Selection bias correction for species sensitivity distribution modeling and hazardous concentration estimation

Area of Science:

• Ecotoxicology
• Environmental Chemistry
• Environmental Risk Assessment

Background:

• Species sensitivity distribution (SSD) is a standard method for determining safe contaminant concentrations in water bodies.
• The statistical validity of SSD relies on the assumption of random species selection, which is frequently not met.
• Existing concerns regarding SSD methodology have persisted for years.

Purpose of the Study:

• To investigate the impact of nonrandom species selection on the accuracy of hazardous concentration estimations derived from SSD.
• To evaluate potential biases introduced by nonrandom sampling in ecotoxicological assessments.

Main Methods:

• Statistical analysis of species sensitivity distribution models.
• Simulation of nonrandom species selection scenarios.
• Evaluation of error propagation under biased sampling conditions.

Main Results:

• Nonrandom species selection can introduce hazardous concentration estimation errors of 20-fold or greater.
• Bias towards more sensitive species in toxicity data can lead to incorrect compensation when using lower confidence interval limits.
• The assumption of random species selection is critical for reliable SSD outcomes.

Conclusions:

• The widespread use of SSD requires addressing the violation of random selection assumption.
• Current practices may underestimate or overestimate safe contaminant levels due to nonrandom species selection.
• Further research is needed to develop robust SSD methods that account for sampling biases.