Improving groundwater flow model parametrization techniques
Date
2020-11-13
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Numerical modeling of groundwater flow is a powerful tool that can assist in hydrogeological analysis and understanding of groundwater system behavior. However, its results are highly dependent on the proper model parameterization (often, at least partially achieved through calibration). The situation is further complicated in transient groundwater models requiring calibration of the additional parameters such as specific storage in addition to standard hydraulic conductivity. Another complication of the transient modelling is a need to define transient boundary conditions. In settings with seasonal recharge variations, the spatial and temporal variation in the recharge boundary conditions becomes an important part of the problem setting. Simultaneous calibration of hydrogeological properties (hydraulic conductivity, specific storage) and recharge is a challenging problem prone to non-uniqueness: similar model results achieved by different combinations of recharge, specific storage and transmissivity. The objective of this study is to improve hydrogeological model parametrization, with the focus on hydrogeological properties governing groundwater system response to transient forcing. It highlights multiple ways of estimating specific storage, which allows to provide better constraints for the calibration problem. In particular, it proposes a new method of specific storage estimation using seismic velocities and compares it with the results of conventional pumping test analysis and analysis of a water level response to atmospheric pressure fluctuations and Earth tides. The resulting values of specific storage have shown reasonable agreement between methods for a sandstone unit of the Paskapoo Formation: the estimates from seismic velocity and water level analyses ranged from 1.25∙10-5 1/m to 1.6∙10-5 1/m, while the specific storage estimate from pumping test was 4.6∙10-5 1/m. This finding suggests that seismic velocity method can be used to evaluate spatial variation of specific storage or to augment the inversion as a part of parameter constraints or regularization term. This study also identified multiple avenues of extracting additional information about aquifers from conventional datasets in order to further reduce the uncertainty. The proposed new framework uses integration of multiple methods for inversion of pumping test in a heterogeneous aquifer. Its application at the study site has shown that the flow from the aquitards bounding the aquifer are a major component of a flow budget during even multi-day pumping test highlighting the limitations of extrapolating pumping test results to long-term transient response without accounting for aquifer geometry and properties of the bounding units. In addition, this study investigated sensitivity of modeled groundwater levels to transient recharge properties. The spatially-variable transient recharge was generated by Versatile Soil Moisture Budget (VSMB) model based on atmospheric forcing and was applied as a specified flux to the upper boundary of a watershed-scale 3-D groundwater model. The model reasonably replicated both observed groundwater level dynamics and 10-year mean baseflow of 20.54 mm/a (modeled recharge value was 23.84 mm/a). It was demonstrated that both seasonal groundwater level rise and long-term trends simultaneously reflect magnitude and timing of the recharge; seasonal propagation of a recharge pulse through the aquifer was dependent on hydraulic diffusivity (a ratio of hydraulic conductivity to specific storage). These findings imply that a detailed transient recharge signal is needed for both forward and inverse groundwater modeling.
Description
Keywords
hydrogeology, groundwater recharge, groundwater modeling, specific storage, pumping test inversion
Citation
Abdrakhimova, P. (2020). Improving groundwater flow model parametrization techniques (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.