Ring of Fire (RoF) = large mineral deposits of Ni/Cu/Zn/Cr and PGM: Located in one of the world's largest peatland system; Sensitive to climate change (Hadley et al., 2019) and anthropogenic stresses (Leclair et al., 2015) Environmental conditions: Additional knowledge required to understand how baseline conditions respond to climate change and new anthropogenic stresses (e.g., roads, mining camps,...); Natural presence and behavior of metal(loid)s in this system needs to be carefully assessed prior to any development; Changes to groundwater flow dynamic, changes in lake sediment conditions and forest fires can potentially enhance the remobilization of metal(loid)s over time; Explore and develop environmental indicators adapted to the RoF environment.

Quasi-point hydraulic properties (K) measured locally under laboratory or field conditions need to be upscaled to block-scale K-tensors for use in flow simulators. The upscaled model also needs to be calibrated to hydraulic head observations. The calibration must preserve spatial covariance, cross-covariance and non-linear relations between tensor components. We apply a new upscaling method that allows to compute and model the covariance between block K-tensor components. We use a gradual deformation method for calibration of simulated K-tensor fields to measured head data. Our method incorporates a new bivariate transform that preserves the non-linear relations between K-tensor components. The ensemble of calibrated realizations allows quantification of uncertainty of groundwater flow models. A comparison with PEST on a test case defining capture zones for water supply wells shows that our method calibrates better to measured heads than PEST, provides more realistic K-tensors and results in larger capture zones.

The focus of this study is to explore and assess the uncertainty related to geological heterogeneity in regional groundwater flow models. A stochastic approach is adopted in which the hydrostratigraphic architecture and respective hydrofacies variability directly controls the three-dimensional hydraulic conductivity (K) field. The following procedural steps were applied to develop the stochastic flow model: 1) generation of multiple equivalent hydrostratigraphic models honouring observed hydrostratigraphic units and stratigraphic transitions; 2) determination of the statistical scalar K distribution for each hydrostratigraphic unit along with the respective spatial correlation at the local-scale; 3) upscaling of the scalar K field (quasi-point) to its full 3-D tensor at the regional scale (block); 4) determination of the K-tensor structures and simulation of the K-tensor for hydrogeological models; 5) calibration of the models by numerical inversion of the K-tensor and preserving their component correlations including the recharge rate; and 6) quantification of the uncertainties in the groundwater flow model. The capability and performance of the suggested workflow is illustrated using a case study in southern Ontario. Results highlight the efficiency of the method in assessing model uncertainties and their impact on regional groundwater flow models.

Groundwater represents an important source of water for agriculture in the Innisfil Creek watershed, southern Ontario, where water vulnerability issues were evident through recent drought conditions. Uncertainties remain regarding the groundwater flow pattern of the multi-layer aquifer system and further understanding is required to plan for sustainable groundwater use. A hydrogeological study based on geochemistry and isotopes (d2H, d18O, d13C,14C-DIC, and 226Ra) is being developed to better characterize and quantify recharge, regional flow patterns, and potential inter-aquifer flow. In this paper, we present on work completed in 2018. The study objectives were to gather existing hydrogeological and geochemical data into a comprehensive database, to analyse the data in order to support targeted additional water sampling, and to develop preliminary interpretations. The ongoing project is now set for the development of an initial conceptual model based on the geochemical and isotopic interpretations combined with statistical analyses and physical hydrogeology processes observed in the region.

Variations of hydraulic conductivities (K) in hydrostratigraphic systems may significantly affect the flow velocity field and mass dispersion. Field data used for assessment of K are generally representative at a local scale only and for one or a few hydrofacies forming a specific hydrostratigraphic unit (HSU). Regional groundwater flow systems encompass a multitude of HSUs, where a given HSU can have K-range variability spanning several orders of magnitude. Therefore, for regional systems, blocks with respective K-equivalent have to be defined as part of each HSU. A major challenge under such conditions is to determine the spatial distribution of the full 3D K-tensor blocks (Kb) considering the effect of local scale variability in K. In this study, an efficient method is developed for regional characterization of the HSUs with 3D Kb. The method was tested for the Innisfil Creek watershed. For each HSU, it consists of the following major steps: (i) assessment of K from measured data; generation of the probability density function of K; definition of the spatial covariance of K; and geostatistical simulation of K at local scale; (ii) upscaling of the local scale realizations of K into full 3D Kb; definition of the spatial covariance of 3DKb; and definition of spatial distribution of the 3DKb. The preliminary results include a K database built using 1694 grain size analyses, 32 HSU borehole samples and 1086 transmissivity measurements in public wells. The grain size samples were collected by Ontario Geological Survey (OGS) from 15 boreholes in the South Simcoe area. Between 28 and 301 samples/HSU were analysed covering all of the 14 HSUs observed in the study area. K from grain size analyses were in good agreement with K based on laboratory permeability test measurements of field core. The database was completed with K assessments from specific capacity analyses in public wells, which have a strong bias from the high permeability formations. To obtain the local scale variability of K, only the results from grain size analyses were used mainly due to their sufficient quantity to define the probability density function of each HSUs and their results reflecting the expected strong variability of hydrofacies within a given HSU. Local scale K fields were simulated with non-conditional turning band simulation. These results were then upscaled to the block regional scale. In that regards, we revisited the Zhou et al. (2010) methodology for non-local 3D hydraulic conductivity full tensor upscaling using flow simulator. Preliminary results suggest no correlation between upscaled blocks within the study area. The upscaling methodology is still in development, upscaling parameters and validation criteria need to be tested. The final outputs of this study will be an ensemble of hydrostratigraphic models with equivalent 3D K-tensor parameters, which will be used to assess the impact of K and HSU uncertainties on groundwater flow modeling. The proposed methodology is appropriate for characterizing the uncertainty of groundwater flow and transport. For example, it can be used for aquifer vulnerability assessment and the delineation of wellhead protection areas.

Understanding the geological uncertainty of hydrostratigraphic models is important for risk assessment in hydrogeology. An important feature of sedimentary deposits is the directional ordering of hydrostratigraphic units (HSU). Geostatistical simulation methods propose efficient algorithm for assessing HSU uncertainty. Among different geostatistical methods to simulate categorical data, Bayesian maximum entropy method (BME) and its simplified version Markov-type categorical prediction (MCP) present interesting features. In particular, the zero-forcing property of BME and MCP can provide a valuable constrain on directional properties. We illustrate the ability of MCP to simulate vertically ordered units. A regional hydrostratigraphic system with 11 HSU and different abundances is used. The transitional deterministic model of this system presents lateral variations and vertical ordering. The set of 66 (11 × 12/2) bivariate probability functions is directly calculated on the deterministic model with fast Fourier transform. Despite the trends present in the deterministic model, MCP is unbiased for the HSU proportions in the non-conditional case. In the conditional cases, MCP proved robust to datasets over-representing some HSU. The inter-realizations variability is shown to closely follow the amount and quality of data provided. Our results with different conditioning datasets show that MCP replicates adequately the directional units arrangement. Thus, MCP appears to be a practical method for generating stochastic models in a 3D hydrostratigraphic context.

The Nanaimo Lowland Groundwater Study is a collaboration between the British Columbia (BC) Ministry of Environment (MoE), Ministry of Forests, Lands, and Natural Resource Operations (MFLNRO), the Regional District of Nanaimo (RDN), and the Geological Survey of Canada (GSC). This Open File reports on the three-dimensional (3D) groundwater flow model (3DGFM) for the Nanaimo Lowland aquifer system developed to support improved sustainable groundwater management of the Nanaimo Lowland. The study area represents a coastal strip running from Nanoose to Deep Bay in the eastern part of Vancouver Island and covers approximately 580 km2. The groundwater flow model was based on the 3D hydrostratigraphic model (3DHSM) developed by Benoit et al. (2015). The hydrostratigraphic units were converted into a numerical grid and representative hydraulic properties were assigned based to each unit on statistical analysis of available values. The groundwater resurgences present along the steep slopes of the main riverbanks and at the shoreline were simulated with seepage boundary conditions allowing for a vertical hydraulic link between the main Quadra Sand aquifer and surface water bodies. Hydraulic conductivity and groundwater recharge rate were further calibrated through numerical inversion using static hydraulic heads recorded in existing wells and estimated baseflow values from streamflow records. Simulation results that two distinct groundwater flow systems: (1) the Quadra aquifer system with groundwater divides closely matching watershed boundaries containing relatively fresh water; and (2) the extensive bedrock aquifer system with older water and a flow patterns generally directed toward the sea coast. Simulations also reveal discharge to rivers and streams mainly from the Quadra Sand aquifer and the Capilano sediments, often locally present along the main river channels, with low contribution from the bedrock. Exception is the Englishmen River where the Quadra Sand is inexistent and the river is in direct hydraulic contact with the bedrock aquifer. The Quadra Sand aquifer is naturally protected from sea intrusion due to the higher altitude with respect to sea level, whereas the bedrock aquifer appears vulnerable to encroachment especially under heavy pumping conditions close to the shoreline. This 3DGFM provides then a meaningful framework for the understanding of the regional hydrogeology of Nanoose-Deep Bay aquifer system that would contributes to support sustainable management of groundwater resources.