Historical variation of elements with respect to different geochemical fractions in recent sediments from Pigeon Lake, Alberta, Canada

Abstract

Geochemical analysis of elements and organic matter were conducted on vertical profiles of the recent sediments from Pigeon Lake, Alberta, Canada, to determine historical variations in elemental content of the sediments as related to their geochemical fractions. The elements are grouped according to their affinity with different geochemical fractions, by using cluster analysis and sequential extraction experiments. As a result, four elemental fractions were identified: clastic mineral detritus; carbonate; organic; and elements that show less similarity to the previous groups perhaps due to anthropogenic input or the influence of other fractions, such as oxyhydroxides. Following the identification of geochemical fractions in the sediments, a three-step normalizing method was applied using parameters that represent each geochemical fraction. These normalizing techniques appear to be important in verifying whether the variation of elements is indeed the result of anthropogenic and/or natural activities. The normalized data are correlated with the recent history of human activity and natural events near Pigeon Lake. Given the age of the lake sediments, this correlation indicates that the depth profiles of elements after being normalized to the organic and carbonate fractions reflect the variation of detrital input into the lake. However, the former mainly corresponds to the coarse-grained clastic minerals originating from high-energy erosion as the result of natural events (e.g., flooding), whereas the latter corresponds to the low-energy erosion of the fine particles (enriched in lithophile elements) due to deforestation in the drainage basin. Normalizing to the clastic mineral detritus fraction results in the increase of heavy metals in the uppermost part of the sediment profiles, which coincides with industrial activities during the past two decades in central Alberta. However, the concentration of these elements is negligible, as compared to the quantities released by geogenic processes (erosion).

---

Introduction

The geochemistry of marine and lacustrine sediments has been used widely as a historic record of anthropogenically induced changes in elemental fluxes into the environment.^1–5 It is known that both natural phenomena and anthropogenic activities contribute to the observed accumulation of elements in the surficial sediments.⁶ Lottermoser et al.⁷ found that the increase in elemental concentrations in sediments from Holzmaar, Germany, is, in fact, due to the increase in erosion and transport of clastic materials into the lake. A number of other studies have reached the same conclusion (e.g. Roulet et al.;⁸ Burden et al.⁹) suggesting that assessment of anthropogenic elements and pollutants in sediments, solely based on their bulk concentrations, can be misleading.¹⁰

Hakanson and Jansson¹¹ stated that differences in sediment parameters such as organic content and mineral matter may markedly effect the concentration of elements in sediments. This is because elements are not primarily in solution but attached to various “carrier particles” (i.e., suspended organic and inorganic particles/aggregates of different origin and chemical character).^11,12 Therefore, elements with different chemical properties may appear with similar distribution patterns in the sediments due to the fact that they are linked to the same carrier particle with similar sedimentologiocal properties.^9,11

The type of carrier particle and the way in which elements are bound to it, determine the element's geochemical fraction in sediments.^11,13,14 Identification of these geochemical fractions is imperative in proper assessment of elements and pollutants in sediments, as each fraction has a suite of naturally occurring elements associated with its carrier particle.

The objectives of this study are: first, to determine the geochemical fractions, and their associated elements, present in the lake core sediments retrieved from Pigeon Lake, Alberta, Canada; and secondly, to determine the vertical distribution of elements as related to the identified geochemical fractions, possible source for these elements, and the significance of anthropogenic inputs to the lake sediments.

Study area

Pigeon Lake is a shallow (maximum depth of 9 m and mean depth of 6.2 m), fresh water lake located in south-central Alberta, Canada (53° 01′ N latitude, 114° 02′ W longitude; Fig. 1). Pigeon Lake has a small watershed (187 km²), which is only twice the size of the lake (96.7 km²). Surficial deposits in the drainage basin are predominantly calcium rich glacial till that originated from the Paskapoo bedrock formation underlying the area. Paskapoo bedrock (Tertiary) is primarily sedimentary and consists of layers of sandstone, siltstone, mudstone, thin limestone, coal, and tuff beds.¹⁵ Several intermittent streams flow directly into the lake¹⁵ (Fig. 1). A small volume of the lake is discharged through Pigeon Lake Creek, located at the south end of the lake.¹⁶ Such an environment causes a high water residence time, and subsequently, a high sedimentation rate.

Fig. 1 Map of the study region, sampling sites, present land use and other hydrological features of Pigeon Lake, Alberta, Canada.

The arrival of settlers to the Pigeon Lake area in 1827 has been followed by extensive deforestation and agricultural activities within the lake's catchment¹⁵ (Fig. 1). Anthropogenic activities that may contribute to the influx of material into lakes include agricultural and residential activities in the lake's surrounding area (e.g., increasing soil erosion), and atmospheric fallout from traffic and industrial activities in vicinity of the lake's catchment.¹⁷

Methodology

Coring and sample preparation

Vertical profiles of the top 150 cm of the sediment were retrieved using a 7.5 cm diameter percussion corer. Core A was obtained from a 9 m water depth and at a distance of 5 km from the shore (off-shore) (Fig. 1). Core B was obtained from the littoral zone (near shore) at a water depth of 4.5 m (Fig. 1). The cores were cut length-wise to study the visible structures and sedimentary features of the sediments. The cores were then sub-sampled at 1 cm intervals to a depth of 4 cm, and at 2–3 cm intervals (based on changes in sediment type and/or sedimentary structures) for the remainder of the cores. Samples were freeze-dried, ground, and mixed before analytical procedures were conducted.

Instrumentation

Geochemical analysis for a suite of elements (Table 1) were conducted using inductively coupled plasma mass spectrometry (ICP-MS) after a hot digestion with nitric, perchloric and hydrofluoric acids. Instrumental neutron activiation analysis (INAA) was conducted for a number of elements (Table 1) which could not be measured by ICP-MS method due to the hot digestion procedure used and the resultant loss of some volatile elements in the sample. The sediment samples were irradiated at 2 MW with a flux of 8 × 10¹² neutrons cm⁻² s⁻¹ in the core of the McMaster University (Hamilton, Ontario, Canada) nuclear reactor. After a waiting period, the sample was placed close to a high resolution instrinsic germanium coaxial detector coupled through a Canberra model 2024 fast spectroscopy amplifier to a model 8715 ADC, nuclear data ND599 loss free counting module, and an Aptec 8 K channel multichannel analyzer. Interactions of the gamma-rays (which continue to radiate from the sample), with the detector, led to discrete voltage pulses proportional in height to the incident gamma-ray energies. The multichannel analyzer sorted out the voltage pulses from the detector and digitally constructed a spectrum of gamma-ray energies versus intensities. By comparing spectral peak positions and areas with library standards, the elements could be qualitatively and quantitatively identified. For determination of boron content, the samples were transferred to the prompt gamma activation site at McMaster University, where the gamma-ray spectra were obtained. The spectra were reviewed and analyzed by Becquerel Laboratories, Mississauga, Ontario, Canada. Analysis of duplicate samples and laboratory standards were used to monitor analytical accuracy and precision of the all analytical methods. The reader is referred to Stoeppler¹⁸ and Sloss and Gardner¹⁹ for further details on sample preparation, instrumentation used in the analytical methods, and detection limits of the applied methods.

Table 1 Analytical methods used for determination of elemental concentration in this study

Analytical method
INAA^a		ICP-MS^b		PG^c
Element	LLD^d	Element	LLD^d	Element	LLD^d
a INAA: instrumental neutron activation analysis. b ICP-MS: inductively coupled plasma mass spectrometry. c PG: prompt gamma. d LLD: lower limits of detection (in ppm unless listed as %).
Al	0.01%	Ba	10	B	5
As	0.1	Be	0.05
Br	0.1	Bi	0.01
Ca	0.10%	Cd	0.02
Ce	0.1	Cr	1
Cs	0.05	Co	0.2
Dy	0.1	Cu	1
Eu	0.2	Ga	0.1
Fe	0.10%	Li	0.2
K	0.10%	Mn	5
La	0.5	Mo	0.2
Lu	0.05	Nb	0.2
Mg	0.05	Ni	0.2
Nd	0.5	P	10
Rb	5	Pb	0.5
Sb	0.05	Sr	0.2
Sc	0.1	Tl	0.02
Sm	0.05	V	1
Th	0.2	Y	0.1
Ti	0.05%
U	0.1
Yb	0.1
Zn	2

---

The geochemical parameters for organic matter were determined using the new version, high resolution, Rock-Eval 6, which is a programmed heating, two-step pyrolysis and oxidation instrument manufactured by Vinci Technologies, Rueil-Malmasion, France. The parameters obtained from this method are the TOC (percentage total organic carbon), HI (hydrogen index, corresponding to the atomic H/C ratio), S2 (the amount of hydrocarbon derived from thermal cracking of organic matter), and MI (percentage total mineral carbon). For details of this method and more description about the above parameters refer to Lafargue et al.²⁰

Sequential extraction

Sequential extractions provide valuable information about the speciation of elements in the sediments by adding the chemical reagents that extract the elements selectively from certain phases.^7,12 The known chemistry of the system permits the phase to be determined.

A wide variety of sequential extraction methods have been employed by researchers (e.g., Tessier et al.,¹³ and Solomons and Forstner¹⁴). In this study, the uppermost 50 cm of sediment of each core was mixed and homogenized. The two resulting samples (5 g) were then subjected to a seven-step sequential extraction experiment, using the following sequence of chemical treatments: (1) 25 ml of deionized water; (2) 25 ml of 1 M MgCl₂·6H₂O (pH 7.0); (3) 50 ml of 1 M NaAc (pH = 5) at 90 °C; (4) 25 ml of 0.1 M NH₂OH·HCl at 25% v/v CH₃COOH at 90 °C; (5) 25 ml of H₂O₂ (pH = 2) at 100 °C; (6) 20 ml of aqua regia (3HCl ∶ 1HNO₃) at 120 °C; and (7) hot digestion by HF–HCl–HNO₃ to dryness. This sequential extraction procedure is able to differentiate between: (1) water-soluble; (2) exchangeable; (3) carbonate; (4) reducible Fe/Mn; (5) oxidizable organic phases; (6) sulfides; and (7) residual elemental fractions, respectively. However, the reagents used are not necessarily as selective as implied by the above schemes, and extra caution should be taken in interpretation of such data.^21–23 The elemental content of the extractants was then determined using ICP-MS.

Cluster analysis

In order to clarify the elemental relationships, the pair linear correlation coefficient was calculated to construct a hierarchic dendogram (cluster analysis), using the method described by Labonte and Goodarzi.²⁴ By identifying relationships among elements and other geochemical indicators in cores, groups of elements were determined. These groups of elements were compared with the results from other comprehensive studies of lake sediment chemistry to illustrate elements linked to natural and anthropogenic activities.

Results and discussion

Sediment descriptions

Sediments in core A are described as fine-grained, organic rich (whole core average organic carbon of 7.9%) with no visible laminae. Core B contains coarse-grained clastic sediments with low organic content (whole core average organic carbon of 3.0%) and some visible sedimentary structures.²⁵

Cluster analysis

Geochemical parameters of organic matter such as total organic carbon (TOC), hydrogen index (HI), and S2 (amount of hydrocarbon), as well as the elemental concentration in sediments for both cores, were subject to cluster analysis. The elements are grouped according to their affinity with different geochemical indicators (Fig. 2a and 2b). Therefore, each group of elements can be linked to a particular sediment type having the same input source, relating to lacustrine conditions either at the time of sedimentation or during an ensuing diagenetic change.⁹

Fig. 2 Cluster analysis on the geochemical data of the sediment samples in core A (a) and core B (b) indicating various geochemical fractions. G I: elements with organic affinity; G IIa: elements associted with carbonates; G IIb: lithophile elements with affinity to the clastic, detrital materials; G IIc: elements with least similarities, possibly from anthropogenic or unknown sources.

As the result, two main groups of elements, of which the second group is further divided into three subgroups, are identified by cluster analysis in cores A and B (Fig. 2a and 2b) and are described below.

Group G I

The first group of elements is called G I, which are the elements with an organic affinity, and show a high correlation with organic indicators such as HI, S2, and TOC (Fig. 2a and 2b). The elements P, B and Br fall into this group. The occurrence of Mn and Pb in this group is likely by coincidence, since these elements mainly concentrate in the uppermost sediments where younger sediments with higher organic matter content exist.

Group G II

The second group, G II, contains the majority of elements which are believed to be allogenic (Fig. 2a and 2b). This group is classified into three subgroups, G IIa, G IIb, and G IIc, based on their origin and their similarity in the dendograph. The G IIa elements show an affinity to carbonate (lime mud association), whereas the elements in G IIb are presumably related to the clastic mineral detritus. These two subgroups are located at the two ends of G II. Between these two subgroups, there is a set of elements, which show less similarity. This set of elements is classified into the third subgroup called G IIc. The overall low similarity of this subgroup as compared to the other groups of elements suggests that these elements are influenced by other fractions such as oxyhydroxide, and anthropogenic input. These subgroups include the following elements.

Subgroup G IIa. Calcium and strontium are classified into subgroup G IIa. These elements show a high similarity in both cores and are attributed to carbonates. Mineral carbon (MIC) obtained from Rock-Eval analysis is also associated with Ca and Sr in the dendograph depicted for both core A and B. This confirms an affinity for this group of elements to carbonate (lime mud). The higher concentration of Mg and Ba along with Ca and Sr is also due to the presence of carbonate minerals.²⁶ In this study, Mg and Ba were not found in the same group possibly due to the effects of the mineral matrix in the analytical process.

Subgroup G IIb. This group consists of lithophile elements such as rare earth elements (REEs), Al, Ti, Fe, Sc, Zn, Co, Th, Ga, Nb, Y, Rb, and Cs, which are the main elements contributing to the chemical composition of clays and silts.²⁷ Various authors have illustrated a link between these elements and clastic input to the lake sediments.^27,28 The group G IIb contains a greater number of elements in the dendograph obtained from core B (Fig. 2b), which is located close to the shore and hence contains more clastic materials.

Subgroup G IIc. The elements of environmental concern such as Cu and Cr are found in G IIc in both core A and core B. Core A, due to its fine-grained sediments, adsorbs greater amounts of elements related to anthropogenic activities.²⁹ As a result, the dendograph obtained from core A includes a greater number of these elements in this subgroup (Fig. 2a). These elements (As, Cd, Mo and Ni) are believed to partially originate from anthropogenic sources.

Sequential extraction

The results of sequential extraction for each element are presented as the mass of the extracted species divided by the mass of the entire starting material before sequential extraction (e.g., Fe in residual fraction per g). Such a presentation illustrates the behavior of one element in the same fraction of all samples or in the other words, indicates the relative proportions of separated element in each fraction. The results indicate three major groups of elements as follows.

Carbonate fraction. A significant amount (50–60%) of the total Ca and Sr is found to be soluble in NaAc, indicating a strong affinity with carbonate fraction (Fig. 3a). This is in good agreement with the results of the cluster analysis (equivalent to the subgroup G IIa).

Fig. 3 Relative proportions of different elements in each of the sequential extraction fractions for the sediments indicating the elements associated with carbonate fraction (a), clastic mineral detritus fraction (b), and organic fraction (c). 1, water soluble (deionized H₂O); 2, exchangeable (MgCl₂); 3, carbonate fraction (NaAc); 4, reducible Fe/Mn (NH₂OH–HCl); 5, organic matter fraction (H₂O₂); 6, sulfide/organic matter fraction (aqua regia); 7, residual elemental fractions (HF–HCl–HNO₃).

Clastic mineral detritus fraction. The results of sequential extraction for this group are shown in Fig. 3b. These elements mainly belong to the group of lithophile elements and correspond to group G IIb in the cluster analysis. Approximately ½ to ⅔ of the total amount of these elements are soluble in strong acids (last two steps). The elements leached in the last two steps are highly immobile and geogenic, indicating the detrital origin of the elements.¹⁸

Organic fraction. This group consists mainly of elements of environmental concern³⁰ with partial organic affinities (Fig. 3c). Significant amounts of these elements are leachable with H₂O₂ indicating an organic fraction. The remaining fraction of the leachable elements is released by aqua regia, which to some extent also represents organic residues. The organic fraction of these elements could possibly have originated from anthropogenic sources in the recent sediments.³¹ The organic association of some elements with expected lower organic affinities, such as Co and Ni, could be due to geogenic input of these elements from coal seams underlying the surficial deposits in Pigeon Lake.

The current sequential leaching experiment was not conducted under inert conditions. Therefore, possible mobility and shifts in elemental fractions could occur due to oxidation processes.²³ The sequential extraction data presented in this paper is only used to define the major geochemical fractions in the sediments, and to confirm the data obtained from the cluster analysis. The detailed discussion regarding the sequential extraction data is the subject of a separate study and is not within the scope of this paper.

Normalizing methods

The vertical variation of elements based on bulk elemental analysis (Fig. 4) shows an erratic pattern throughout the sediment profile. Comparison between each pair of these elements as expressed by correlation coefficient matrix indicates a relatively low similarity of r_mean = 0.86 (Fig. 4). This is due to the fact that the bulk chemistry of sediments is influenced by various geochemical fractions.^8,10 In order to eliminate the effects of these fractions, the bulk concentration of elements can be corrected with respect to the geochemical fractions present in sediments. Following the identification of our elemental fractions in the sediments, a three-step normalizing procedure was applied as follows.

Fig. 4 Vertical variation of lithophile elements in core A based on their bulk concentration (before normalization) and the mean correlation coefficient between each pair of elements.

Step 1: Normalization to organic matter (TOC and S2)

The method of normalizing to organic content of sediments has been widely used to correct for the strong affinity between some elements and organic matter in the sediments (e.g., Burden et al.;⁹ Hakanson and Jansson¹¹).

A schematic model is proposed (Fig. 5a), which tentatively classifies the sediments into four fractions (organic, clastic mineral detritus, carbonate, and others) based on our previous disscussions. Theoretically, each fraction can be eliminated by normalizing the bulk data to the geochemical parameters, which best represent, that particular fraction. This is analogous to “sieving elements” based on the geochemical fractions present in sediments.

Fig. 5 (a) The schematic model for Step 1 normalizing method, indicating the main elemental fractions and their geochemical indicators before and after the normalizing procedure. The bulk concentrations of elements are normalized to the organic fraction using the organic parameters S2 and TOC. (b) Vertical variation of lithophile elements normalized to the organic fraction (the mean of the sum of S2 and TOC; Step 1) and their correlation with Rock-Eval parameters for the sediments obtained from core A. The Rock-Eval parameters are reported in mg HC g⁻¹ rock for S2 and weight (%) for TOC. MH = marker horizons, indicating the historical flooding and subsequently high input of clastic materials from the watershed into the lake; “Primary productivity cycle” corresponds to the period of high algal growth in the lake.

In this study, the bulk concentration of elements are normalized to the organic fraction using parameters such as TOC and S2 (the mean of the sum of TOC and S2 for each sample). According to the model, the resulting profiles after this normalizing step likely reflect the variation of elements belonging mainly to the clastic mineral detritus fraction, with some influence of the carbonate fraction (Fig. 5a). This model is confirmed by Sanei et al.,²⁵ who found similar vertical variations in REEs (Ce, Yb, Sm, La, Lu, Nd, Eu, and Dy), and Th, Sc, Al, Fe, Zn, K, Ti, and Rb (r_mean = 0.97) after being normalized to the mean of the sum of TOC and S2, indicating the same source of input for these elements. They conclude that the resulting profiles likely correspond to the rate of high-energy (flood-related) erosion, which transports coarse-grained, calcium-rich, clastic particles (clastic mineral detritus fraction + carbonate fraction; Fig. 5a) into Pigeon Lake (Fig. 5b).²⁵

The profile shapes of these elements are inversely proportional to TOC and S2 from Rock-Eval analysis (Fig. 5b), likely reflecting dilution by different sources. For instance, as the lake progresses through a high algal productivity period, detrital elements (clastic mineral detritus fraction + carbonate fraction; Fig. 5a) comprise a smaller portion of the total input of elements in the sediments than the elements associated with organic matter (Fig. 5b). On the other hand, the high input of clastic material to the lake during a flood event can attenuate the organic content of the sediments.

Step 2: Normalization to carbonate fraction (Ca and Sr)

The bulk concentrations of elements are normalized to the mean of the sum of Ca and Sr (elements, representing the carbonate fraction) in each sample, in order to eliminate the effect of carbonate fraction. The schematic model, shown in Fig. 6a predicts the group of elements remaining after this step. It is expected that the resulting profiles from this step of normalizing will correspond to the variation of elements belonging mainly to the clastic mineral detritus fraction and, to a lesser extent, the organic and other fractions (Fig. 6a).

Fig. 6 (a) The schematic model for Step 2 normalizing method, indicating the major elemental fractions and their geochemical indicators before and after the normalizing procedure. The carbonate fraction is eliminated as a result of normalization of the bulk data to the carbonate fraction (the mean of the sum of Ca and Sr). (b) Vertical variations of lithophile elements normalized to the carbonate fraction (the mean of the sum of Ca and Sr; Step 2) for the sediments obtained from core A and the mean correlation coefficient calculated for each pair of elements. “Deforestation cycle” corresponds to the period of high lithophile element input due to the anthropogenic increase of erosion by deforestation and agricultural activity in the drainage basin.

The depth profiles of lithophile elements were depicted after this normalizing step (Fig. 6b). The resulting profiles indicate a remarkable improvement in their correlation coefficients (r_mean = 0.92) as compared to the profiles before normalizing (r_mean = 0.86) (Figs. 6b and 4). The pattern of variation in this series of elements is entirely different from those obtained from Step 1 (normalization to organic matter). The most remarkable difference is that the marker horizons (MH 1 and MH 2) observed in Step 1 (Fig. 5b) are eliminated in this step (Fig. 6b). This is likely because these two horizons correspond to the input of calcium-rich materials originating from the erosion of glacial tills covering the drainage basin. Therefore, by correction of bulk concentration data for carbonate fraction, these marker horizons are eliminated. This is in good agreement with the schematic model described in the above paragraph.

Since these two marker horizons represent significant flooding in the lake's area,²⁵ an increase of elements in the marker horizons (MH 1 and MH 2) can be related to a rapid, high-energy runoff event transferring the detrital calcium-rich material into the lake system. Such detrital materials have a greater particle size, as compared to the detrital particles originating from gradual weathering and erosion of rock in the drainage basin. On the other hand, the fine-grained particles contain higher amounts of REEs as compared to coarse-grained sand.²⁷ Therefore, the increases of lithophile elements in the vertical profiles could be due to the high elemental input resulting from an increase in the low-energy erosion of the drainage basin, which transports small particles enriched in these elements to the lake.

The above results indicate that although the elemental profiles after Step 1 and 2 normalization, both represent the variation in detrital input to the lake; the former mainly corresponds to the coarse-grained clastic minerals originating from high-energy erosion (flooding), whereas the latter corresponds to the fine particles originating from the low-energy erosion of the drainage basin.

Step 3: Normalization to clastic mineral detritus fraction (rare earth elements)

The Fe and Al concentrations are frequently used for the normalization of naturally occurring elements to their anthropogenic counterparts.^32,33 The use of Al and Fe normalization can correct the data for the quantity of aluminosilicates, and compensate for the effects of grain size and sedimentation on the concentration of metals in the sediments.⁸ However, the use of the total concentration of Fe and Al for normalization of the sediments does not yield a satisfactory estimation of the quantity of aluminosilicates, since the oxyhydroxide fraction of Fe and Al often interferes with the siliceous fraction.⁸

In this study, a number of REEs are used for normalizing since these elements also represent the quantity of aluminosilicates, the natural adsorption matrix for elements originating from the erosion of soils (clastic mineral detritus fraction).^22,27,28 The advantage of using REEs over Fe and Al is that they are less influenced by other fractions such as the oxyhydroxides. Therefore, with normalization of the bulk concentration data to REEs, the group of clastic mineral detritus is eliminated. This assumption has been depicted in the schematic model shown in Fig. 7a. According to this model, the expected elemental fractions remaining after this step of normalizing should be mainly anthropogenic, with minor portions of organic, carbonates, redox, and unknown fractions (Fig. 7a).

Fig. 7 (a) The schematic model of Step 3 of the normalizing method, indicating the major elemental fractions and their geochemical indicators before and after the normalizing procedure. Clastic mineral detritus fraction is eliminated as a result of the normalization of bulk data to mineral indicators (rare earth elements). (b) Vertical variation of “elements of environmental concern” in core A based on their bulk concentration (before normalization, values are reported in (ppm). (c) Vertical variation of “elements of environmental concern” normalized to the clastic mineral detritus fraction (rare earth elements; Step 3) for the sediments obtained from core A. “Industrial cycle” corresponds mainly to the atmospheric fallout of anthropogenic elements in the uppermost part of sediments.

The REEs chosen for normalization were La, Sm, Ce, Lu and Yb, because they are (i) good indicators of the clastic mineral detritus fraction,^22,27,28 (ii) show a consistent pattern throughout their profile, (iii) have a good correlation with each other, and (iv) are more accurately measured (Fig. 7a). The mean of the sum of these five elements was calculated for each sample, and the bulk concentration of elements in each sample were normalized to this value. The resulting profiles for the elements, such as As, Cu, Cr, Co and Ni, exhibit the surficial enrichment towards the top sediments (Fig. 7b and 7c). The significant change in shape of the element profiles after this step of normalization indicates the importance of the natural aluminosilicates in the flux of trace metals into the lake, which is in good agreement with other studies.^32,34–39 This is especially evident in the following example.

The vertical variation of total phosphorus, iron, and manganese obtained from bulk elemental analysis of the sediment samples is shown in Fig. 8. Correlation coefficients between Fe–Mn, Fe–P, and Mn–P indicate a significant correlation between P and Mn. In contrast, Fe shows a correlation with neither P nor Mn. This is contrary to the results of other studies, which indicate iron, manganese, and phosphorus are all related to the same “fraction” of the sediment.^40,41 This is likely due to the nature of Fe, Mn, and P, which are mainly related to the oxide portion of the sediment and, therefore, by implication are influenced by the non-oxide phases (mineral matter). Thus, the relationships between Fe, Mn, and P concentrations can be recast after correction for the clastic mineral detritus fraction using Step 3 (REEs) normalization method (Fig. 8). The correlation coefficient for Fe–Mn and Fe–P significantly improved (r = 0.90) after being corrected for the mineral matrix. This is due to the fact that ferromagnesian silicates from relatively unweathered quaternary glacial deposits constitute a significant fraction of the sediments in central Alberta lakes⁴² and hence, silicate iron (not associated with either phosphorus or manganese) is included in the analysis. These results indicate that using bulk elemental concentrations (i.e., treating the sediment as if it were a single phase), is not sufficient for the evaluation of elemental variation, and that the influence of all elemental fractions must be taken into account.

Fig. 8 Depth profiles of P, Fe, and Mn in core A before and after Step 3 normalizing (to REEs) and the inter-relationship between each profile.

In this paper, the results for core B are not presented, as the elemental distribution patterns in core B closely resembled those in core A. This similarity indicates that our interpretation is valid for various sediment types deposited in different part of the lake.

History of the lake

The results of the study were compared and correlated with the available data on the sedimentation rate (< 1.1 cm year^{−1  137}Cs dating)²⁵ and the recent history of human settlement and natural events in the area surrounding Pigeon Lake.¹⁵ The following interpretations are the results of such work:

Deforestation cycle

As a result of the normalization to Ca and Sr (Step 2), a cycle of high elemental input is identified from 69–101 cm in core A (Fig. 6b). The elements showing this increase are mainly lithophile elements. The correlation of the results with a sedimentation rate of 1.1 cm year⁻¹ obtained from ¹³⁷Cs dating and historical marker horizons²⁵ indicates that the beginning of this cycle coincided with the history of human settlement¹⁵ in the catchment area and the subsequent deforestation for agricultural and residential purposes. A high concentration of the lithophile elements is due to weathering of rocks in the drainage basin and their subsequent transport to the lake.^27,28 On the other hand, the absence of forest cover and the development of agriculture accelerates the surface erosion of fine mineral particles rich in trace elements.^7,33 Therefore, the increases in concentration of lithophile elements is likely due to the high soil erosion in the Pigeon Lake catchment caused by human settlement, deforestation and agricultural activities.

Primary productivity cycle

Following the deforestation cycle, the trophic status of the lake changed due to an increase in nutrient input.^15,43 Coincident with the Rock-Eval parameters (TOC, S2 and HI), which showed an increase in algal productivity, the sediments have reduced quantities of elements linked to erosion (Fig. 5b). This period is called the “primary productivity cycle” and occurs approximately 10 cm above the beginning of the deforestation cycle. This delay after the onset of deforestation likely reflects the time necessary for adequate nutrients to be transported to the lake in order to change the trophic status. The comprehensive discussion regarding the primary productivity cycle and historical variation of organic matter in Pigeon Lake, Alberta can be found in Sanei et al.²⁵

Industrial cycle

This cycle is defined by the increase in the amount of elements of environmental concern (e.g., As, Cd, Cr, and Cu) in the uppermost part of the sediment profiles after being normalized for the mineral matrix (Fig. 7c). Correlation of the results with the lake's sedimentation rate and the history of settlement around the lake area^15,25 indicates that the increase in the amount of heavy metals in the top sediments coincides with onset of industrial activities in the lake's surrounding area. Elements and particulates can be introduced to lakes as atmospheric fallout from automobiles and industry.^44,45 With an expanding industrialized society in central Alberta, atmospheric fallout continues to contribute material to the lakes in this region. These small quantities of elements of environmental concern found in the surface sediments of Pigeon Lake probably originated from these sources.

The results of this study indicate that the amount of these elements is negligible as compared to the quantities released by geogenic processes (erosion), since they can be identified only after being normalized to the mineral matrix (Step 3). The mean concentrations of these elements are within the range of natural values for North American and other lake sediments around the world.¹²

Acknowledgements

We would like to thank the efforts of the following people for their contribution to this project: Dr. Anthony Foscolos (Technical University of Crete, Greece) for direction on the sequential extraction experiment; Mr. Julito Reyes (GSC, Calgary) for his assistance in the field; and Mr. Marcel Labonte (GSC, Calgary), Ms. Michelle Hawke (University of British Columbia), and Mr. Travis Ferbey (University of Victoria) for their scientific and editorial help.

References

1. K. W. Bruland, K. Bertine, M. Koide and E. D. Goldberg, Environ. Sci. Technol., 1974, 8, 425.
2. K. K. Bertine and M. F. Mendeck, Environ. Sci. Technol., 1978, 12, 201.
3. A. Ng and C. Patterson, Geochim. Cosmochim. Acta, 1982, 46, 2307.
4. B. Rippey, R. J. Murphy and S. W. Kyle, Environ. Sci. Technol., 1982, 16, 23.
5. W. B. Lyons, P. B. Armstrong and H. E. Gaudette, Mar. Pollut. Bull., 1983, 14, 65.
6. P. E. Rasmussen, D. J. Villard, H. D. Gardner, J. A. C. Fortescue, S. L. Schiff and W. W. Shilts, Environ. Geol., 1998, 33, 170.
7. B. G. Lottermoser, U. Schutz, J. Boenecke, R. Oberhansli, B. Zolitschka and J. F. W. Negendank, Environ. Geol., 1997, 31, 236.
8. M. Roulet, M. Lueotte, R. Canuel, N. Farella, M. Courcelles, J.-R. D. Guimaraes, D. Mergler and M. Amorim, Chem. Geol., 2000, 165, 243.
9. E. T. Burden, G. Norris and J. H. McAndrew, Can. J. Earth Sci., 1986, 23, 55.
10. G. C. Bortleson and G. Lee, Limnol. Oceanogr., 1974, 19, 794.
11. L. Hakanson and M. Jansson, Principles of lake sedimentology, Springer Verlag, Berlin, 1983, pp. 65–69 and 258–282..
12. J. E. Fergusson, The heavy elements: chemistry, environmental impact and health effects, Pergamon Press, Oxford, 1990, pp. 306–313 and 243–329..
13. A. Tessier, P. G. C. Campbell and M. Bisson, Anal. Chem., 1979, 51, 844.
14. W. Solomons and U. Forstner, Environ. Technol. Lett., 1980, 1, 506.
15. P. Mitchell and E. Prepas, Atlas of Alberta Lakes, The University of Alberta Press, Edmonton, 1990, pp. 483–488..
16. Alberta Environment, Pigeon Lake regulation and feasibility study, Alberta Environment Planning Division, Edmonton, AB, 1982..
17. F. Goodarzi, The Impact of Anthropogenic Activities on the Wabamun Lake Area, GSC, Calgary, 1997, Project No. 950001 06-07..
18. M. Stoeppler, Hazardous metals in the environment, Elsevier, London, 1992, pp. 97–122, and 157–164..
19. L. L. Sloss and C. A. Gardner, Sampling and analysis of trace emissions from coal-fired power stations, IEA Coal Res. Publ., London, 1995..
20. E. Lafargue, F. Marquis and D. Pillot, Rev. Inst. Fr. Pet., 1998, 53/4, 421.
21. P. S. Rendell, G. E. Batley and A. J. Cameron, Environ. Sci. Technol., 1980, 14, 314.
22. E. Tipping, N. B. Hetherington, J. Hilton, J. W. Thompson, E. Bowies and J. Hamilton-Taylor, Anal. Chem., 1985, 57, 1946.
23. U. Forstner, in Speciation of Metals in Water, Sediments and Soil Systems, Lecture notes in earth sciences, ed. L. Lander, Springer-Verlag, Berlin, 1986, pp. 13–41..
24. M. Labonte and F. Goodarzi, Fuel, 1985, 64, 1177.
25. H. Sanei, F. Goodarzi, L. R. Snowdon, L. D. Stasiuk and E. Van Der Flier-Keller, Environ. Geosci. (AAPG), 2000, 7, 177.
26. W. E. Dean, E. Gorham and D. J. Swaine, Geol. Soc. Am. Spec. Pap., 1993, 276, 115.
27. N. I. Volkova, Chem. Geol., 1998, 147, 265.
28. W. E. Dean and E. Gorham, Liminol. Oceanogr., 1976, 21, 259.
29. U. Forstner and W. Salomons, Trace metal analysis on polluted sediments, Delft Hydraulics Lab. Publ., 1981, No 248, pp. 1–13..
30. F. Goodarzi, J. Brown, J. P. Charland, F. Huggins, D. Deshpandi and S. Pollock, GSC, Calgary, GSC-Bulletin, 2000..
31. M. Sager, R. Belocky and R. Pucsko, Acta. Hydrochim. Hydrobiol., 1990, 18, 157.
32. M. Ravichandran, M. Baskaran, P. H. Santschi and T. S. Bianchi, Environ. Sci. Technol., 1995, 29, 1495.
33. M. Roulet, M. Lueotte, A. Saint-Aubin, S. Tran, I. Rheault, N. Farella, E. De Jesus da Silva, J. Dezeneourt, C. J. Sousa Passos, G. Santos Soares, J. R. D. Guimaraes, D. Mergler and M. Amonm, Sci. Total Environ., 1998, 223, 1.
34. E. D. Goldberg, J. J. Griffin, V. Hodge, M. Koide and H. Windom, Environ. Sci. Technol., 1979, 13, 588.
35. H. L. Windom, S. J. Schropp, F. D. Calder, J. D. Ryan, R. G. Smith, L. C. Gurney, F. G. Lewis and C. H. Rawlinson, Environ. Sci. Technol., 1989, 23, 314.
36. S. J. Schropp, F. G. Lewis, H. L. Windom, J. D. Ryan, F. D. Calder and L. C. Gurneys, Estuaries, 1990, 3, 227.
37. C. R. Alexander, R. G. Smith, F. D. Calder, S. J. Schropp and H. L. Windom, Estuaries, 1993, 16, 627.
38. J. W. Morse, B. J. Presley, R. J. Taylor, G. Benoit and P. H. Santschi, Mar. Environ. Res., 1993, 36, 1.
39. K. D. Daskalakis and T. P. O'Connor, Environ. Sci. Technol., 1995, 29, 470.
40. J. K. Syers, R. F. Harris and D. E. Armstrong, J. Environ. Qual., 1973, 2, 1.
41. J. O. Nriagu and C. I. Dell, Am. Mineral., 1974, 59, 934.
42. Geology of Alberta Research Council, Geological Map of Alberta, Nat. Resour. Div., Alta. Geol. Surv., Edmonton, 1972..
43. J. Crosby, in Proc. Third Annual Alberta Lake Management Society Conf., Alberta Environmental Protection, Edmonton, AB, 1994, vol. 3, pp. 19–22..
44. H. Erlenkeuser, E. Suess and H. Willkomm, Geochim Cosmochim Acta, 1974, 38, 823.
45. F. Oldfield, R. Thompson and K. E. Barber, Science, 1978, 199, 679.

---

Footnote

† Presented at the Whistler 2000 Speciation Symposium, Whistler Resort, BC, Canada, June 25–July 1, 2000

---