Aquifer vulnerability assessment to heavy metals using ordinal logistic regression

Area of Science:

• Environmental Science
• Hydrogeology
• Statistical Modeling

Background:

• Groundwater contamination by heavy metals is a significant environmental concern.
• Understanding the factors influencing heavy metal occurrence is crucial for water resource management.
• Existing methods may not fully capture the probabilistic nature of heavy metal presence.

Purpose of the Study:

• To develop and validate a methodology for predicting heavy metal occurrence in groundwater.
• To identify key influencing variables on heavy metal contamination.
• To assess the economic impact of land use changes on heavy metal occurrence.

Main Methods:

• Ordinal logistic regression was employed to model the probability of heavy metal occurrence.
• The model incorporated variables such as land use, soil hydrologic group (SHG), elevation, clay content, hydraulic conductivity, and well depth.
• The methodology was applied to the Sumas-Blaine Aquifer in Washington State.

Main Results:

• Predicted probabilities of heavy metal occurrence aligned with observed data.
• Aquifer vulnerability varied for different heavy metals, influenced by distinct variable sets.
• Land use and SHG were consistently strong predictors for all heavy metals studied.

Conclusions:

• The developed model effectively predicts heavy metal occurrence and provides insights into contributing factors.
• Land use and SHG are critical factors in heavy metal contamination of groundwater.
• The study demonstrates the utility of statistical modeling for groundwater quality assessment and management, including economic evaluations.