Causal Analysis/Diagnosis Decision Information System (CADDIS)

• Data Analysis Methods
• Fundamentals of Data Analysis
• Get Data Analysis Tools

DA.7. Extrapolation

DA.7.1. Extrapolation Issues

Relating stressor-response data from the laboratory or from other field studies to the case at hand requires extrapolation. Common extrapolation issues include:

Extrapolating from Analogous Stressors

• For biological stressors such as invasive species or pathogens, consider whether the species have the same mode of action. For example, data concerning effects of introduced rainbow trout on invertebrates might be extrapolated to introduced brown trout but not to introduced carp.

• For physical stressors such as increased temperature, low DO, or altered flow, data concerning the effects of natural fluctuations or disturbances may have different patterns of variance than anthropogenic sources.

Extrapolating Laboratory Results to Field Effects

• Choose test data with similar water chemistry to the site with respect to parameters that influence speciation, bioavailability, etc.

• Choose test data with similar exposures to the site including routes and duration.

• Choose test data with similar organisms including taxonomy and life stage.

• Consider whether the impairment involves types of effects that do not occur in laboratory tests including predation, competition, avoidance, and use of refugia.

Extrapolating between Geographical Areas

• Choose data from sites with similar environmental conditions (climate, disturbance, geomorphology, hydrology, etc.).

• Consider which differences in environmental conditions are likely to significantly influence the impairment of the biotic community.

• Use categorization or normalization to reduce uncertainty due to extrapolation among sites.

DA.7.2. Extrapolation Approaches

Judgment Approaches

• Judgment approaches rely on professional expertise to relate the information that is available to observed biological impairments.

• Judgment is essential when databases are inadequate to support empirical models, and process models are either unavailable or inappropriate.

• Judgments are typically performed ad hoc, but formal expert elicitation methods may be used.

Empirical Approaches

• Empirical extrapolation models have been developed to relate laboratory test data to organisms in the field.

• They include factors, regression models, allometric scaling, and species sensitivity distributions (see Ch. 26 in Suter, 2007).

• Empirical models are used to extrapolate between species, life stages, exposure durations, body sizes, and types of effects.

• EPA's Acute-to-Chronic Estimation (ACE) and Interspecies Correlation Estimation (ICE) applications may be helpful tools if extrapolations from laboratory data are necessary.

• These empirical extrapolation models were developed for toxic chemicals.

Process-Based Approaches

• If impairments are defined by properties of populations, a Leslie matrix or other population process model may be used to determine whether those effects can be explained by the effects of a candidate cause on individual organisms (see Ch. 27 in Suter, 2007).

• If impairments are defined by ecosystem properties or if effects on populations are possibly mediated by competition, predation, or other ecosystem processes (i.e., the effects are indirect), ecosystem process models such as AQUATOX can be used to determine whether the effects can be explained by the effects of a candidate cause on individual organisms (see Ch. 28 in Suter, 2007).

• These models allow the quantification of causal relationships that have not been measured and therefore cannot be empirically modeled. However, they require knowledge of biological processes and their functional relationships to environmental entities and processes.

Suter, GW II. (2007) Ecological risk assessment, 2nd ed. Boca Raton, FL: CRC Press/Taylor & Francis.