An evaluation of conditioning data for solute transport prediction

Area of Science:

• Geosciences
• Hydrogeology
• Environmental Engineering

Background:

• Subsurface characterization generates extensive data crucial for understanding groundwater flow and contaminant transport.
• Field research sites provide unique opportunities to evaluate data conditioning impacts on predictive models.

Purpose of the Study:

• To assess the influence of various data types and scales on the accuracy of subsurface flow and transport predictions.
• To compare model performance using different conditioning strategies, from simple to complex.

Main Methods:

• Utilized a three-dimensional numerical model to simulate bromide breakthrough curves (BTCs).
• Experimented with diverse hydraulic conductivity spatial distributions and conditioning data (small-scale measurements vs. geophysical interpretations).
• Evaluated six simulation cases, ranging from homogeneous to stochastic indicator simulations.

Main Results:

• Conditioning models with numerous small-scale measurements did not significantly enhance predictive accuracy and risked introducing bias.
• Geophysical interpretations, despite having larger spatial support, substantially improved prediction accuracy and precision.
• Model error remained a significant factor, comparable to parameter uncertainty.

Conclusions:

• The scale and type of conditioning data critically influence subsurface model reliability.
• Geophysical data integration offers a more effective approach for improving flow and transport predictions compared to dense, small-scale measurements.
• Acknowledging and quantifying model error is essential alongside parameter uncertainty analysis.