Modeling contaminant concentration distributions in China's centralized source waters

Area of Science:

• Environmental Science
• Water Quality Assessment
• Statistical Modeling

Background:

• Understanding contaminant occurrences in China's centralized source waters is crucial for effective water quality management.
• Current single-factor assessment methods offer limited insight into the national extent of specific contaminant issues.
• Scientifically sound management strategies require detailed information on contaminant concentration distributions.

Purpose of the Study:

• To present a Bayesian hierarchical modeling approach for estimating contaminant concentration distributions in China's centralized source waters.
• To assess the national probabilities of arsenic and fluoride exceeding regulatory standards.
• To explore the application of Bayesian methods for source water quality information systems.

Main Methods:

• Utilized data from the 2006 national census of centralized source waters in China.
• Employed a Bayesian hierarchical modeling approach to estimate contaminant concentration distributions.
• Applied three common source water stratification methods to establish alternative hierarchical structures.

Main Results:

• Estimated the national probability of arsenic exceeding 0.05 mg/L at 0.96-1.68%.
• Estimated the national probability of fluoride exceeding 1 mg/L at 9.56-9.96%.
• Observed strong spatial patterns in the occurrences of both arsenic and fluoride exceedances.

Conclusions:

• The Bayesian hierarchical model provides a robust method for characterizing contaminant occurrences in source waters.
• The findings highlight specific risks associated with arsenic and fluoride contamination nationally.
• The approach supports the development of data-driven source water quality management systems.