Variabilité chimique de l’eau souterraine

Microcline weathering

Mineralogical analyses performed on deep marine deposits in some areas of the SLSJ region showed that microcline and illite represent 41% and 21% of the argillaceous fraction of clay, respectively (Gravel; 1974). K-feldspar (microcline) transformation into the clay mineral (illite), is represented by:
According to the PCA results (Figure 5), SiO2 and K+ are related to samples in Cluster 4. The influence of SiO2 and K+ on the chemistry of Cluster 4 results from the chemical breakdown of microcline and it confirms that the weathering of silicate minerals occurs more efficiently within the regional aquitard than in the bedrock.

General geochemical evolution path in the SLSJ aquifer systems

The combination of HCA, PCA and binary graphs demonstrates that the occurrence of two distinct salinization paths depends on the hydrogeological context. Based on the content of major elements, the general geochemical evolution paths of groundwater within the SLSJ aquifer systems are presented in the Piper diagram in Figure 3.10. The samples are presented according to the HCA cluster to which they belong.
Cluster 2 samples (Ca,Na-HCO3) reflect the first stage of groundwater evolution as the cation composition changes from Ca2+ dominant to Na+ dominant, which is explained by ion exchange processes in the presence of the marine clay aquitard and further explained by the feldspar dissolution/precipitation kinetics in the bedrock. Then, the evolution of groundwater follows two possible paths.
The first path occurs as groundwater flows through the crystalline bedrock aquifers, which induces a change in the anion composition from HCO3- -dominant (Clusters 2) to Cl- -dominant end-members (Clusters 3 and 4). According to the first path (Path 1), groundwater tends to evolve by water/rock interactions toward a Ca-Cl end-member. This evolution is accompanied by the simultaneous increase of the Ca2+, Sr2+ and Ba2+ contents of groundwater.
The second path begins with Ca2+water-Na+mineral ion exchange and results in the transformation of the samples from Ca-HCO3 to Na-HCO3 (Figure 10; Path 2). Path 2 represents the evolution of groundwater due to a salinization path within the confined aquifers that are in contact with the regional aquitard and possible groundwater mixing with the Pleistocene Laflamme Sea end-member. The seawater could be trapped in the regional aquitard (solute diffusion) or may be stagnant (mixing) in some part of the aquifer (Cloutier et al., 2010). This latter combination of processes (cation exchange and/or leaching of saltwater trapped in the regional aquitard) is grouped in this study under the term “water/clay interactions”.
Ion exchange, solute diffusion from the Laflamme Sea clay aquitard and/or mixing with Laflamme seawater cause the evolution of Cluster 2 (recharge Ca-HCO3 groundwater to Na-HCO3 groundwater) to Cluster 4 (brackish groundwater Na-Cl in composition). Following this latter trend, the chemistry of the groundwater tends to evolve toward an end-member having a composition that is like that of present seawater, i.e., Na-Cl-rich. According to the PCA, this evolution is accompanied by a simultaneous increase of the Mg2+, SiO2, K+ and SO42- contents of groundwater.
Figure 3.11 is a generalized cross-section showing the different salinization pathways occurring in the SLSJ area. The topography caused by a graben structure (>1000 m) displays a connection between water composition and groundwater flow driven by the topographic gradient. Bedrock groundwater evolves from Cluster 2 (Ca,Na-HCO3, unconfined environment) to Cluster 3 (Ca,Na-Cl; rock dominated) by interactions with basement fluids (water-rock interactions, e.g. Cl-/Br- mass ratios of 88 derived from Figure 3.8) coming up along the graben fault system. The distribution of Cluster 3 samples along major faults in Figure 3.6 also suggests the upwelling of basement fluids toward the surface. In the crystalline basement, the dissolution/precipitation kinetics of feldspars induce a gradual changing of the chemical facies from Ca-HCO3 to Na-HCO3 (Cluster 2).
Confining conditions dominate Cluster 4 (Figure 3.4; Table 3.2). The groundwater in the granular aquifers exhibits an evolution from the recharge groundwater of Cluster 2 by ion exchange (Cluster 2 Na-HCO3) in a confined environment and possible mixing with the Laflamme seawater end-member (e.g., Cl-/Br- mass ratios of 266 derived from Figure 3.8; Cluster 4). This latter evolution might also be observed in bedrock aquifers where confining conditions prevail.
Cluster 1 samples are dominated by Na-Cl waters from confined bedrock aquifers (Figure 3.4). The samples from Cluster 1 are predominately distributed near limestone units (24/30 samples) (Figure 3.6) and may then represent a seawater end-member of groundwater evolving in contact within the confined Ordovician limestone.
The end-members of the salinization paths, represented by Clusters 3 and 4, are thus identified and can be distinguished based on the relative concentrations of the following trace elements: Ca2+, Sr2+, Ba2+ for Cluster 3 and Mg2+, SiO2, K+, SO42-, and HCO3- for Cluster 4 (Figure 3.5). Mixing and dilution (Figure 3.10) occur at different rates during the evolution of groundwater having different origins in response to the specific hydrogeological context prevailing locally in the region.

SUMMARY AND CONCLUSION

In this paper, groundwater evolution paths that account for the chemical characteristics of a large dataset obtained from groundwater analyses in a regional-scale study has been described. The dissolution of calcite and/or plagioclase minerals controls the chemical background of recharge groundwater. Anorthite plagioclase weathering is more effective in bedrock aquifers, and the Ca2+water-Na+mineral ion exchange process is dominant in confined aquifers. In addition, there is mixing with a seawater end-member where confining conditions by the regional aquitard are prevailing. In the crystalline bedrock, mixing with deep brackish groundwater by topographically driven groundwater flow appears to be a possibility in the Lake St. John area.
The combination of HCA, PCA and binary graphs identifies two distinct salinization paths occur: 1) an evolution by water/rock interactions; and 2) an evolution from the recharge groundwater by water/clay interactions and groundwater mixing. The first evolution path is specific to fractured rock aquifers. The term water/clay interactions was introduced in this paper to account for a combination of processes, namely: ion exchange and/or leaching of salt water trapped in the regional aquitard. Mixing with fossil seawater present in the aquifer might also increase the groundwater salinity.
Based on the results obtained using PCA, clusters can be distinguished according to the relative concentrations of the following trace elements: Ca2+, Sr2+ and Ba2+ for Cluster 3 (water/rock interactions) and HCO3-, Mg2+, SiO2, K+ and SO42- for Cluster 4 (water/clay interactions). One of the main processes controlling the groundwater chemistry in confined aquifers is believed to be the degradation of organic matter and/or any other reaction that produces CO2. Another process could be the silicate alteration process, more specifically microcline plagioclase weathering.
The percentage of cumulative variance obtained via PC 1 and 2 is 52.6%. This suggests that 50% of the variance of the dataset cannot be explained by the first two components (salinity). Other components need to be investigated to establish a more global portrait of regional hydrogeochemistry. Further work should also focus on identifying reliable tracers of the two salinization paths using Factorial Analysis (FA). In contrast to PCA, the emphasis of FA is explaining covariances instead of the total variance of a dataset. Thus, FA could be used to highlight highly correlated chemical parameters within the dataset and shed new light on the chemical nature of the samples.
Next steps should prioritize an improved knowledge of the different rock types and their geochemical properties. Compilation of the existing geochemical data on rocks and sediments should be coupled with the sampling of outcrops to investigate the mineralogical assemblages of the various regional rocks (via thin section analysis). Leaching experiments should also be performed to obtain different specific ionic ratios, as for instance the Cl-/Br- ratio. Finally, a new sampling campaign focused on isotopic data (18O, 2H, 13C, etc.) and residence time constraints could contribute to deciphering the data set, especially the origins of the saline end-members.
This study—derived from a routine chemical analysis of water quality—provides a new understanding of the chemical evolution of groundwater that is of interest in conceptualizing the dynamics of groundwater, such as flow patterns and hydraulic connectivity between aquifers. With a better understanding of the chemical characteristics of the salinization paths, the major, minor and trace element chemistry of groundwater becomes a relevant tool for investigating groundwater dynamics, such as the interconnectivity of aquifer systems or groundwater flow dynamics according to gravity-driven flow processes. With a better knowledge of the geochemistry of the porous and the fractured matrix of regional aquifers, as well as of the geochemistry of aquitards present in the SLSJ region, this study may leads to discussions on the spatial and temporal evolution of groundwater quality with respect to the prevailing geochemical paths (water/rock and/or water/clay interactions).

ACKNOWLEDGEMENTS

This project was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC), the Fonds de recherche du Québec – Nature et technologies (FRQNT), the Fondation de l’université du Québec à Chicoutimi (FUQAC), Rio Tinto Alcan (RTA), and the Programme d’acquisition de connaissances sur les eaux souterraines of Quebec (PACES), with contributions from the Quebec Ministère du Développement durable, de l’Environnement, de la Faune et des Parcs (MDDEFP), and the five county municipalities of the SLSJ region (Domaine-du-Roy, Du-Fjord-Du-Saguenay, Lac-Saint-Jean-Est; Maria-Chapdelaine and the City of Saguenay). The authors wish to acknowledge the anonymous reviewers for their meaningful input, Professor Philippe Page from the University du Quebec à Chicoutimi (UQAC) who gave us the benefit of his extensive experience in geochemistry, technician David Noël for his multiple skills in field hydrogeology, all the students involved in the project as field assistants and, finally, the research assistants M. Lambert, A. Moisan, M.L. Tremblay and D. Germaneau for their help with the PACES project. In addition, the authors wish to thank the well owners who graciously provided access to their groundwater intakes for sampling.

STUDY AREA

The Saguenay-Lac-Saint-Jean (SLSJ) region of Quebec, Canada covers 13,210 km2 and contains two important surface water features: Lake Saint-Jean (surface 1200 km2) and the Saguenay River (Figure 4.1). Regional topography is controlled by the Phanerozoic Saguenay Graben (180 Ma), which is approximately 30 km wide. The northern and southern walls of the Saguenay Graben are bounded by WNW fault systems [19] that mark the limits between the highlands (up to 1000 m asl) and the lowlands (0 to ca. 200 m asl) (Figure 4.1B). The regional physiography deeply affected by the Saguenay graben has controlled the emplacement and the formation of large accumulations of Quaternary deposits (sand, gravel, clay-silt) to a thickness of up to 180 m in the central lowlands [15], where the most populated areas are located. The highlands are characterized by rugged terrain dominated by thin glacial drift (till) deposits and outcropping areas.
A simplified representation of the basement geology of the SLSJ area is shown in Figure 4.1A. The basement is made up of plutonic rocks belonging to the Precambrian Canadian Shield [20]. Around Lake Saint-Jean and in the lowlands, there are several remnants of an Ordovician platform composed of a series of stratified sedimentary rocks, including siliciclastic strata, micritic limestones, and highly fossiliferous alternating limestones and shales [21]. The Ordovician sequence has a maximum thickness of 110 m [15].
The principal hydrogeological contexts encountered in the study area are presented in a conceptual cross-section in Figure 4.2A. The most recent glaciation left behind a discontinuous and heterogeneous layer of till, several terminal moraines and important esker deposits [22, 23] (fluvio-glacial sediments; Figure 4.2A). When the glaciers retreated approximately 10,000 years ago, the lowlands were invaded by the Laflamme Sea [24]. This marine incursion deposited a semi-continuous extensive layer of deep water sediments consisting of laminated clayey silt and grey silty clay [23, 25] (Deep water sediments; Figure 4.2A). Various facies of deltaic, littoral and pre-littoral sediments were deposited during the phase of isostatic rebound that forced the retreat of the Laflamme Sea [23, 26] (Deltaic and shore deposits; Figure 4.2A).
The geological history of the SLSJ region has created several hydrogeological environments that can be simplified into six principal stratigraphic assemblages (Figure 2B). The definition of the six hydrogeological environments was based on drilling data obtained at different locations over the study area (Figure 4.1B). Confined and/or unconfined aquifers occur in the Pleistocene deposits and combine locally to form multilayered aquifers [15] that can be either disconnected from or interconnected with Ordovician and/or Precambrian bedrock aquifers [27, 28]. Fluvio-glacial sediments are the most productive regional aquifers; they are frequently covered by the regional marine clay aquitard [13, 15]. Deltaic and shore deposits constitute the other main aquifer type [15].

METHODOLOGY

REGIONAL GROUNDWATER SAMPLING

Groundwater samples were collected from private wells [11]. The sampling protocol included the recording of temperature (T), redox potential (Eh), pH, dissolved oxygen (DO) and electrical conductivity (EC) while purging the wells. In situ parameters were measured using a Hanna Instruments HI 9828 multi-parameter probe. Purging was completed and samples were collected after the field parameters were stabilized. The sampled water was filtered in the field through a 0.45 μm membrane filter prior to the analysis of metallic ions. Nitric acid stabilized the samples for determining major cation and trace element concentrations. Zinc acetate was applied to determine sulphides. Samples were kept at 4 °C until they were sent to an accredited commercial laboratory for all chemical analyses including major, minor and trace levels of 38 inorganic constituents. Trace elements were analyzed by ICP-MS and total alkalinity was determined by titration (final pH 4.5). Ion exchange chromatography measured chloride, bromide, sulphate, sulphide, and nitrate levels whereas ammonium, fluoride and inorganic phosphorous were determined using a specific probe. The threshold for ionic balance for all the selected samples was set at less ±10% [29, 30]; samples exceeding this threshold were discarded from the chemical sample dataset. The original dataset contains 321 samples: 170 samples from bedrock aquifers and 151 samples from granular deposits (Figure 4.1A).

CLASSIFICATION OF SAMPLES

Walter et al. [15] proposed that the chemistry of groundwater found in bedrock is chemically different from groundwater found in granular aquifers. Generally speaking, bicarbonate groundwater evolves towards a chloride endmember. Considering this fact, the database was divided into four groups depending on the permutations of 2 water types (HCO3- or Cl-) combined with 2 aquifer types (bedrock or granular aquifer). The 2 groundwater types were identified using a Durov plot based on the relative concentrations of the major elements of chloride (Cl-) and bicarbonate (HCO3-) as expressed in milliequivalent per liter (meq/L). The Durov classification [31] provides a means of identifying a dominant cation, for instance, calcium (Ca2+) or sodium (Na+). Information on aquifers making it possible to divide them into 2 types were obtained from the owners of the wells. No distinction was made between limestone aquifers and crystalline bedrock aquifers, due to a lack of information.

PREPARATION FOR MULTIVARIATE ANALYSIS: DEFINING THE SUBSET OF DATA

Farnham et al. [32] performed data experiments to determine the best constant to use for substitution of analytical values below the detection limit (<DL) in statistical multivariate processing. The simulation experiments showed that substitution of <DL values with DL/2 was superior to substitution with 0 or DL. Results also showed that the performance for all substitution methods in statistical multivariate processing was weakened when the number of analytical values <DL exceeded approximately 30% of the dataset. Including these analytical values in the multivariate study would add substantial noise to the analysis [32]. In this study, chemical parameters with a number of analytical values <DL exceeding 25% were discarded.
In some cases, the chemical parameter to be discarded was a key parameter considered in this study (for instance, fluoride or iron). A key parameter corresponded to a chemical parameter for which concentration exceeded the limits for groundwater quality established by Canadian policy. To ensure that at least one of the key chemical parameters were included in the statistical multivariate processing, samples in which the key parameter was not detected (analytical value <DL) were withdrawn from the database. As a first step, sample selection was made by screening (using Excel software) analytical values > or <DL for each of the 38 chemical parameters. The screening of the data is completed when the number of analytical data >DL (i.e., valid data) for the key parameters reaches (or exceeds) 75% of the new set of data.
The screening step is iterative. It can be repeated with more than one screened chemical parameter. The choice of screened parameter is subjective and may vary depending on the group of groundwater (brackish or fresh groundwater; bedrock or the granular aquifers). In this study, correlation matrices constructed for each group of groundwater (supplementary material; Appendix 12) support the choice of the screened chemical parameters in each group of groundwater samples. The screening of the data reduced the total number of samples which could then be subjected to statistical multivariate processing. In this study, the number of screened samples varied between 10 and 15 samples per sub-group, for a total number of samples varying between 40 and 60 samples. The newly formed set of data is called the 51-sample subset.

HIERARCHICAL CLUSTER ANALYSIS AND FACTORIAL ANALYSIS

The Box-Cox power transformation was applied to the 51 samples of the subset to ensure a normal distribution [33] for the data. The variables (major, minor, and trace elements) were standardized by subtracting the mean concentration of a given element from each measured concentration and dividing by the standard deviation of the distribution [34]. The Statistica software version 6.1 [35] was used to perform multivariate statistical analyses. Euclidian distance was used as a distance measure of similarities and Ward’s method was applied as a linkage rule.
Factorial analysis (FA) was then applied to the subset. FA attempts to explain covariances [36] expressed as linear combinations of a small number of highly correlated variables (i.e., the chemical elements). The combinations of variables are called factors. Normalized Varimax rotation was applied to the factors [34, 36, 37]. The Kaiser criterion [34, 38] helps to determine the number of factors to be retained; finally, only those components with loadings greater than 1 were retained.

RAINWATER SAMPLING AND COMPILATION

Existing data for rainwater composition, primarily extracted from the National Atmospheric Chemistry Database (NatChem database) [39] were initially compiled. In addition, four samples were collected from two stations (two from each) where handmade rain collectors had been installed (Figure 1A). The samples were sent to a commercial lab for analysis, for the same chemical parameters as for groundwater. Rainwater chemistry used in this study is presented in Table 4.1.

STANDARDIZATION TECHNIQUE

The degree of enrichment and/or depletion of an element is calculated by dividing the concentration of an element found in a sample by the concentration of the same element found in a selected reference water sample. The gain or the loss of one or a combination of chemical elements during the process of evolution is interpreted to be a consequence of the interaction of the groundwater with its environment — in this case the “environment” is a combination of aquifer type and hydrogeological context. Another possible explanation for the increase of chemical content in groundwater during its evolution over time could be the mixing of several different groundwaters having different origins.

STANDARDIZATION TECHNIQUE

The degree of enrichment and/or depletion of an element is calculated by dividing the concentration of an element found in a sample by the concentration of the same element found in a selected reference water sample. The gain or the loss of one or a combination of chemical elements during the process of evolution is interpreted to be a consequence of the interaction of the groundwater with its environment — in this case the “environment” is a combination of aquifer type and hydrogeological context. Another possible explanation for the increase of chemical content in groundwater during its evolution over time could be the mixing of several different groundwaters having different origins.
where [X]Step y is the mean concentration of the chemical element X of the water that characterizes the step of evolution y, and [X]Step y+n is the mean concentration of the chemical element X of the water of reference that characterizes a more advanced stage of evolution (+ n). A value of one denotes no enrichment nor depletion during the process of evolution. A value of MI greater than 1 indicates that the chemical element X is gained in groundwater from the environment. A value of MI lower than 1 suggests the loss of the chemical element from groundwater (i.e., precipitation/adsorption) or by mixing with groundwater whose content in the chemical element has been depleted.

COMPILATION OF CANADIAN SHIELD BRINE DATA

Chemical data for brines found at depth in the Canadian Shield were compiled from the existing literature so as to investigate the geochemical origin of the end-members in this study. A total of 137 chemical analyses (i.e. samples) from five studies were included: Frape and Fritz [40], 33 samples; Bottomley et al. [41], 24 samples; Frape and Fritz [42], 36 samples; Frape et al. [43], 35 samples; and Gascoyne and Kamineni [44], 17 samples. The compiled data are presented as supplementary material (Appendix 2).

RESULTS

DATABASE VERSUS SUBSET

By combining the 2 groundwater types (HCO3- or Cl-) and the 2 aquifer types (bedrock or granular aquifer) considered in this study, the 321-sample dataset was divided into four groundwater groups: Group 1: bicarbonate groundwater from granular aquifers (Ca(Na)-HCO3_ GranAq; 132 samples); Group 2: chloride groundwater from granular aquifers (Na(Ca)-Cl_GranAq; 19 samples); Group 3: bicarbonate groundwater from bedrock aquifers ((NaCa)-HCO3_RockAq; 124 samples); Group 4: chloride groundwater from bedrock aquifers ((NaCa)-Cl_RockAq; 46 samples). For each group, values above the detection limit (N), the median, the first (25) and third (75) quartiles, the maximum (Max) and the minimum (Min) are presented in supplementary material (Appendix 3). Conventional statistics calculated for the entire database are also presented and described in Walter et al. [11].
Figure 4.3 presents the frequencies of detection of the chemical elements for each of the four groundwater groups. Along the X-axis, ionic species are sorted in descending order of abundance for Group 1 (Ca(Na)-HCO3_ GranAq). Depending on the type of groundwater (HCO3- or Cl-) and the type of aquifer (bedrock or granular aquifer), each chemical element presented a different number of valid data units. For instance, boron (B3+) and molybdenum (Mo6+) are commonly detected in groundwater from bedrock aquifers; lithium (Li+) and bromide (Br-) were frequently detected in brackish groundwater, particularly in bedrock aquifers; and copper (Cu2+), lead (Pb2+) and nitrate (NO3-) were more commonly detected in bicarbonate groundwater, particularly in granular aquifers. These observations suggest that the salinization path of bicarbonate groundwater toward a chloride endmember is characterized by several different chemical fingerprints inherited from the aquifer matrix (bedrock or granular aquifer).
Figure 4.3 also shows that only 10 chemical parameters (K+, HCO3-, Mg2+, SiO2, Na+, Ca2+, Ba2+, Sr2+, SO42-, and Mn2+) presented less than 25% of valid data and can be directly used in multivariate statistical process of the dataset (c.f. 3.3 Preparation for multivariate purposes: the subset). On the contrary, more than 25% of the bicarbonated water (Group 1 and 3) showed fluoride and iron contents to be <DL. Again for iron, more than 25% of the chloride groundwater from granular aquifers (Group 2) was <DL. To include fluoride and iron in the multivariate processing, a subset of samples was defined (c.f. 3.3 Preparation for multivariate purposes: the subset). The subset included samples that were screened in each of the four groups of groundwater: 16 samples were screened in group 1 (bicarbonate groundwater from granular aquifers – (CaNa)-HCO3_ GranAq); 9 samples were screened in group 2 (chloride groundwater from granular aquifers – (CaNa)-Cl_GranAq); 16 samples were screened in group 3 (bicarbonate groundwater from bedrock aquifers – (CaNa)-HCO3_RockAq); and 10 samples were screened in group 4 (chloride groundwater from bedrock aquifers – (CaNa)-Cl_RockAq).
By screening the samples, the ionic content of the 51-sample subset increased, although the m.c of some chemical elements remained unchanged or decreased (Figure 4.5). Variations of the median of the 2 datasets calculated in percentages are presented in supplementary material (Appendix 4). Chemical parameters with decreasing content were vanadium (88% the m.c of the 321-sample database), uranium (86% the m.c of the 321- sample database), phosphate (71% the m.c in the 321-sample database) and copper (63% the m.c of the 321-sample database). Selenium, beryllium, titanium and bismuth were all <DL in the 51-sample subset. For silver, nitrate and nickel, the m.c for the 51-sample subset was equal to the m.c for the 321-sample database. The most significant increase is for chloride (370% the m.c of the 321-sample database) and ammonium, (288% the m.c of the 321-sample database). The correlation matrices for the subset and the database (supplementary material; Appendix 5) showed that ammonium was highly correlated to major elements (K+, HCO3-, Mg2+, Na+, Ca2+, Cl- and SO42-), as well as some minor elements (Ba2+, Sr2+ and B3+) and some trace elements (Li2+, Br- and Pb2+).

HIERARCHICAL CLUSTER ANALYSIS

Hierarchical cluster analysis (HCA) is a semi-objective method [45], where the number of clusters of similar samples is determined by the position of the phenon line that cuts across the resulting dendrogram (Figure 4.6A). In our HCA, the distance linkage (Dlink) of the samples is expressed as a percentage of the maximum distance (Dmax) between the most dissimilar samples ([Dlink/Dmax]*100). At this scale, 100% linkage (Dlink = Dmax) would indicate that all samples are grouped into a single cluster that corresponds to the 51-sample subset. For this study, four clusters (Dlink <30% of Dmax) provided the most satisfactory result and fulfilled the objectives of our classification method (Figure 4.6A). Major cation and anion concentrations for the 4 clusters were projected on a log-scale spider diagram in ppm on Figure 6B. Chloride, sulfate and sodium dominated clusters 3 and 4. Calcium dominated cluster 3. Except for calcium, cluster 4 is enriched in major elements when compared to cluster 1, 2 and 3. Clusters 1 and 2 correspond to more diluted groundwater where bicarbonate ion dominated.
The HCA clusters showed some similarities with the classification of the 321 samples from the database (c.f. 3.2 Classification of samples), particularly for clusters 3 and 4. For each cluster, Table 4.4 shows the median concentrations for the physical and the analytical parameters and presents the water type, the aquifer type and the hydrogeological context. The HCA results for brackish groundwater clearly distinguish bedrock (cluster 3: 7/8 samples) from granular groundwater (cluster 4: 7/8 samples). The chemistry of these two latter clusters appears related to the nature of the porous matrix of aquifers with samples containing Na-Cl (cluster 4) mainly found in confined environments. Cluster 1 samples (Ca-HCO3) originated mainly from unconfined environments.
Cluster 2 samples (Ca(Na)-HCO3) originated equally from both confined and unconfined environments. Clusters 1 and 2 are largely composed of bicarbonate waters (32/35 samples) while clusters 3 and 4 are composed of chloride waters (16/16; Table 4). All cluster 1 samples are composed of Ca2+-dominant groundwater, whereas cluster 2 samples are primarily composed of Ca-HCO3 (11/20) and Na-HCO3 (6/20) groundwater with a limited number of Ca-Cl (1/20) and Na-Cl (2/20) groundwater samples. Cluster 4 is exclusively a Na-Cl type groundwater and cluster 3 is a mixture of Na+ (3/8) and Ca2+ (5/8) chloride groundwater.
Consequently, the key finding of the HCA is that the subset brackish groundwater samples collected from the bedrock aquifers (cluster 3) are chemically distinct from the subset brackish groundwater samples collected in the granular aquifers (cluster 4).

FACTOR ANALYSIS

Factors were calculated on the symmetrical correlation matrix computed for 16 variables (K+, HCO3-, Mg2+, SiO2, NH4+, F-, B3+, Mo6+, Na+, Ca2+, Ba2+, Sr2+, SO42-, Fe2+ and Mn2+). Chloride has been consciously discarded from the factorial analysis after observation of the correlation matrix (supplementary material; Appendix 5) as Cl- was highly correlated with all the other major elements and some trace elements and would induce a redundancy [53]. The correlation matrix for the 51-sample subset also revealed that aluminium and zinc had no significant correlation with other chemical elements. To prevent noise introduction in the multivariate statistical treatment, aluminium and zinc have been discarded too.
The first four factors present loadings (i.e. explained variance; Table 4.5) greater than 1 and account for 78.8% of the total variance of the subset. The first factor (25.1% of the total variance) is characterized by high loadings for K+, HCO3, Mg2+, SiO2, NH4+, and SO42-. Loadings for F-, B3+, Mo6+, and Na+ explain the second factor (22.2% of the total variance). Loadings for the third factor (20.6% of the total variance of the subset) are dominated by Ca2+, Ba2+, Sr2+, and SO42-. Finally, the fourth factor (10.8% of the total variance of the subset) is strongly influenced by Fe2+ and Mn2+.

DISCUSSION

EVOLUTION OF RECHARGE GROUNDWATER (CLUSTERS 1 AND 2; FACTOR 4)

Clusters 1 and 2 appear to belong to a “recharge” type groundwater. The fourth factor explains only 10.8% of the total variance of the data set and thus characterizes only a few samples among the 51-sample subset. In Figures 4.7(c), 4.7(e), and 4.7(f), sample scores for the fourth factor are projected onto the Y-axis. This factor presents high loadings for Fe2+ and Mn2+ (Table 4.5). These are redox sensitive species [46, 47] and thus, their concentrations in groundwater are primarily influenced by the amount of available dissolved oxygen [48]. In the presence of dissolved oxygen, principally near surface environments, Fe2+ and Mn2+ will be solubilized by oxidizing and acidic conditions that tend to increase the Fe2+ and Mn2+ content in groundwater. This may explain why Factor 4 predominantly influences (CaNa)-HCO3 groundwater (clusters 1 and 2; Figures 4.7(c), 4.7(e), and 4.7(f)).
Negative values for the fourth factor can be interpreted as a lack of correlation between the two chemical elements. Based on the Eh-pH diagram (Supplementary material; Appendix-6), the range of stability for dissolved Fe2+ differs from that of Mn2+. Depending on the Eh-pH conditions, Fe2+ and Mn2+ may not coexist in the soluble form. This might be at the origin of the negative values obtained for the fourth factor in Figures 4.7(c), (e), and (f). For instance, negative values are recorded in the recharge type of groundwater (predominantly cluster 2) that follows the freshwater influx vector interpreted at the bottom left quarter in Figures 4.7(c), (e), and (f).

FINGERPRINT OF THE LAFLAMME SEAWATER (CLUSTER 4; FACTOR 1)

Positive scores for the first factor are recorded for samples from cluster 4 (Figures 4.7(a) – 4.7(c)) highlighting the influence of the first factor on cluster 4 samples. The first factor is characterized by high loadings of K+, HCO3-, SiO2, NH4+, SO42-, and Mg2+ (Table 4.5).

FINGERPRINTING OF THE PRECAMBRIAN SHIELD BRINES (CLUSTER 3; FACTOR 3)

In Figures 4.7(b), 4.7(d), and 4.7(f), samples from cluster 3 are predominantly influenced by the third factor with loadings dominated by Ca2+, Ba2+, Sr2+, and SO42- (Table 4.5). Log-log plots of [Ca2+, Ba2+, Sr2+, and SO42-] vs. Cl- concentrations (Figures 4.8(a)–8(d)) for the subset of samples in this study were coupled with the data from other Canadian Shield brines.
Frape et al. [43] determined that the geochemistry of Ca2+ plays a major role in controlling the chemistry of mineralized groundwater in the crystalline bedrock of the Canadian Shield. Frape et al. [43] described very old stagnant groundwater that may have undergone prolonged chemical alteration since its original emplacement. The similar Ca2+/Cl- trend for the cluster 3 and the Precambrian Shield brines (PSB) (Figure 4.8(a)) suggest a common origin for Ca2+ and Cl-. In Figures 4.8(a) and 4.8(d), Ca2+ and Sr2+ concentrations for clusters 3 and 4 show similar trends. This suggests a common origin for Sr2+ and Ca2+ in cluster 3. For Sr2+, as well as Ba2+, it is commonly found incorporated as a trace element in the crystalline structure of feldspar plagioclase minerals, [53, 54]. In Figure 4.8(b), the trend for Ba2+ is different than that of Sr2+ and Ca2+. Edmunds and Smedley [61] showed that barite solubility controls Ba2+ concentration and that Ba2+ quickly reaches saturation or supersaturation in recharge groundwater. In Figure 4.8(b), Ba2+ concentrations for clusters 1 and 2 reach levels similar to the concentrations for clusters 3 and 4. Which suggest that maximal Ba2+ concentrations are attained in “recharge” groundwater. Unlike Ba2+, Sr2+ is not limited in terms of solubility [55]. Some of the Sr2+ may be added by the dissolution of anhydrite or gypsum. However, SO42- vs Cl- (Figure 4.8(c)) shows a different trend than Sr2+ vs Cl- (Figure 8(d)) that refutes this possibility. According to Gascoyne and Kamineni [44], most of the added Cl- and some SO42- in groundwater collected in selected granitic, gabbroic, and gneissic plutons in the Canadian Shield are derived from the rock matrix. Ions are present either as soluble salts in fluid inclusions and on grain surfaces [56] or as structurally bound elements in micas and amphiboles [57]. Sulphate may then be more derived from the oxidation of sulphide minerals in the rock matrix rather than due to the presence of gypsum-infilling minerals in the fractures.

Table des matières

CHAPITRE 1 – INTRODUCTION
1.1 PROBLÉMATIQUE GÉNÉRALE
1.1.1 Variabilité chimique de l’eau souterraine
1.1.2 Modèles d’évolution de la chimie de l’eau souterraine
1.2 PROBLÉMATIQUE SPÉCIFIQUE
1.2.1 Contexte hydrogéologique régional et variabilité chimique de l’eau souterraine au Saguenay Lac-Saint-Jean
1.2.2 Hypothèses de recherche
1.3 OBJECTIFS
1.4 METHODOLOGIE
1.4.1 Base de données hydrogéochimiques existantes et statistiques univariées
1.4.2 Analyses statistiques multi-variées et Pôles compositionnels régionaux
1.4.3 Représentations graphiques
1.4.4 Méthode normative et modèle d’évolution
1.5 FORMAT DE LA THÈSE
1.6 RÉFÉRENCES
CHAPITRE 2 – CHARACTERIZATION OF GENERAL AND SINGULAR FEATURES OF MAJOR  SYSTEMS IN THE SAGUENAY-LAC-SAINT-JEAN AREA
2.1 MISE EN CONTEXTE
2.2 ABSTRACT
2.3 INTRODUCTION
2.4 METHODOLOGY
2.5 THE SLSJ HYDROGEOLOGICAL FRAMEWORK
2.6 BEDROCK AQUIFERS
2.6.1 Precambrian rocks
2.6.2 Ordovician sedimentary rock units
2.6.3 Faults and fractures network within the bedrock
2.6.4 Mapping of bedrock topography
2.7 GRANULAR AQUIFERS
2.7.1 Surface topography and Quaternary granular deposits
2.7.2 Quaternary aquifers and municipal groundwater supply
2.8 INTERCONNECTION BETWEEN FRACTURED BEDROCK AND GRANULAR AQUIFERS
2.9 GROUNDWATER GEOCHEMISTRY
2.10 DISCUSSION
2.10.1 Database updating
2.10.2 Groundwater quality
2.10.3 Groundwater protection
2.10.4 Groundwater flow in a graben environment
2.10.5 Expertise required of groundwater database users
2.11 CONCLUSION
2.12 ACKNOWLEDGEMENTS
2.13 REFERENCES
CHAPITRE 3 – THE INFLUENCE OF WATER/ROCK – WATER/CLAY INTERACTIONS AND MIXING IN THE SALINIZATION PROCESSES OF GROUNDWATER 
3.1 MISE EN CONTEXTE
3.2 ABSTRACT
3.3 INTRODUCTION
3.4 GEOLOGY OF THE STUDY AREA
3.4.1 FRACTURED ROCKS
3.4.2 BEDROCK TOPOGRAPHY
3.4.3 PLEISTOCENE DEPOSITS
3.5 HYDROGEOLOGICAL BACKGROUND
3.5.1 HYDROGEOCHEMICAL BACKGROUND
3.6 METHODOLOGY
3.6.1 Sampling sites and groundwater sampling location
3.6.2 Data processing
3.6.2.1 STEP 1: SELECTION OF THE CHEMICAL ELEMENTS USED IN MULTIVARIATE ANALYSIS
3.6.2.2 STEP 2: MULTIVARIATE STATISTICAL ANALYSIS
3.6.2.3 STEP 3: GRAPHICAL REPRESENTATIONS OF THE INVESTIGATIONS
3.7 RESULTS
3.7.1 Hierarchical cluster analysis
3.7.2 Principal component analysis
3.7.3 Geographical distribution of clusters
3.8 DISCUSSION
3.8.1 Recharge groundwater and water/rock interactions – Ca2+ vs HCO3-
3.8.2 Salinization processes
3.8.3 Ca2+WATER-Na+MINERAL ion exchange
3.8.4 Microcline weathering
3.8.5 General geochemical evolution path in the SLSJ aquifer systems
3.9 SUMMARY AND CONCLUSION
3.10 ACKNOWLEDGEMENTS
3.11 REFERENCES
CHAPITRE 4 – CHEMICAL PATHFINDERS FOR THE NATURAL EVOLUTION OF GROUNDWATER TOWARD BRACKISH END-MEMBERS IN PRECAMBRIAN BEDROCK AQUIFERS AND PLEISTOCENE GRANULAR AQUIFERS 
4.1 MISE EN CONTEXTE
4.2 ABSTRACT
4.3 INTRODUCTION
4.4 STUDY AREA
4.5 METHODOLOGY
4.5.1 Regional Groundwater Sampling
4.5.2 Classification of Samples
4.5.3 Preparation for multivariate analysis: defining the subset of data
4.5.4 Hierarchical Cluster Analysis and Factorial Analysis
4.5.5 Rainwater Sampling and Compilation
4.5.6 Standardization technique
4.5.7 Compilation of Canadian Shield Brine Data
4.6 RESULTS
4.6.1 Database versus subset
4.6.2 Hierarchical Cluster Analysis
4.6.3 Factor Analysis
4.7 DISCUSSION
4.7.1 Evolution of Recharge Groundwater (Clusters 1 and 2; Factor 4)
4.7.2 Fingerprint of the Laflamme Seawater (Cluster 4; Factor 1)
4.7.3 Fingerprinting of the Precambrian Shield Brines (Cluster 3; Factor 3)
4.7.4 The Source of Dissolved Fluoride in Groundwater (Factor 2)
4.8 CONCEPTUAL MODEL OF THE GROUNDWATER CHEMICAL EVOLUTION IN THE STUDY AREA
4.8.1 The gravity-driven flow model
4.8.2 Groundwater Chemical Facies and Chemical Pathfinders
4.8.2.1 Step 1 and 2: Rainwater Infiltration and Groundwater Evolution Towards Cluster 2 Chemistry
4.8.2.2 Step 3: Groundwater Maturation Towards Cluster 3 and 4 Brackish End-Members
4.9 CONCLUSION
4.10 ACKNOWLEDGEMENTS
4.11 References
CHAPITRE 5 – DISCUSSION ET RECOMMANDATIONS
5.1 Connaissances hydrogéologiques régionales
5.1.1 Une physiographie particulière : le Graben du Saguenay
5.1.2 Le roc fracturé précambrien et ordovicien
5.1.3 Les dépôts meubles quaternaires
5.1.4 Cadre hydrostratigraphique régional
5.1.5 Contribution au cycle des connaissances hydrogéologiques régionales
5.2 Pôles de l’évolution géochimique naturelle de l’eau souterraine
5.2.1 Qualité des données hydrogéochimiques et limites de détection
5.2.2 Regroupement des échantillons et réduction du jeu de données
5.3 Méthode normative et empreinte géochimique des pôles régionaux
5.4 Portée des résultats
5.4.1 Variabilité spatiale et temporelle
5.4.2 Nouveaux champs d’interprétation graphique
5.4.3 Contextes naturels favorables au fluor
5.5 Travaux de recherche complémentaires
5.5.1 Études minéralogiques
5.5.2 Modélisation hydrogéochimique
5.5.3 Montages expérimentaux
5.5.4 Analyses isotopiques
5.5.5 Projet d’envergure
5.6 Références
CHAPITRE 6 – CONCLUSION 
ANNEXES