Modeling Hydrostratigraphy on Quadra Island, BC

This work was motivated by the Quadra Island Climate Action Network (ICAN) and their Water Security Team, who are concerned about water resources. The Quadra Island communities wish to take an active role in community management of groundwater resources and require basic information about the local groundwater resources.

As is often the case for small communities in British Columbia, limited information on groundwater resources is available. In the case of Quadra Island, two aquifers are mapped and recorded in the Provincial Databases. This mapping consists of polygons and basic descriptions of materials and hydraulic characteristics (aquifer vs. aquitard; aquifer type: confined or unconfined).

The goals of this project were to improve on the existing aquifer mapping by:

1) Developing a three-dimensional subsurface model of hydrostratigraphic units

2) Determining the characteristics of each hydrostratigraphic units, including thickness, extent, geometry, and stratigraphic relationships between units.

3) Performing basic modeling of groundwater flow and assessing the potential vulnerability of mapped aquifers

Quadra Island consists of three main lithologies: Triassic age basalt (Karmutsen Formation), Limestone (Quatsino Formation), and Plutonic rocks of the Coast Belt intruding some of the younger units. Within the study area, bedrock is largely obscured by Quaternary-age sediments. Bedrock crops out as individual nobs and, progressively, as more continuous outcrops in the northern portions of the study area. Based on the regional continuity of bedrock units and some well records, it is assumed that Karmutsen basalt occurs below the Quaternary sediments on the southern peninsula of Quadra Island. Limestone, interbedded within the basalt, may occur in the northeast portion of the study area.

The southern peninsula of Quadra Island is covered in a succession of unconsolidated sediments with a thickness ranging from 1-68 m in total. These sediments represent deposition during glacial and interglacial periods. Most sediments on Quadra Island are associated with the onset, climax, and decay of the Cordilleran Ice Sheet during Late Wisconsin. Holocene non-glacial sediments also occur on the landscape.

On Quadra Island, groundwater occurs within these sedimentary deposits. Permeable water-bearing units (aquifers) and less permeable units (aquitards) reflect the sedimentary characteristics of the subsurface materials. Sediments are grouped into lithostratigraphic units consisting of varied materials that reflect changing depositional environments during the phases of glaciation. The lithostratigraphic units encountered in this study, and their characteristics are summarized below.

This study integrates subsurface data from water well logs accessed through  iMap BC . Subsurface materials were also sampled and characterized during field visits to Quadra Island where sediments are exposed along coastal cliffs, gravel pits and roadside exposures.

Groundwater well logs are the primary data source for this project. Well logs provide a description of geologic materials (and their depth below surface) encountered during drilling, as reported by the driller. Well log descriptions have variable reliability due to numerous factors, including inconsistent geologic terminology. For this reason, well log descriptions need to be standardized to ensure consistent use of geologic descriptors and reliable stratigraphic occurrence of the described materials.

A standardization scheme was developed to ensure consistent materials descriptions and to ensure appropriate modeling of the subsurface materials. The standardization scheme used in this project was modified from Russell et al. (1998) and uses the first descriptor of materials as the dominant material to be modeled. Additional descriptors were considered in the assignment of materials to be modeled. Materials were also assigned to known lithostratigraphic units as part of the standardization process. This assignment was informed by the regional Quaternary stratigraphic context and by field observations of materials.

In some cases, assigning stratigraphic units is relatively simple. For example, till has diagnostic and relatively consistent sedimentologic characteristics and it can be reliably assigned to the Vashon Drift stratigraphic unit. Other materials, like sandy intervals, are more equivocal in terms of their stratigraphic units. In such cases, assignment was informed by interval thickness, relative position below surface, and position relative to more reliable materials like till. For example, if a sandy interval was found below a till section, the sandy section was assigned to the Quadra Sand unit. Alternatively, if a sandy section was found above a till section, the sandy section was assigned to a stratigraphic unit above the Vashon Drift - in this case Capilano Sediments or Salish Sediments.

Hydraulic conductivity (K) is a measure of the ease with which water can flow through a porous material (bedrock and sediment), where high values represent easier flow. This is an important characteristic because it contributes to determining the viability of water extraction from an aquifer. Hydraulic conductivity is a first-order determinant on whether a stratigraphic unit is considered an aquifer or an aquitard.

Hydraulic conductivity is a difficult parameter to determine. However, it can be estimated from grain size distributions obtained from sieved sediment samples. Two separate methods were used to estimate the hydraulic conductivity of five samples (three from Quadra Sand units, two from Vashon till units).

Method 1: Hazen Method

The Hazen method determines hydraulic conductivity using the Hazen equation:

K=C(D ₁₀ ) ² 

Where,

• K = Hydraulic conductivity
• C = Hazen coefficient
• D ₁₀  = Effective grain size

Effective grain size (D ₁₀ ), the grain size that contributes the most to fluid flow, is the diameter at which 10% of the material is finer by weight.

The Hazen coefficient (C) is determined using a simple analytical method. First, a “uniformity coefficient (C _u )” is calculated by the ratio of D ₆₀ /D ₁₀ . A C _u  value greater than 6 indicates that the material is poorly sorted and a value less than 4 indicates that the material is well sorted. Subsequently, the material is classified as fine, medium, or coarse. Both pieces of information are thereafter used in the following table (figure 1) to determine C.

Method 2: Alyamani and Sen Method

 The Alyamani and Sen method was used to compare the Hazen Method results and to assess the variability of results from different methods. 

The Alyamani and Sen equation is:

K (m/day)= β [I _o +0.025(d ₅₀ -d ₁₀ )] ² 

Where,

• is a constant with a value of 1300
• I _o  represents the intercept in millimeters of a line formed by the d50 and d10 values along the grain-size axis.

Estimates of hydraulic conductivity for Quadra Sands samples range from 0.016 cm/s to 0.072 cm/s, and reflect grain size variability within this unit. Both methods produce comparable results for sand samples. Overall, the estimated values fall within known ranges for these materials.

Conductivity estimates for Vashon till samples range between 0.0019 and 0.0046 cm/s. while published values for glacial till range between 10 ^-10  to 10 ^-4  cm/s.

Subsurface materials are represented as three-dimensional surfaces that capture the extent, thickness and geometry of the stratigraphic units. In the images below, the materials are represented according to their Quaternary stratigraphic units. The modeling was performed using the LeapFrog software provided by Seequent.

Hydrostratigraphic units are assigned based on the hydraulic conductivity of respective materials. In this case, the materials of Quaternary stratigraphic units correlate strongly to the hydrostratigraphic units, and can be summarized as follows:

Confined Aquifer: Quadra Sand is the most extensive aquifer and is considered to be largely confined due to its depth, its stratigraphic position, and a near-continuous cover of Vashon till. Preliminary examination of static water level (SWL) reported in some water well logs indicates that SWL is frequently above the elevation of the upper boundary of this aquifer.

Aquitard: Vashon till is considered an aquitard due to it's comparatively low hydraulic conductivity. It is regionally extensive, though variable in thickness. Importantly, till on Quadra Island has variable grain size and exhibits varying degrees of stratification. Its hydraulic conductivity is likely to be more variable than the provided estimates can capture. and this reflects the heterogeneity of this unit at a spatial scale beyond what can be determined by grain size analysis.

Unconfined aquifer: Capilano sediments (and Salish sediments in isolated places) form an unconfined aquifer near the ground surface. The thickness and extent of this aquifer is highly variable.

Water well logs are the basis for developing this subsurface model. Well distribution is uneven and frequently forms clusters. This has important implications for interpreting the extent of modeled subsurface units since some modeled areas are not well constrained by well data. In other cases, clustered wells provide very good constraints on the interpolated surfaces.

Over 400 water well logs were examined to develop this model. Of those, 140 were of sufficiently high quality to be standardized. The LeapFrog model was developed using 130 standardized well logs. The remaining 10 well logs, and sediment exposures, were used to validate the interpolated surfaces.

LeapFrog models subsurface units by defining unit contacts documented in the well records. Contacts are then mathematically interpolated to define surfaces that join the gaps between wells. Based on known geologic relationships between units, modeled surfaces can be represented in relation to each other. From this, the volume of each unit can be represented, revealing the overall geometries, extents and distributions of hydrostratigraphic units.

The patchy nature of the well data means that the accuracy of the model varies greatly. We are confident that the model is accurate in areas with high well densities, but the areas between the clusters are a generalization, and likely miss some of the inherent subsurface variability. There are several areas where the model predicts 'holes' in the layer of Vashon till, exposing the Quadra Sand aquifer, but there is no well data to confirm if this is accurate. More field data are needed to assess the areas between well clusters. In areas of high data density, more detailed modeling of subsurface materials and variability could be performed.

The extent of each hydrostratigraphic units is represented below:

Groundwater flow modeling offers basic visualization of subsurface water flow paths. Basic groundwater flow modeling was performed using TopoDrive software. The method uses graphical modelling to represent the topographic surface and geological materials. The basis of the model is the topographic relief which determines the distribution of hydraulic head. Subsurface materials are represented and assigned hydraulic conductivity and porosity values.

This software provides preliminary representations of groundwater flow paths. It should be noted that this software is designed to model unconfined aquifers where hydraulic head is assumed to mimic surface topography. In this project, the confined aquifer (Quadra Sand materials) was also modeled. It is assumed that head distribution in this confined aquifer also mimics surface topography. This is probably not accurate and any modeled flow paths for the confined aquifer need to be examined in light of this considerable limitation.

Groundwater flow was modeled along two cross-sections. Model parameters were determined as follows:

• The topographic surface was derived from LiDAR, sourced from OpenLiDAR BC.
• Groundwater well logs and LeapFrog modeled surfaces were used to define the thickness and extent of subsurface geological units.
• Hydraulic properties were assigned to each unit based on hydraulic conductivity estimates and reported values of porosity for these materials.

Model Assumptions: 

• Top and side boundaries are no-flow boundaries
• The surface of the water table matches the topographic surface

Limitations:

• Groundwater flow is forced to follow the side and bottom boundaries,
• Designed for unconfined aquifers and not designed for confined aquifers

The DRASTIC method is a standardized way to assess aquifer vulnerability to soluble pollutants on the surface. DRASTIC is an acronym meaning: Depth to water, net Recharge, Aquifer media, Soil media, Topography, Impact of vadose zone, and hydraulic Conductivity.

DRASTIC computes vulnerability by assigning a value and a pre-determined weight to each parameter.

In the form of an equation, this is represented as:

(Dw × Dv) + (Rw × Rv) + (Aw × Av) + (Sw × Sv) + (Tw × Tv) + (Iw × Iv) + (Cw × Cv).

The final sum of this equation is then compared to known aquifer vulnerability values (index). 

The DRASTIC method was used to compute vulnerability for the unconfined and the confined aquifers on Quadra Island. The resulting vulnerability ratings are 130 and 160 for the confined and unconfined aquifers, respectively.

These ratings are most influenced by the hydraulic parameters of the aquifers, the high recharge, and the relatively shallow position of the water table. However, DRASTIC is ultimately designed to assess the potential for surface contamination, and is predicated on the presence of contaminant sources. Quadra Island has minimal industrial areas and limited land-use leading to significant surface contaminants.

Therefore, despite moderate-high DRASTIC ratings, the true vulnerability is much less due to a general absence of contaminant sources.

• This project has refined the three-dimensional mapping and characterization of hydrostratigraphic units on the southern peninsula of Quadra Island.
  • Aquifers and aquitards are now better represented in terms of their extent, thickness, geometry and subsurface relationships to other hydrostratigraphic units.
  • Most drilled wells obtain water from the lower confined aquifer (Quadra Sands)
  • Most shallow/dug wells obtain water from the upper unconfined aquifer (Capilano Sediments).

• The regional-scale model provides a general view of hydrostratigraphic units but does not capture all the subsurface variability of materials.
  • Smaller aquifers may also occur (Aquifer 974)
  • The main aquitard (Till) has variable texture that is not well represented and likely affects its hydraulic conductivity.
  • The interpolated surfaces could be improved with additional well data in areas where large data gaps occur.

• DRASTIC vulnerability rating is moderate-high and on par with other areas of coastal Vancouver Island.
  • This rating must be considered in light of the fact that Quadra Island has very limited sources of surface contaminants.
  • The true vulnerability to surface contaminants is therefore much lower.

• This subsurface model is a framework for future work
  • More refined subsurface modeling could be performed in areas of higher data density.
  • More refined groundwater flow modeling could be performed to better represent flow pahs in the confined aquifer.

Aller, L. Bennett, T. Lehr, J. Petty, R. and Hackett, G. (1987) DRASTIC: A Standardized System for Evaluating Ground Water Pollution Potential Using Hydrogeologic Settings. United States Environmental Protection Agency, 641 Pg.

Alyamani, M.S. Zekâi, Ş. 1993. Determination of Hydraulic Conductivity from Complete Grain-Size Distribution Curves. Groundwater. 551-555https://doi.org/10.1111/j.1745-6584.1993.tb00587.x

Hazen, A. (1892), Physical Properties of Sands and Gravels With Reference to Their Use Infiltration, Massachusetts State Board of Health, Boston, Mass

Hsieh, P. (2020) TopoDrive and ParticleFlow - Online Version. https://tdpfonline.net/index.html

Newton, P. and Gilchrist, A. (2012) Technical Summary of Intrinsic Vulnerability Mapping Methods for Vancouver Island, Vancouver Island Water Resources Vulnerability Mapping Project –Phase 2. Vancouver island University, 51pp.

Roddick, J.A., Woodsworth, G.J. (2006) Geology, Bute Inlet, British Columbia, Geological Survey of Canada, Open File 5073. https://doi.org/10.4095/221570

Russell, H.A.J. , Brennand, T.A., Logan, C. and Sharpe, D.R. (1998) Standardization and assessment of geological descriptions from water well records, Greater Toronto and Oak Ridges Moraine areas, southern Ontario. Current Research 1998-E; Geological Survey of Canada, p. 89–102.

Woessner WW., Poeter EP. 2020. Hydrogeologic Properties of Earth Materials and Principles of Groundwater Flow. The Groundwater Project, Guelph, Ontario, Canada.