The effects of numerical-model complexity and observation type on estimated porosity values

J Jeffrey Starn,Christopher T Green,Amvrossios C Bagtzoglou

doi:10.1007/s10040-015-1289-3

J Jeffrey Starn, Christopher T Green + Show 1 more

Open Access

https://doi.org/10.1007/s10040-015-1289-3

Copy DOI

Abstract

The relative merits of model complexity and types of observations employed in model calibration are compared. An existing groundwater flow model coupled with an advective transport simulation of the Salt Lake Valley, Utah (USA), is adapted for advective transport, and effective porosity is adjusted until simulated tritium concentrations match concentrations in samples from wells. Two calibration approaches are used: a “complex” highly parameterized porosity field and a “simple” parsimonious model of porosity distribution. The use of an atmospheric tracer (tritium in this case) and apparent ages (from tritium/helium) in model calibration also are discussed. Of the models tested, the complex model (with tritium concentrations and tritium/helium apparent ages) performs best. Although tritium breakthrough curves simulated by complex and simple models are very generally similar, and there is value in the simple model, the complex model is supported by a more realistic porosity distribution and a greater number of estimable parameters. Culling the best quality data did not lead to better calibration, possibly because of processes and aquifer characteristics that are not simulated. Despite many factors that contribute to shortcomings of both the models and the data, useful information is obtained from all the models evaluated. Although any particular prediction of tritium breakthrough may have large errors, overall, the models mimic observed trends.

Highlights

Background and motivationModel uncertainty in transport simulations is affected by how transport parameters are represented in the model and by the number and type of data available for model calibration
Tritium concentrations above background levels have persisted in the Salt Lake Valley aquifer, Utah (USA) for decades, a time period over which changes in total dissolved solids concentrations have been observed (Starn et al 2014); one would expect that calibration to tritium data should serve to decrease uncertainty in model predictions of changes in water quality
The approach used in this study is to convert an existing groundwater flow model to simulate advective transport in the Salt Lake Valley, Utah (Fig. 1), by adjusting effective porosity to match tritium concentrations in samples from wells and, alternatively, interpreted tritium/helium apparent ages (Starn et al 2014)

Summary

Introduction

Background and motivationModel uncertainty in transport simulations is affected by how transport parameters are represented in the model and by the number and type of data available for model calibration. Tritium concentrations above background levels have persisted in the Salt Lake Valley aquifer, Utah (USA) for decades, a time period over which changes in total dissolved solids concentrations have been observed (Starn et al 2014); one would expect that calibration to tritium data should serve to decrease uncertainty in model predictions of changes in water quality.

Results

Conclusion