Phosphorus load estimation in the Saginaw River, MI using a Bayesian hierarchical/multilevel model.

Area of Science:

• Environmental science
• Statistics
• Water resource management

Background:

• Estimating annual riverine phosphorus loads is crucial for water quality management.
• Traditional methods face challenges with varying flow-concentration relationships and small sample sizes.

Purpose of the Study:

• To develop and apply a Bayesian hierarchical ratio approach for estimating annual riverine phosphorus loads in the Saginaw River.
• To explicitly quantify uncertainty in long-term load estimations.

Main Methods:

• Utilized a Bayesian hierarchical/multilevel ratio approach combining ratio estimation with Bayesian modeling.
• Employed posterior predictive distributions to address prediction uncertainty.
• Leveraged hierarchical techniques to overcome small sample size limitations by borrowing strength across years.

Main Results:

• A slight decrease in total phosphorus load was observed over the study period (1968-2008).
• The flow-weighted concentration (ratio parameter) showed a clearer decreasing trend, mitigating noise from flow variations.
• Saginaw Bay's phosphorus load likely exceeded the 440 tonnes/yr target in most years, with probabilities of non-compliance estimated at 1.00, 0.50, 0.57, and 0.36 for 1977, 1987, 1997, and 2007, respectively.

Conclusions:

• The Bayesian hierarchical ratio model provides a robust framework for estimating long-term riverine phosphorus loads with quantified uncertainty.
• The model demonstrates good goodness-of-fit to observational data.
• This approach effectively reduces uncertainties associated with small sample sizes in parameter and load estimations, improving reliability for water quality assessments.