After biocontrol: Assessing indirect effects of insect releases
Introduction
Development of biological control agents for weeds has been motivated by the need to reduce the abundance and distribution of a pest plant where chemical and mechanical controls were not cost-effective (Harris, 1993, Pemberton and Turner, 1990). Widespread weeds of uncultivated lands, including rangeland weeds, such as Hypericum perforatum L. (Harris et al., 1969), Euphorbia esula L. (Harris, 1993), and Opuntia spp. (Moran and Zimmermann, 1984) are good examples. Historically, the primary objective of weed biological control has been the direct reduction in density, cover, and range of a target weed that has suppressed desirable forage species or was toxic or unpalatable to livestock. A secondary objective, therefore, has been to improve pasture or rangeland productivity or to promote the replacement of the target with something more desirable. Target selection has been influenced strongly by client groups, such as cattlemen’s associations, whose sustained interest has been critical to the continued development of such publicly funded projects. Yet increasingly, public, government, and scientific communities in the USA have become skeptical of biocontrol programs because of their potential to cause both direct and indirect nontarget impacts (e.g., Beardsley, 1997, Louda et al., 1997, Louda et al., 2003a, Louda et al., 2003b, Louda et al., 2005, McEvoy and Coombs, 2000, Simberloff and Stiling, 1996). Such skepticism raises the question of how clearly the goals and benefits of biocontrol programs are articulated in advance and the degree to which direct and indirect impacts can be predicted and measured.
With increasing appreciation of the impact of invasive plants on native ecosystems, and heightened value placed on wildlands, weeds of natural areas recently have become an important focus of biological control projects. As is the case for rangeland weeds, the potential effectiveness of biological control over large areas of poorly accessible terrain is attractive economically (Malecki et al., 1993). Moreover, biocontrol agents are usually highly host-specific and thus produce fewer nontarget impacts than does widespread use of chemical and mechanical control methods. Recent projects for control of Melaleuca quinquenervia (Cav.) Blake in the Florida Everglades, Lythrum salicaria L. in the northern wetlands, and Tamarix spp. along western stream courses are driven by perceived threats these species pose to ecosystem services and native species. Desired outcomes of the biological control of such wildland weeds, therefore, include both direct reduction of the target as before as well as indirect effects such as increased abundance of native species, increased species diversity, restoration of vegetation structure, and restoration of ecosystem processes and services, such as water yield, that presumably had been provided by the pre-invasion community (Harris, 1993, Lesica and Hanna, 2004). Expansion of use of biological control to target wildland weeds, therefore, has increased the potential list of benefits of biocontrol while simultaneously increasing the complexity of measuring, monitoring, and modeling responses and potential benefits.
In the current climate of skepticism, it is increasingly important for scientists to quantify ecological responses to control activities and for biocontrol projects to be developed as part of integrated weed management and restoration programs, rather than as stand-alone projects. For meaningful risk/benefit analyses to be conducted, ecologists must be able to assign probabilities to outcomes of insect releases. We should be able to estimate not only the probabilities of the successful establishment of the released agent and subsequent depression of the target, but also the probabilities associated with desired community and ecosystem responses, particularly as they relate to the ultimate management goal that motivated control in the first place (Fig. 1).
Our goal in this paper is to evaluate the extent to which several seemingly successful biological control projects measure the full range of impacts of the biocontrol agents on the invaded communities and ecosystems. We are interested specifically in what was monitored after release of the control agent(s) and the extent to which indirect impacts (e.g., Fig. 1) were measured and ultimate goals achieved. We do not focus on feeding of the agent on nontarget plant species since that has been covered well elsewhere (e.g., Louda et al., 2003a, Louda et al., 2003b, McEvoy and Coombs, 2000, Pemberton, 2000, Simberloff and Stiling, 1996; and others). Rather, we evaluate how fully control projects measured desired or even undesirable outcomes and the extent to which true benefits were realized even with only partial control of the pest. Fig. 1 provides a conceptual presentation of potentially desirable and undesirable outcomes. Such information would be useful to determine whether a release is justified when full control is not expected. We have reviewed only successful projects, defined as those for which authors report noticeable reductions in the cover or distribution of the pest plant. We sought insights into ways in which to better measure community and ecosystem responses to biocontrol releases. Ultimately, such information should improve the utility of risk/benefit analyses as well as the outcomes.
Section snippets
Materials and methods
We consulted the published literature on nine biocontrol projects for which some measure of success in reducing the density, range, or abundance of the target pest has been claimed. Projects were selected from those listed in Julien and Griffiths (1998) for which one or more biological control agents released at least 10 years ago were described as substantially reducing the cover of the target weed. The project list includes target weeds of rangeland as well as wildlands and herbaceous, woody,
Results
Table 1 lists a range of impacts described in the literature with associated codes, which we then use to summarize published findings in Table 2. Where a citation claimed simply “substantial control” or similar phraseology, we assumed this meant reduction of cover or biomass (COV) unless reduction of density was specified. The history of most planned introductions of biological control agents describes highly variable results, including failure to establish; establishment but no effective
Conclusions
The above snapshots of large biocontrol projects reveal several common themes.
(1) Despite the varying quality of data, all of these examples document variation in the impacts of agent(s) across the invasive range of the target plant. Compensation for this variation often is made in the form of multiple agents targeted at specific climatic or habitat conditions. With better tools for predicting potential environmental ranges of agents (e.g., with the climate matching model, CLIMEX, e.g.,
Acknowledgments
We are grateful to the US Department of Agriculture CSREES program for support of the conference “Science and Decision Making in Weed Biological Control” held in Denver, Colorado January 27 to 29, 2004. This paper is the product of presentations and discussions at that conference. T. Dudley provided helpful discussion in the development of ideas presented here, P. McEvoy steered us to critical publications, and S. Louda and two external reviewers provided helpful reviews of the manuscript.
References (81)
Melding ecology, classical weed biocontrol, and plant microbial ecology can improve practices for controlling invasive plant species
Biol. Control
(2005)- et al.
Host specificity of the leaf beetle, Diorhabda elongata deserticola (Coleoptera: Chrysomelidae) from Asia, a biological control agent for saltcedars (Tamarix: Tamaricaceae) in the Western United States
Biol. Control
(2003) Effects, constraints and the future of weed biocontrol
Agric. Ecosyst. Environ.
(1993)- et al.
An ecological economic simulation model of mountain fynbos ecosystems. Dynamics valuation and management
Ecol. Econ.
(1997) - et al.
Biology of Diorhabda elongata deserticola (Coleoptera: Chrysomelidae), an Asian leaf beetle for biological control of saltcedars (Tamarix spp.) in the western United States
Biol. Control
(2003) - et al.
Biological control of jointed cactus, Opuntia aurantiaca (Cactaceae), in South Africa
Agric. Ecosyst. Environ.
(1991) History and success of plant pathogens for biological control of introduced weeds in Hawaii
Biol. Control
(2005)- et al.
The potential for indirect effects between a weed, one of its biocontrol agents, Mesoclanis polana and native herbivores: a food-web approach
Biol. Control
(2005) - et al.
Biological control of prickly pear, Opuntia ficus-indica (Cactaceae), in South Africa
Agric. Ecosyst. Environ.
(1991) - Bangsund, D.A., Leisritz, F.S., 1991. Economic impact of leafy spurge in Montana, South Dakota, and Wyoming....