Multimodel Validity Assessment of Groundwater Flow Simulation Models Using Area Metric Approach

Area of Science:

• Environmental modeling
• Hydrogeology
• Geoscience

Background:

• Model validation is crucial for reliable environmental assessments.
• Traditional binary validation (valid/invalid) can be limiting.
• Multimodel approaches are needed to address epistemic and aleatory uncertainties.

Purpose of the Study:

• To demonstrate the application of the Area Metric for multimodel validity assessment.
• To quantify the replicative validity of groundwater flow models.
• To rank multiple model representations based on their validity scores.

Main Methods:

• Application of the Area Metric developed by Ferson et al. (2008).
• Quantification of model-data agreement using empirical cumulative distribution functions (ECDFs).
• Ranking of multiple landfill groundwater flow model representations.

Main Results:

• The Area Metric effectively quantified the degree of replicative validity.
• Multiple model representations were ranked based on their Area Metric scores.
• The approach allows for blind assessment, avoiding model overfitting.

Conclusions:

• The Area Metric provides a robust framework for multimodel validity assessment.
• Explicit incorporation of epistemic and aleatory uncertainties enhances validation.
• This approach increases confidence in model representativeness and reduces user risk.