Research Themes: (B) Improving theory and models for ecological risk assessment

ERA models for exotic species, GEOs and biological control agents are sparingly quantitative and extremely diverse in degree of sophistication. In the USA, exotic species ERA generally relies on qualitative expert opinion, while for GEO ERA, models are more developed but vary from somewhat qualitative food web models to more quantitative migration-selection-population dynamics models (Table 1). Strategies for model construction range from induction from empirical results (e.g., colonization and establishment models) to derivation from well-established population genetic theory. This diversity of modeling schemes poses challenges that we will address during our IGERT (Table 1).

We illustrate some of these challenges below. To prepare to strengthen and unify theory, students will gain broad exposure to the diversity of existing models in the modeling workshop, the Risk Analysis Survey course and the modeling course. Several problems that have not yet proven amenable to modeling useful for ERA (e.g., delayed impacts, indirect effects) will be described in the Survey Course and will serve as a focus for discussions during the roundtables and symposia.

Colonization.

Existing models have considered the arrival process empirically, relying on known transport pathways and assuming the probability of arrival is proportional to propagule pressure. In fact, arrival may be a non-linear function of the number of propagules and depend on propagule aggregation. Several more rigorously quantitative models of exotic species colonization assess marine species invasions via ballast water and shipping patterns and may guide improvement of the mostly qualitative models.

Economists have begun to investigate how best to intervene when international trade increases the risk of invasions. For example, McAusland and Costello (2004) found that the threat of new invasions depends on the past trade level with a region and the past exposure to exotic species. Identifying the relative risk of trade partners based on these aspects and then targeting specific regions can reduce inefficiencies resulting from certain market-based mechanisms, such as non-specific tariffs. Knowler and Barbier (2005) have demonstrated that taxes can produce a socially optimal level of exotic plant imports. Costello and McAusland (2003) have shown that protectionism may not mitigate invasion risks, and failure to account for agricultural damages skews the interpretation of the efficacy of these mechanisms. Students will have the opportunity to explore how these and related models may link to biological colonization models so that management of ISGs can be integrated economically into broader discussions about trade policy.

Establishment.

The establishment process is modeled as a function of the intrinsic growth rate, r. If r >0, establishment occurs, otherwise it does not. Climate matching is currently one of the main considerations in predicting r. Students will participate in discussions of additional ecological aspects to consider improving predictions of r in the new environment. An extensive literature notes characteristics associated with invasiveness, but most of these have little predictive value. We will hold taxon-specific discussions of organismal characteristics that may help predict r in new environments, and these characteristics will be evaluated systematically through literature reviews and experiments. The taxa we will examine includes soil microbes, plants, insects and fish. Characteristics that prove useful may be incorporated into existing climate-matching models. A novel element of this work will be the cross-taxon comparisons that will become possible as students progress in their research.

Spread of introduced organisms and gene flow risks.

odels of both spread and gene flow are based on reaction-diffusion and migration-selection models. Spread models have been improved via more realistic population growth components, such as Allee effects, and gene flow models via inclusion of spatially restricted dispersal. In addition, Sharov and Liebhold (1998) have emphasized how to use spread models to “slow the spread” of an invading species in an economically optimal way. As with establishment, however, ecological factors that affect the key parameters have not been incorporated into the models. For example, the shape of the dispersal kernel, the rate of population growth at low density, and the selective advantage of a rare trait are all affected by ecological factors, but these have not been incorporated into models, thereby limiting accuracy of prediction. One specific area for student research will be gene spread models based on sexual selection. These models have been used to model gene spread in fish and may be more widely applicable.

Direct ecological impacts.

One of the most challenging aspects of invasion biology is to predict the ecological effects of a new species or genotype. Following taxon-specific models, Tilman (2004) developed a model for plant invasions based on the ‘R*’ rule - a species (or genotype) that can persist at the lowest level of a limiting resource will displace other species or genotypes. Consistent with this model, Fargione et al. (2003) showed that plant species most strongly inhibited the establishment and growth of invading species with similar resource requirements. This experimentally observable R* can be used to predict the effects of introduced genotypes or the efficacy of biological control agents in suppressing pests, as in the case of the California red scale. We will encourage students to test and extend this theory in their own research.

Minimizing direct ecological impacts via design.

Our students will investigate strategies to breed non-invasive horticultural crops and farmed fish to minimize potential ecological impacts. In breeding programs, selected traits that confer market value generally constitute the basis for domestication. It may be possible to establish a ‘non-invasive crop ideotype’ and breed against invasiveness. Invasion models that associate species traits to ecological impacts in heterogeneous environments could inform breeding objectives. The net fitness-Trojan gene model offers another perspective on breeding objectives. Since breeding programs are long-term, IGERT students would conduct research on the design of non-invasive horticultural crops and fish with known invasive types, using field trials to evaluate invasion risk in multiple environments.

Indirect non-target impacts and food webs.

Assessment of harm to biological diversity is typically indirect, relying on indicators of potential harm. The use of indicators has a long history and has proven valuable in several cases (e.g., mayflies as indicators of acidification of streams), but it has little scientific support as it is applied to invasive species, biological control agents and GEOs, for which ERA should be case-specific. Alternatives, however, have not been fully developed and validated. This gap offers rich research opportunities for IGERT students. The Andow & Hilbeck (2004) model classifies biological diversity according to ecological function (e.g., herbivory). For each function, worst-case risks in the local environment are identified (e.g., increased crop losses from enhanced herbivory), and species that are most likely causes are identified and used to assess the risk. This model allocates effort to the most serious concerns, uses financial resources efficiently and allows flexibility in developing a strategy for assessing risk. Another kind of model quantifies the probability of harm to a particular non-target species that is of special concern. For example, IGERT students could further develop the quantitative monarch butterfly model.

The integration of quantification into decision-making.

Increased quantification may not improve social deliberation and decision-making unless it is done in an iterative, deliberative process that involves diverse stakeholders. Quantification could enrich the decision-making process by informing deliberation on comparative futures, whereas quantification might be ignored if it fails to clarify cause-effect linkages. A methodology to improve social deliberation is Problem Formulation and Options Assessment (PFOA), which establishes context for societal dialogue concerning the use of a proposed GEO technology, such as Bt maize in East Africa. This multi-stakeholder approach to deliberation offers a rational, science-driven planning process by which stakeholders can assess their needs, evaluate the risks related to various options, and recommend to decision-makers policies to reduce societal risks and to enhance the benefits of various options. Improved quantification, in conjunction with stakeholder conceptual models, offers the greatest potential for strengthening ERA and decision-making for biological safety.

In this Section

A) Evaluating effectiveness of regulation using retrospective analysis

B) Improving theory and models for ecological risk assessment

C) Addressing uncertainty in risk assessments

D) Managing introduced species and genotypes and post-removal strategies