Related Experiment Videos
Bacterial density in water determined by poisson or negative binomial distributions
A H El-Shaarawi1, S R Esterby, B J Dutka
1National Water Research Institute, Canada Centre for Inland Waters, Burlington, Ontario, Canada L7R 4A6.
Applied and Environmental Microbiology
|January 1, 1981
Summary
This study explores bacterial density in water using Poisson and negative binomial distributions. It provides methods to characterize bacterial counts, especially for varied sampling conditions, improving water quality assessments.
Area of Science:
- Environmental microbiology
- Statistical ecology
Background:
- Bacterial density in water is crucial for ecological assessment.
- Accurate characterization requires appropriate statistical models for count data.
- Variability in sampling locations and times complicates density estimation.
Purpose of the Study:
- To evaluate Poisson and negative binomial distributions for characterizing bacterial density from small-volume water samples.
- To develop and illustrate a procedure for managing bacterial count variability.
- To compare the implications of using different distributions on sterile sample probability.
Main Methods:
- Applied Poisson distribution for controlled, replicate analyses.
- Applied negative binomial distribution for spatially and temporally varied samples.
- Developed a procedure to group bacterial counts into homogeneous subsets for negative binomial data.
- Calculated probabilities of sterile samples under different distribution assumptions.
Main Results:
- Poisson distribution is suitable for highly controlled replicate samples.
- Negative binomial distribution is appropriate for diverse sampling conditions.
- The grouping procedure effectively characterizes variability in bacterial counts.
- Distribution choice significantly impacts the probability of observing sterile samples.
Conclusions:
- The choice of statistical distribution (Poisson vs. negative binomial) is critical for accurate bacterial density assessment in water.
- The negative binomial model and associated grouping procedure offer a robust method for handling complex environmental sampling data.
- Understanding these statistical nuances enhances the reliability of water quality monitoring and ecological studies.