Related Experiment Videos
Better bootstrap estimation of hazardous concentration thresholds for aquatic assemblages.
Eric P M Grist1, Kenneth M Y Leung, James R Wheeler
1School of Biological Sciences, Royal Holloway, University of London, Egham, Surrey, United Kingdom. e.grist@rhul.ac.uk
Environmental Toxicology and Chemistry
|July 12, 2002
Summary
Species sensitivity distribution (SSD) methods improve ecological risk assessment. Advanced bootstrap techniques offer more reliable estimates for safe environmental concentrations, moving beyond traditional, less robust statistical models.
Area of Science:
- Environmental toxicology
- Ecological risk assessment
- Statistical modeling
Background:
- Species sensitivity distributions (SSD) are used in ecological risk assessment for deriving predicted no-effect concentrations.
- Conventional SSD methods often rely on standard distributions (e.g., log-logistic, log-normal) lacking robust ecological or statistical justification.
- Bootstrap resampling offers an alternative to distributional assumptions in SSD analysis.
Purpose of the Study:
- To describe an advanced bootstrap methodology for deriving improved point estimates and confidence intervals for SSDs.
- To explore a hybrid bootstrap regression approach for SSD estimation when standard models are inadequate.
- To highlight the need for consensus on appropriate SSD derivation methods in ecological risk assessment.
Main Methods:
- Application of advanced bootstrap resampling techniques for SSD analysis.
- Development and application of a hybrid bootstrap regression approach.
- Comparison of bootstrap methods with conventional parametric curve approaches.
Main Results:
- Advanced bootstrap methods provide better point estimates and confidence intervals for SSDs.
- The hybrid bootstrap regression approach yields substantially different SSD estimates compared to basic bootstrap and parametric methods.
- The true SSD may not conform to standard statistical distribution categories.
Conclusions:
- Advanced bootstrap methodologies offer a more scientifically sound basis for SSD derivation in ecological risk assessment.
- Hybrid bootstrap regression is a valuable tool when empirical data do not fit standard SSD models.
- Establishing consensus on robust SSD derivation methods is crucial for the advancement of ecological risk assessment practices.