Inferences about the variability of means from censored data

Area of Science:

• Environmental Science
• Analytical Chemistry
• Statistics

Background:

• Intensive biological monitoring of the Niagara River by the Ontario Ministry of the Environment.
• Focus on determining relative bioavailability and sources of trace contaminants.

Purpose of the Study:

• To address challenges with data below detection limits in environmental monitoring.
• To develop and illustrate a statistical method for analyzing censored environmental data.

Main Methods:

• Application of a likelihood ratio test for type I censored data.
• Assumption of log-normal distribution for trace contaminant concentrations.
• Evaluation of spatial variability of contaminants in the Niagara River.

Main Results:

• Derivation of a statistical test suitable for environmental data with values below detection limits.
• Demonstration of the test's utility in assessing spatial contaminant patterns.
• Improved analysis of trace contaminant bioavailability and source identification.

Conclusions:

• The developed statistical method effectively handles censored environmental data.
• This approach enhances the understanding of trace contaminant distribution in aquatic ecosystems.
• Provides a robust tool for environmental agencies in monitoring and source tracking.