Predicting groundwater redox status on a regional scale using linear discriminant analysis

Area of Science:

• Environmental Science
• Hydrogeology
• Geochemistry

Background:

• Denitrification, crucial for removing nitrates, requires reducing subsurface conditions.
• Groundwater redox status is a key indicator for identifying potential nitrate reduction zones.
• Understanding redox conditions is vital for managing groundwater quality and nutrient cycling.

Purpose of the Study:

• To classify groundwater redox status in contrasting New Zealand regions.
• To relate groundwater chemistry and redox status to mappable geological and soil factors.
• To develop predictive models for subsurface reduced groundwater zones.

Main Methods:

• Groundwater samples from 568 (Waikato) and 2223 (Canterbury) wells were analyzed for redox status.
• Discriminant analysis was employed to link redox status with geology, topography, and soil characteristics.
• Models were developed and validated for different well depths (<25m, 25-100m, >100m).

Main Results:

• Waikato: 18% reduced, 18% mixed, 64% oxic conditions. Canterbury: 5% reduced, 10% mixed, 84% oxic conditions.
• Reduced groundwater prevalence increased with well depth in both regions.
• Predictive models achieved 63% (Waikato) and 70% (Canterbury) agreement between predicted and measured redox status.

Conclusions:

• Geological and soil factors significantly influence groundwater redox status.
• GIS-based predictive models enhance spatial identification of reduced groundwater zones.
• Improved prediction of reduced zones, combined with flow path analysis, refines denitrification estimates.