Simulating solute transport in a structured field soil: uncertainty in parameter identification and predictions

Area of Science:

• Environmental science
• Soil science
• Hydrology

Background:

• Dual-permeability models are crucial for understanding contaminant transport in structured soils, but parameter estimation challenges limit their application.
• Macropore flow significantly influences solute movement, necessitating accurate model parameterization.

Purpose of the Study:

• To evaluate data requirements for parameter identification in dual-permeability modeling using the MACRO model.
• To compare Sequential Uncertainty Fitting (SUFI) and Generalized Likelihood Uncertainty Estimation (GLUE) for parameter estimation and uncertainty analysis.

Main Methods:

• Application of the MACRO dual-permeability model to a field data set of bromide and bentazone transport in a structured soil.
• Comparison of SUFI and GLUE methods for parameter conditioning using various observation combinations.
• Investigation of six parameters controlling macropore flow, pesticide sorption, and degradation.

Main Results:

• Both resident and flux concentrations are necessary for well-conditioned and unbiased parameters; tracer transport data improve macropore flow parameter conditioning.
• GLUE exhibited wider parameter ranges than SUFI, but similar 5th-95th percentile uncertainty ranges.
• SUFI's neglect of parameter correlations led to incorrect posterior uncertainty domains and larger prediction uncertainty ranges for bentazone losses compared to GLUE.

Conclusions:

• GLUE is better suited for uncertainty estimation in predictive modeling with dual-permeability models compared to SUFI.
• Accurate parameter identification for macropore flow and pesticide fate requires comprehensive field data, including both tracer and pesticide concentrations.