activities that disrupt the ecological functions species perform, we will need to better understand why species are at risk and how this risk covaries with their functional traits.

Supporting Online Material

Materials and Methods

Equations S1 and S2

References and Notes

View Abstract

Navigate This Article