Indirect effects of contaminants in aquatic ecosystems

Abstract

Contaminants such as petroleum hydrocarbons, heavy metals and pesticides can cause direct toxic effects when released into aquatic environments. Sensitive species may be impaired by sublethal effects or decimated by lethality, and this ecological alteration may initiate a trophic cascade or a release from competition that secondarily leads to responses in tolerant species. Contaminants may exert direct effects on keystone facilitator and foundation species, and contaminant-induced changes in nutrient and oxygen dynamics may alter ecosystem function. Thus, populations and communities in nature may be directly and/or indirectly affected by exposure to pollutants. While the direct effects of toxicants usually reduce organism abundance, indirect effects may lead to increased or decreased abundance. Here we review 150 papers that reference indirect toxicant effects in aquatic environments. Studies of accidental contaminant release, chronic contamination and experimental manipulations have identified indirect contaminant effects in pelagic and benthic communities caused by many types of pollutants. Contaminant-induced changes in behavior, competition and predation/grazing rate can alter species abundances or community composition, and enhance, mask or spuriously indicate direct contaminant effects. Trophic cascades were found in 60% of the manipulative studies and, most commonly, primary producers increased in abundance when grazers were selectively eliminated by contaminants. Competitive release may also be common, but is difficult to distinguish from trophic cascades because few experiments are designed to isolate the mechanism(s) causing indirect effects. Indirect contaminant effects may have profound implications in environments with strong trophic cascades such as the freshwater pelagic. In spite of their undesirable environmental influence, contaminants can be useful manipulative tools for the study of trophic and competitive interactions in natural communities.

Introduction

Indirect effects in ecological communities (complex relationships involving three or more species, Strauss, 1991, Wootton, 1994a) recently have been the subject of increased scrutiny. Trophic cascades (indirect effects mediated through consumer-resource interactions) are a well-studied type of indirect effect (Pace et al., 1999), and are generally considered in terms of ‘top–down’ (predator influence on lower trophic levels) and ‘bottom–up’ (nutrient/food/prey influence on higher trophic levels) causes. Numerous top–down effects have been detected in aquatic communities including planktonic (Brett and Goldman, 1996), nektonic (Estes et al., 1998), reef (Hay, 1997), intertidal and subtidal hard-bottom (Witman, 1987, Menge, 1995) and soft-sediment systems (Kneib, 1991, Posey and Ambrose, 1994). Some aquatic systems respond to both bottom–up and top–down factors (Posey et al., 1999).

When pollutants are released into aquatic habitats, direct (toxic) effects on aquatic biota are possible. Direct effects vary with the intensity and duration of exposure to a toxicant, and they are frequently studied, in part, because predictive criteria to estimate risk and establish permissible levels of contamination are based on species responses to contaminants (Long et al., 1995). These criteria are derived from laboratory toxicity tests that usually employ model species in single-toxicant exposures. The direct effects of toxicants typically reduce organism abundance (by increased mortality or reduced fecundity). Biota from a given habitat often exhibit a wide range of tolerance to specific toxicants (e.g. insecticides and herbicides target specific organisms in an interacting community) with the consequence that a toxicant may exert lethal effects on some species, but cause no observable effects on others. Direct sublethal effects, e.g. behavioral impairment or physiological stress, are also possible.

Pollutants may, however, exert effects on tolerant species by a number of ecological mechanisms. Such effects are called indirect (or secondary) contaminant effects. Single species, laboratory-based toxicity tests cannot detect indirect contaminant effects; studies at the population, community or ecosystem level (most commonly conducted in field or microcosm settings) are required (Cairns, 1983, Clements and Kiffney, 1994). The direct influences of contaminants on predators/grazers (e.g. through lethality or altered behavior) can lead to cascading indirect effects on resistant species in other trophic levels. The direct effects of contaminants on sensitive species may also alter competitive interactions within the resistant portions of producer and consumer communities. In addition, toxicants may directly influence ‘keystone facilitator’ or ‘foundation’ species (cf. Bruno and Bertness, 2001, species that positively affect the fitness of other species through their modification of the environment), and thereby lead to changes in the abundance of associated species. Similarly, disturbance rates or resource availability may be influenced by contaminants, which may in turn modify important ecosystem functions (e.g. decomposition rates, oxygen dynamics and nutrient cycling). Finally, Spromberg et al. (1998) used a theoretical analysis to suggest that localized toxicant-induced mortality may alter metapopulation dynamics, and that the contaminant-induced loss of subpopulations may have ecologically significant impacts on non-exposed groups. Thus, the mechanisms associated with population and community change following contaminant exposure in the field are potentially complex and quite varied. Indirect toxicant effects may lead to increased (e.g. via reduced competition) or decreased (e.g. via reduced availability of preferred food) abundance.

The purpose of this review is to evaluate the indirect effects of toxicant exposure in aquatic communities. We have found reference to indirect toxicant effects in 150 refereed papers published since 1970. Earlier references and reports in the gray literature are available in reviews of specific toxicants, including Hurlbert, 1975, Barron and Woodburn, 1995, Pratt et al., 1997, Graymore et al., 2001. Approximately 10% of these references examined accidental contaminant releases, chronic contamination or model development. Thirty papers examined the influence of contaminants on some aspect of behavior in aquatic biota that might contribute to an indirect effect. Nearly 100 experimental studies have considered indirect effects after contaminants were amended in field or microcosm experiments. The purpose of most of these studies was to quantify toxicant (i.e. direct) effects, although toxicants were applied in a few studies to manipulate the abundance of target taxa to test hypotheses regarding species interactions. We focus on papers that tested for contaminant effects on food webs (by surveying organism abundance at more than one trophic level, see Table 1), and thus yield information on trophic cascades. Only 10% of the experimental studies considered indirect effects and reported that none occurred, or that observed effects may not be related to trophic cascades (e.g. Gruessner and Watzin, 1996 reported that an herbicide caused early emergence of aquatic insects but could not be sure the effect was due to a reduction in algal food supply). Eighty-three papers identified possible indirect effects of contaminants. Of these, 60% examined freshwater pelagic and 20% freshwater benthic responses to contaminants. Reports of indirect effects from marine/estuarine habitats are rare; the majority have been conducted in marine benthic systems.

More than half of all studies involved the use of insecticides (and most of these were conducted in freshwater habitats). Indirect effects of herbicides were noted in less than 20 investigations, and metals and hydrocarbons in less than 10 each. The indirect effects of many contaminants (e.g. fungicides, surfactants) have only rarely been examined.

There are a number of ways this collection of references may be biased. Direct effects of contaminants are reported in 100's of papers in which possible indirect effects may not have been considered or noted. Furthermore, studies are subject to inherent biases. Microcosm experiments facilitate the examination of short-term responses to contaminants while field studies after an accidental release of a contaminant may miss short-term indirect effects, but reveal longer-term or chronic effects. Although microcosm studies are often criticized for not adequately representing natural communities to the point that they lose ecological relevance (Gray and Pearson, 1982, Carpenter, 1996), they are uniquely capable of generating and testing hypotheses regarding the mechanisms by which indirect effects may occur.