Technical ReportPDF Available

Chemostratigraphy of the late Pleistocene Dashwood Drift to Capilano Sediment succession using portable XRF spectrometry, Nanaimo, British Columbia, Canada

Authors:
  • Geological Survey of Canada
GEOLOGICAL SURVEY OF CANADA
OPEN FILE 7651
Chemostratigraphy of the late Pleistocene Dashwood Drift to
Capilano Sediment succession using portable XRF
spectrometry, Nanaimo, British Columbia, Canada
R.D. Knight, A.M.G. Reynen, E.C. Grunsky,
H.A.J. Russell
2015
GEOLOGICAL SURVEY OF CANADA
OPEN FILE 7651
Chemostratigraphy of the late Pleistocene Dashwood Drift to
Capilano Sediment succession using portable XRF
spectrometry, Nanaimo, British Columbia, Canada
R.D. Knight, A.M.G. Reynen, E.C. Grunsky, H.A.J. Russell
2015
© Her Majesty the Queen in Right of Canada, as represented by the Minister of Natural Resources Canada, 2015
doi:10.4095/295688
This publication is available for free download through GEOSCAN (http://geoscan.nrcan.gc.ca/).
Recommended citation
Knight, R.D., Reynen, A.M.G., Grunsky, E.C., and Russell, H.A.J., 2015. Chemostratigraphy of the late Pleistocene
Dashwood Drift to Capilano Sediment succession using portable XRF spectrometry, Nanaimo, British Columbia,
Canada; Geological Survey of Canada, Open File 7651, 1 .zip file. doi:10.4095/295688
Publications in this series have not been edited; they are released as submitted by the author.
Table of Contents
1.0
Introduction ...........................................................................................................................
1
2.0
Surficial stratigraphy …………………………………………….………………………….
2.1 Dashwood Drift
2.2 Cowichan Head Formation
2.3 Quadra Sand
2.4 Vashon Drift
2.5 Capilnao Sediments
2.6 Salish Sediments Postglacial
1
3.0
Methods .................................................................................................................................
3.1 Reproducibility and precision of standards
3.2 Limits of detection
3.3 Erroneous analysis
4
4.0
Results ....................................................................................................................................
4.1 Chemostartigraphic trends Cochrane borehole
4.1.1 Dashwood Drift
Unit 1
Unit 2
Unit 3
Unit 4
Unit 5
4.1.2 Cowichan Head Formation
4.1.3 Quadra Sand
4.1.4 Vashon Drift
4.1.5 Capilano Sediments
4.2 Chemostartigraphic trends Spider borehole
4.2.1 Dashwood Drift
Unit 1
Unit 2
Unit 3
Unit 4
Unit 5
4.2.2 Cowichan Head Formation
4.2.3 Quadra Sand
4.2.4 Vashon Drift
4.2.5 Capilano Sediments
12
5.0
Summary .............................................................................................................................
26
6.0
Acknowledgements .............................................................................................................
28
7.0
References ...........................................................................................................................
28
1
1.0 Introduction
The Geological Survey of Canada in collaboration with the Regional District of Nanaimo has been
completing a geological investigation in the Nanaimo Lowland region (Fig. 1) from Parksville to
Bowser since 2009 in support a regional hydrogeology study. To better define the aquifer potential of
the region a seismic reflection survey was completed and ground truthed with five boreholes (Fig. 1).
Geochemical analyses by pXRF spectrometry were completed on the sediments collected from two of
the five boreholes (GSC-BH-CHR and GSC-BH-SPI).
Results of these analyses are crucial to defining chemical and related mineralogical variations within
sediments and augments sediment description, grain size data, downhole geophysical and stratigraphic
correlations. Geochemical data also provides an opportunity for spatial correlation and establishment of
a chemostratigraphic framework that complements other stratigraphic correlation techniques, for
example lithostratigraphy and biostratigraphy. For groundwater studies the collection of geochemical
data is often beyond the budget of most programs. However recent advances in portable X-ray
flourescence spectrometry (pXRF) as well as numerous studies on the comparison of these data with
laboratory chemical analysis (Knight et al, 2013, Morris, 2009; Radu and Diamond, 2009, Stanley et al,
2009) indicate that this type of analyses is appropriate for the fine sand to clay-size fractions. This
method has been successfully used to define the chemostratigraphy from a borehole in a Champlain Sea
aquitard in southeastern Ontario (Knight et al., 2012) and for a borehole from the Spiritwood buried
valley in southern Manitoba (Crow, et al., 2012 and Plourde et al., 2012). The integrated data sets
provide fundamental information that can be used for defining chemical and mineralogical variations
within aquifers and aquitards and are crucial to the development of basin stratigraphy, provenance and
the production of accurate 3-dimensional basin models.
The primary objective of this Open File is to release data obtained from the use of a portable X-ray
fluorescent spectrometer from a 130 m deep borehole (Cochrane) and a 117 m deep borehole (Spider)
located near Nanaimo, B.C. (Fig. 1). Secondary objectives include analysis of the geochemical data
within a stratigraphic context using principal component analysis and reference to individual elemental
trends.
2.0 Surficial stratigraphy
The investigated well sites are located within the Georgia Basin (Gabrielse and Yorath, 1992), a
structural depression that contains the Georgia Strait. An overview of the basins regional setting,
structural history and late Cretaceous infill is presented in Mustard (1994). Quaternary fill in the Strait
of Georgia has been extensively studied since Dawson (1878) reported shells in massive diamicton on
Vancouver Island. The primary source of information on the surficial geology of the area is by Hicock
(1980) and Hicock and Armstrong (1983) with recent work being carried out by Bednarski (2014) in the
area from Deep Bay to Nanoose Bay. A number of authors have reviewed the glacial history of the area,
(e.g. Fyles,1963; Clague et al., 2005) and others (Hicock and Armstrong, 1981, 1983, 1985) have
worked on aspects of the stratigraphy, chronostratigraphy, sedimentary facies, and process
interpretations (Trettin, 2004; Clague et al., 2005 and references there in). Seven regional stratigraphic
units are summarized here.
2
Figure 1. Location of the Nanaimo boreholes (from Crow at al, 2014).
GSC-BH-CHR = Cochrane, GSC-BH-SPI = Spider.
2.1 Dashwood Drift
The oldest Quaternary deposits in the study area are exposed at the base of a few sea cliff sections.
These sediments consist of unfossiliferous interbedded clay, silt, sand, and minor gravel. They were
classified as the Maplegaurd Sediment by Fyles (1963) and later incorporated as the basal unit of the
Dashwood Drift by Hicock and Armstrong (1983). Trettin (2004) identifies a number of facies changes
regionally from previous work and highlights a lower and upper member of predominantly sand and
diamicton, respectively. Fyles (1963) suggested a variety of depositional setting including, fine-grained
river flood-plain deposits, lake sediments and glacial fluvial deposits. These lowermost sediments were
not encountered in either the Cochrane or Spider boreholes.
The sediments of the Dashwood Drift that are encountered in the two boreholes consist of diamicton
overlain by mud and gravel. The gravel is a mixture of plutonic lithologies derived from the Coast
3
Mountains and volcanic and sedimentary lithologies from the east coast of Vancouver Island.
Depositional environments include a complex of glacial, glaciofluvial, ice-contact and glaciomarine to
marine sediments (Hicock and Armstrong, 1983).
2.2 Cowichan Head Formation
Cowichan Head Formation is up to 21 m thick and unconformably overlies the Dashwood Drift (Hicock
and Armstrong, 1983). The formation has been divided into a lower member of clayey silt and sand
with marine shells, and an upper member of sandy silt and gravel from volcanic and sedimentary rocks,
commonly with reddish oxidized hues rich in fossil plant remains (Armstrong and Clague, 1977, Clague,
1976). The lower member is interpreted as glaciomarine whereas the upper member is attributed to
estuarine and fluvial environments (Armstrong and Clague, 1977).
2.3 Quadra Sand
The Quadra Sand may either sharply (erosionally - unconformably) or conformably overlie the
Cowichan Head Formation and occurs over a broad area in the Georgia Depression, British Columbia
and Puget Lowland, Washington (Clague 1976; Armstrong and Clague,1977). It is up to 75 m thick and
consists of horizontally and cross-stratified sand composed of quartz, feldspar, and lithic fragments of
granitic provenance (Clague, 1976, 1977). The provenance and well sorted character imparts a ivory
(white) colour to the unit. The sand is diachronous, becoming younger to the south away from its source
in the Coast Mountains (Clague et al., 2005). The Quadra Sand is interpreted to be proglacial outwash
deposited subaerially on floodplains and locally as deltaic deposits (Clague et al., 2005). This unit is a
significant aquifer.
2.4 Vashon Drift
Vashon Drift deposits are up to 60 m thick and unconformably overlie either Quadra Sand or older units.
The Vashon Drift consists of sandy diamicton, with local mud rich, and sand and gravel facies (Fyles,
1963). Vashon Drift is interpreted as glacial till formed in a variety of ice-contact depositional
(landform) settings including esker, deltas, and fans (Hicock and Armstrong, 1985).
2.5 Capilano Sediment
Capilano Sediment unconformably overlies Vashon Drift and locally Quadra Sand. The sediment is up
to 25 m thick (Bednarski unpublished) and consists of sand and gravel with minor diamicton. The
sediment commonly fines upward. It is interpreted as a deglacial succession of coarse glaciofluvial
outwash that becomes more distal upward.
2.6 Salish Sediments Postglacial Modern
Salish Sediments are the youngest unit in the area, are generally < 5 m thick (commonly < 2 m) and
consist of gravel, sand and mud (Fyles, 1963). Depositional processes are predominantly channel and
floodplains along river valleys, deltaic sediments related to the modern sea, river and lake levels, and
recent mass wasting (e.g. alluvial fans). These sediments were not encountered in either the Cochrane or
Spider boreholes.
4
3.0 Methods
In 2011 the Geological Survey of Canada collected 17 seismic reflection profiles in the Nanaimo
lowlands of British Columbia (Pugin et al., 2009). These profiles where used to locate targets for
drilling three cored boreholes: Cochrane, Spider, and Hillier and one non-cored borehole: Qualicum.
The boreholes are located between 40 and 60 km northwest of Nanaimo, B.C. (Fig. 1).
The boreholes were drilled in the winter of 2012-13 by Mud Bay Drilling (Borat Longyear) using a full-
sized, truck-mounted Roto-Sonic drill (Fig. 2). Water with a biodegradable viscosity agent was used as
drilling fluid. Onsite geological observations were carried out by GSC staff. Sediment core was
collected from the Cochrane borehole starting at ground level. For the Spider borehole core was
collected starting at 8 feet below ground level. Both holes where terminated before contact with
bedrock. For each borehole, core was collected in 10 ft runs. The sediment cores were placed in a plastic
sleeve, sealed with tape, boxed and shipped to GSC Ottawa for further logging and sampling. A suite of
downhole geophysical logs, including magnetic susceptibility, apparent conductivity, P- and S- wave
velocities, Gamma, and fluid temperatures where obtained for each borehole in March of 2013 (Crow et
al. 2014).
Figure 2. Truck-mounted Roto-Sonic drill located at the Spider borehole.
(photo by H.A.J. Russell).
Grain size was determined for each sample being analysed for geochemistry using a Camsizer particle
scanner and a Lecotrac LT100 laser diffractometer. All grains > 2 mm in diameter were removed prior
to this analysis. Numerical results are available in Appendix A and graphically displayed in Appendix A
and B.
5
Prior to pXRF analyses the sediment was freeze-dried, disaggregated and sieved to <63 µm (silt + clay)
at the GSC Sedimentology Laboratories in Ottawa. The processed samples were placed in plastic vials
and sealed with 4 µm thick SpectroCertified® mylar. Data were acquired using a handheld Thermo
Scientific, Niton XL3t GOLDD XRF spectrometer equipped with Cygnet 50 kV, 2 watt Ag anode X-ray
tube and a XL3 silicon drift detector (SDD) with 180,000 counts per second (cps) throughput, mounted
to a test stand (Fig. 3). Samples were analyzed in Soil Mode which is recommended for elements
expected to occur with < 1% concentration. For the Spider borehole samples were also analyzed in
Mining Mode, which is recommended for elements expected to occur with > 1% concentration.
Although these data are included in Appendix A they are not discussed. In order to honor the protocol
used for previous borehole studies (Knight et al., 2012, Plourde et al., 2012) a dwell time of 60 seconds
was used for each filter (Main, Low, and High), for a total of 180 seconds per analysis. However Knight
et al., (2013) has found that dwell times as low as 30 seconds for some elements can return precise and
accurate results.
Figure 3. Analysis of processed sediment samples using the pXRF spectrometer mounted
in a test stand.
PXRF spectrometry was completed on 111 of the 115 sediment samples for the Cochrane borehole, four
samples had an inadequate residue quantity for pXRF analyses. For the Spider borehole 119 samples
were analysed (Appendix A and B).
6
The pXRF data are interpreted using single element trends from the base to the top of the boreholes.
These distributions are then integrated using Principal Component Analysis (PCA) to reduce the number
of variables and to explore multi-elemental associations within the data (Grunsky, 2010). The original
geochemical data was transformed using a centered log ratio (CLR) to overcome the problem of
closure. The PCA transformation uses linear combinations of elements and a measure of element
association to identify geochemical variations over the depth of the drill core, giving a score for each
sample for the given combination. Principal components are ordered from largest to smallest eigenvalue
with each subsequent orthogonal component (eigenvector) displaying less variation in the data than the
previous component (Fig. 4). The value of a point is defined as the score of the data as it is subsequently
projected onto a new coordinate space representing both positive and negative relationships for each
principal component. The weight or size of contribution of each element of the principal component
calculation is defined as the loading. Comparing the score and loading plots identifies co-relationships
between elements and positive and negative relationships of the elements. The relative change of these
values and their trends display the behavior of the weighted grouping of elements (i.e. a change in the
trend displays a change in the relationship of elements at that depth). Use of principal components can
reveal relationships between elements that are co-related and most likely represent a variation in
lithologic and provenance of the sediments (Grunsky, 2010). These scores can be scaled, fitted and
plotted using a method described by Zhou et al., (1983) and implemented by Grunsky (2001). As the
amount of variation for each principal component decreases the transformed data becomes noisy, which
can be attributed to random error or under sampled processes. For this study scores for PC1 and PC2 are
examined for the Cochrane borehole whereas PC1, 2, and 3 are examined for the Spider borehole.
These principal component scores are plotted with respect to the stratigraphic section where it was
observed that a cyclic variation occurred. Natural cubic splines (Hastie, 1992) were fitted to the scores
as a function of stratigraphic depth.
Fifteen elements (As, Ba, Ca, Cr, Cu, Fe, K, Mn, Rb, S, Sr, Pb, Ti, V, Zn, and Zr) were detected in
sufficient quantities to produce results using the pXRF spectrometer. The X-ray emission lines used to
determine elemental concentrations in Soil Mode are listed in Table 1. Results are presented in
Appendix A and displayed graphically in Appendix B.
3.1 Reproducibility and Precision of Standards
Results of all analyses of standards are included with the sediment datasets in Appendix A. Two
standards (Till-1 and Till-4) were analyzed after every 8 to 10 analyses of the borehole samples. A SiO2
blank was analysed to determine the cleanliness of the pXRF window and sample stand environment.
When data for the SiO2 blank returned values for elements that should not be detected in greater
amounts than trace the operating environment (test stand) was purged with compressed air and wiped
clean with Isopropyl alcohol or Methanol until the operating environment was not contaminated. The
SpectroCertified® Mylar polyester contains trace amounts of Ca and Fe. Some elements such as Ni,
V,and Zr that are not listed as known impurities returned values above the detection limit on a few
occasions and most likely represent internal detector noise. A study into the precision, accuracy,
instrument drift, dwell time optimization and calibration of pXRF spectrometry for reference materials
including Till-1 and Till-4 is available from Knight et al., (2013).
7
Summary statistics for Till-1 and Till-4 determined during analyses of samples from both boreholes are
presented in Tables 2 to 5. For each element detected in a given standard, the count, minimum value,
maximum value, mean, standard deviation, relative standard deviation (%RSD), and %error are listed.
The %error column contains the difference between the mean and recommended value. Low absolute
values in this column indicate that the element is measured accurately; high absolute values indicate that
a calibration curve is required to correct the data or that the data is not reliable. As an example Ni values
obtained from Till-1 during analyses of the Cochrane samples (Table 2) has a %error of 246. Values
returned by the pXRF were often greater than 3 times the recommended value. Thus chemostratigraphy
for Ni was not plotted.
Figure 4. A hypothetical data set projected in 3-dimensions displaying the relationships
between the 3 main principal components. From Whitbread, (2004) after Rock, (1988).
Similarly V values from Till-4 obtained during the collection of Spider data has a %error of 85, a mean
value of 124 ppm compared to a recommended value of 67 ppm. However it is useful to plot V since
chemostratigraphy utilizes the relative changes in concentration, making high precisions more important
than accuracy. It is important to note that the precision and accuracy are affected by concentration.
Lower concentrations tend to result in lower precision, and thus higher %RSD.
8
Element
Line
Energy (keV)
Window Low (keV)
Window High (keV)
Filter
As
1
10.54
10.33
10.73
Main
Ba
1
32.19
31.70
32.70
High
Ca
1
3.69
3.50
3.89
Low
Cd
1
23.17
22.60
23.60
High
Co
1
6.93
6.73
7.13
Main
Cr
1
5.41
5.24
5.59
Low
Cs
1
30.97
29.50
31.50
High
Cu
1
8.05
7.84
8.24
Main
Fe
1
6.40
6.20
6.60
Main
Hg
1
9.99
9.79
10.18
Main
K
1
3.31
3.10
3.49
Low
Mn
1
5.90
5.70
6.10
Main
Mo
1
17.48
17.23
17.68
Main
Ni
1
7.48
7.35
7.67
Main
Pb
1
12.61
12.40
12.80
Main
Rb
1
13.39
13.18
13.60
Main
S
1
2.31
2.20
2.45
Low
Sb
1
26.36
25.90
26.90
High
Sc
1
4.09
3.90
4.19
Low
Se
1
11.22
11.01
11.41
Main
Sn
1
25.27
24.70
25.70
High
Sr
1
14.16
13.95
14.38
Main
Te
1
27.47
27.00
28.00
High
Th
1
12.97
12.80
13.15
Main
Ti
1
4.51
4.21
4.70
Low
U
1
13.61
13.48
13.90
Main
V
1
4.95
4.80
5.10
Low
W
1
8.40
8.26
8.49
Main
Zn
1
8.64
8.49
8.83
Main
Zr
1
15.77
15.53
15.98
Main
Table 1: X-ray energy intensities used to determine elemental concentrations
in Soil Mode, as provided by Niton.
3.2 Limit of Detection
For this study the limit of detection (LOD) using pXRF spectrometry is defined as 2 standard deviations
of the individual measurements taken throughout the 180 second analysis. It should be noted that for
some sediments (e.g. Quadra Sand) two elements (As and S) returned analyses lower than the LOD.
When this occurred, the point was plotted on the chemostratigraphy graphs using the LOD value.
Elements detected by each filter and the corresponding lower limits of detection are listed in Table 6.
9
3.3 Erroneous Analysis
During analyses of the Cochrane borehole samples three results obtained for Till-1 returned either very
low or very high concentrations for the majority of elements and are considered to be erroneous. The
results were removed from the dataset presented in Appendix A. This emphasizes the importance of
monitoring results in real time to ensure that operating conditions are optimal. No unexplained or
erroneous data was collected from the borehole samples.
Recommended
Value (ppm)
Count
Mean
(ppm)
%error
Std Dev
(ppm)
%RSD
Minimum
(ppm)
Maximum
(ppm)
As
18
22
18
0
1.2
6.61
16
20
Ba
702
22
860
22.5
18
2.05
817
887
Ca
19440
22
17536
-9.8
163
0.93
17237
17862
Co
18
14
158
777.8
27
17.11
121
226
Cr
65
22
34
-47.7
5
13.31
24
41
Cs
1
22
46
4502
2.41
5.24
42
51
Cu
47
22
61
29.8
5
8.77
53
71
Fe
48100
22
40820
-15.1
256
0.63
40383
41287
K
18429
22
15797
-14.3
2
1.38
15427
16290
Mn
1420
22
1367
-3.7
35
2.54
1284
1420
Mo
2
22
5
150
1.5
33.2
1.5
8.2
Ni
24
22
83
246
7.4
8.95
70
97
Pb
22
22
13
-40.9
1.66
12.89
10
16
Rb
44
22
41
-6.8
0.9
2.28
39
43
S
< 500
22
185
-63
90
48.55
138
380
Sr
291
22
268
-7.9
3
1.12
262
274
Th
5.6
21
4.27
-23.7
0.92
21.61
2.83
6.67
Ti
5990
22
5441
-9.2
95
1.75
5272
5579
U
2.2
4
5.35
143
0.72
13.40
4.87
6.42
V
99
22
155
56.6
10
6.41
122
169
Zn
98
22
91
-7.1
2.8
3.02
86
95
Zr
502
22
574
14.3
6.0
1.04
560
585
Table 2. Summary statistics of Till-1 by pXRF spectrometry for the Cochrane borehole. Italics represent data
obtained from reference materials that were less than the limits of detection for borehole samples.
10
Recommended
Value (ppm)
Count
Mean
(ppm)
%error
Std Dev
(ppm)
%RSD
Minimum
(ppm)
Maximum
(ppm)
As
18
17
18
0
1.2
6.93
16
21
Ba
702
17
867
23.5
18.5
2.14
835
896
Ca
19440
17
17470
-10.1
254
1
169327
17808
Co
18
3
169
838.9
65.6
38.82
128
245
Cr
65
17
44
-32.3
4.74
10.75
35
51
Cs
1
17
47
4600
3.02
6.40
40
52
Cu
47
17
59
25.5
4.87
8.25
51
67
Fe
48100
17
40934
-14.9
253
1
40369
41272
K
18429
17
15749
-14.5
157
1
15433
15957
Mn
1420
17
1368
-3.7
29
2.1
1317
1415
Mo
2
3
5
150
1.93
38.2
3.1
7
Ni
24
17
84
250
8.1
9.58
61
98
Pb
22
17
12
-45.5
1.69
13.97
8
15
Rb
44
17
41
-6.8
0.63
1.55
40
42
S
< 500
7
406
-18.8
75
18.48
329
516
Sr
291
17
270
-7.2
1.75
0.65
267
272
Th
5.6
17
4.19
-25.5
1.11
26.57
2.7
5.9
Ti
5990
17
5459
-8.9
113
2
5137
5611
U
2.2
10
5.62
155.5
0.99
17.57
4.58
7.8
V
99
10
165
66.7
14
8
132
187
Zn
98
17
92
-6.1
3.35
3.65
84
97
Zr
502
17
569
13.3
7.72
1.36
556
581
Table 3. Summary statistics of Till-1by pXRF spectrometry for the Spider borehole. Italics represent data
obtained from reference materials that were less than the limits of detection for borehole samples.
Recommended
Value (ppm)
Count
Mean
(ppm)
%error
Std Dev
(ppm)
%RSD
Minimum
(ppm)
Maximum
(ppm)
As
111
19
102
-8.1
2.6
2.58
98
108
Ba
395
19
454
14.9
18.7
4.11
421
486
Ca
8934
19
7824
-12.4
123
1.57
7651
8032
Co
395
10
146
-63
23.8
16.28
108
185
Cr
53
19
26
-50.9
4.3
16.57
18
34
Cs
12
19
25
108
3.4
13.75
18
32
Cu
237
19
215
-9.3
5.77
2.65
206
225
Fe
39700
19
33039
-16.85
222
0.67
32597
33504
K
26980
19
23709
-12.1
377
1.59
23192
24449
Mn
490
19
442
-9.8
19.3
4.37
405
473
Mo
16
19
20
25
1.5
7.78
17
22
Ni
17
19
55
224
9.1
16.62
44
79
Pb
50
19
43
-14
2.47
5.74
39
47
Rb
161
19
152
-5.6
1.7
1.12
148
155
S
800
19
576
-28
104
18.03
320
762
Sr
109
19
105
-3.7
1.23
1.13
103
108
Th
17.4
19
42
141
1.5
3.61
39
45
Ti
4840
19
5951
23
79
1.32
5712
6127
U
5
9
8.15
63
1.5
18.63
6.11
10.52
V
67
19
117
74.6
8.6
7.32
101
137
W
204
19
186
-8.8
11.9
6.40
164
203
Zn
70
19
65
-7.1
2.9
4.46
61
71
Zr
385
19
437
13.5
12.1
2.77
419
454
Table 4. Summary statistics of Till-4 by pXRF spectrometry for the Cochrane borehole. Italics represent data
obtained from reference materials that were less than the limits of detection for borehole samples.
11
Recommended
Value (ppm)
Count
Mean
(ppm)
%error
Std Dev
(ppm)
%RSD
Minimum
(ppm)
Maximum
(ppm)
As
111
17
104
-6.3
2.11
2.03
100
108
Ba
395
17
452
14.4
10.6
2.34
429
471
Ca
8934
17
7849
-12.1
87
1.1
7647
7989
Co
395
3
146
-63
23.8
16.28
108
185
Cr
53
17
32
-39.6
5.14
16.06
17
41
Cs
12
17
23
91.7
2.46
10.52
20
27
Cu
237
17
217
-8.4
5.41
2.50
208
230
Fe
39700
17
33239
-16.3
203
0.6
32916
33656
K
26980
17
23811
-11.7
214
0.9
23521
24406
Mn
490
17
443
-9.6
25
5.6
399
498
Mo
16
17
17.9
11.9
1.38
7.70
15
20
Ni
17
17
57
235
7.42
13.12
46
71
Pb
50
17
41
-18
1.31
3.20
39
43
Rb
161
17
152
-5.6
1.64
1.08
148
154
S
800
17
624
-22
95
15.16
422
802
Sr
109
17
106
-2.8
1.01
0.95
104
107
Th
17.4
17
42.7
145
1.25
2.94
40.7
45.2
Ti
4840
17
4612
-4.7
56
1.2
4530
4725
U
5
16
13
160
2.91
22.72
6
18
V
67
17
124
85
9.6
7.76
106
146
W
204
17
179
-12.3
8.66
4.83
162
194
Zn
70
17
65
-7.1
3.10
4.76
61
70
Zr
385
17
426
10.6
12.3
2.89
409
467
Table 5. Summary statistics of Till-4 by pXRF spectrometry for the Spider borehole. Italics represent data
obtained from reference materials that were less than the limits of detection for borehole samples.
Element
Matrix
Filter
SiO2
SiO2 + Fe +Ca
As
4
7
High
Ba
35
45
Low
Ca
40
N/A
Low
Cu
10
13
Low
Cr
10
22
Main
Fe
25
N/A
Main
K
45
150
Low
Mn
35
50
Main
Mo
3
3
Main
Ni
25
30
Main
Rb
3
3
Main
S
75
275
Low
Sc
10
75
Main
Sr
3
3
Low
Ti
20
60
Low
V
10
25
Low
Zn
7
10
Main
Zr
3
4
Main
Table 6. Elements detected in the Nanaimo and Spider boreholes with corresponding detection limits for
the pXRF using two matrix configurations and the filters used to detect these elements, Thermo Scientific
12
4.0 Results
In glaciated terrains of the late Pleistocene the <0.063mm size fraction of unconsolidated sediment
represents crushed bedrock detritus, mineral grains, and grain fragments that are often un-weathered
(McMartin and McClenaghan, 2001). For sediments analysed in the two boreholes Na and K likely
represent granitic provenance, Ca represents carbonate terrains, while Fe, Co, Cr, Ni, V, and Zn
represent volcanic source rocks. Until mineralogy (X-ray diffraction) of the sediments is undertaken
pXRF spectrometry can only infer generalizations with regards to source rock types.
The relationship of grain size to the interpretation of the pXRF derived geochemistry has been discussed
in Zhu et al., (2011) and Knight et al., (2012). Downhole grain size data are presented in Appendix A
and plotted with pXRF derived concentrations in Appendix B. For both boreholes the Dashwood Drift
has a greater abundance of silt and clay compared to the overlying sediments. The Quadra Sands display
lower abundance of silt and clay with commensurate greater abundance of sand compared to the
underlying and overlying sediments. The Vashon Drift displays an increase in clay compared to both
the underlying Quadra Sand and the overlying Capilano Sediments. Sand content of the Capilano
Sediments is greater that the underlying Vashon Drift.
4.1 Chemostratigraphic Trends Cochrane borehole
Basal sediments of the Cochrane borehole are assigned to the Dashwood Drift and consist of 2
diamicton horizons separated by a mud/clay horizon. These sediments are overlain by a coarsening
upwards sequence (mud to medium grained sand) through the Cowichan Head Formation to the
overlying Quadra Sand. The lower half of the Quadra Sand consists of medium grained sand while the
upper half consists primarily of fine grained sand with coarse grained sand horizons. The top 10 meters
of the Quadra Sand consist of medium grained sand with pebbles. These sediments are overlain by the
Vashon Drift, a sandy diamicton. Uppermost sediments consist of a lower mud/clay unit overlain by
fine-grained sand with organic matter assigned to the Capilano Sediments.
Chemostratigraphy of the Cochrane borehole is discuss with respect to changes in the trend of
concentration of single element chemistry (Appendix B) as well as multi-element principal component
analyses. Individual elemental concentrations are plotted with depth adjacent to the stratigraphic section
for the Cochrane borehole and displayed in Appendix B. Figure 5 displays the centered log ratio
transformed geochemical data for principal components (PC) 1 and 2 plotted adjacent to the Cochrane
borehole stratigraphic section.
Elements of similar relationship are plotted on Figure 6 as a positive correlation and elements with an
opposite relationship display a negative correlation. The inset plot on Figure 6 displays the variation of
the lengths of axes of each principal component. In this inset diagram little variation in the data occurs
from PC3 onwards, thus only PC1 and PC2 are considered to be significant.
PC1 is dominated by a positive co-relationship for Ba, Ca, K, Mo, Sr and Zr with high scores being
given to Ca and Sr (Fig. 6) and negative scores for As, Cu, Cr, Fe, Mn, Ni, Ti, V, and Zn (Fig. 6).
PC2 is dominated by a positive co-relationship of scores for Ca, Fe, Mo, Sr, Ti, V, and Zr with high
scores for Fe (Fig. 6) and a negative relationship for As, Ba, K, S, Ni, Rb, and Zn (Fig. 6).
13
Figure 5. Centered log ratio transformed geochemical data for PC1 and PC2 plotted adjacent to
the Cochrane borehole stratigraphic section.
4.1.1 Dashwood Drift
For the basal Dashwood Drift there were 49 analyses determined from the Cochrane borehole. These
sediments can be subdivided into five units based on changes in the trend single elemental
concentrations, principal component analyses results of multiple related elements and/or changes in
grain size. The units may not correspond to visual changes in sediments (eg. Unit 4/5 boundary) but do
represent a shift in provenance and/or depositional processes. For most elements their concentrations
are proportional to changes in abundance of silt and/or clay. However there are sediment horizons
where a change in a single elemental concentration is not reflected by a change in grain size. For
example, there is a sharp decrease in abundance of Ca and Sr and an increase in abundance for K, Mn,
14
Rb, and Zn at a depth of 115.5-121.5 meters (Fig. 7) where there is little to no change in grain size from
the unit below or above. Some of these unit subdivisions are further substantiated by downhole
magnetic susceptibility and apparent conductivity (Crow et al, 2014). For the Cochrane borehole the
contact between the Dashwood Drift and the overlying Cowichan Head Formation brown muds consists
of a sharp boundary.
Figure 6. Principal component loading and scores for PC1 and PC2 for the Cochrane borehole
data. Elements of similar relationship are plotted as a positive correlation while elements with
an opposite relationship plot as a negative correlation. The inset diagram displays the
eigenvalues and the number of eigenvectors.
Unit 1
Nine analyses were determined from the approximately 7 + metres thick unit. A thin horizon of
sand at a depth of 127 meters (Fig. 7) is reflected by an increase in concentration of Ca, Cr, Fe, V, and
Zr, with a corresponding decrease in both K and Rb. It should be noted that the concentration of Sr does
not seem to be affected by this sand horizon (Appendix B). Unit 1 is differentiated from overlying Unit
2 by changes in the trends of Ca, Cu, K, Mn, Rb and Zn.
15
Unit 2
The contact between the underlying Unit 1 and Unit 2 occurs at the base of a gravel / pebble
horizon. This unit is 7 meters thick, and displays a minor decrease in sand content with an increase in
silt compared to unit 1. Analyses were determined for 9 samples. Results display a marked decrease in
Ca from a mean of 33520 ppm for unit 1 to mean of 27255 ppm for unit 2, and a slight decrease in Sr
from a mean of 238 ppm compared to a mean of 298 ppm for Unit 1 (See Appendix A). The transition
from Unit 1 to Unit 2 is reflected in a change from a negative to more positive co-relationship of PC2
(Fig.5). Single element concentrations for K, Rb, (Fig. 7) Mn, and Zn all have higher values in Unit 2
compared to the underlying Unit 1. However some elements such as Fe, Ti, V, and Zr display little to no
changed between the units (Appendix B).
Figure 7. At a depth from115.5 and 121.5 meters there is a decrease in Ca concentrations and
an increase in K and Rb concentrations without any significant change in silt or clay content. At
a depth of 127 meters a decrease in silt and clay (sand horizon) is marked by a significant
change in several elements including Ca, K, and Rb.
Unit 3
Sixteen analyses were determined from this 15 m thick unit. The unit displays a slight
coarsening upwards trend from a decrease in clay and a corresponding increase in silt (Fig. 7). The
inflection points of the trend curves for both PC1 and PC2 (Fig. 5) do not correspond to the exact depth
of the change from unit 2 to unit 3, as defined by the increase in concentration of Ca (Fig. 7), but do
display a shift for PC1 just below the contact and for PC2 just above the contact. There is a small
increase in silt and sand in the upper 3-5 m of unit 3 (101-106 meters in depth, that is reflected in a
decrease in concentration of Cr, Cu, Fe, Ti, V, and Zn (Fig. 8) and a minor increase in the concentration
of Ba, Ca, K, Sr and Zr. Across the lower 3 units Sr, Ti, and V display little change until the top of unit
3 where Sr decreases and Ti and V increase in concentrations. PC 1 however shows a marked change
from positive scores to negative scores corresponding to the gravel horizon observed at a depth of 108
meters (Fig. 5).
Unit 4
The contact between Unit 3 and the overlying Unit 4 occurs between the top of the diamicton
and the overlying clay (Fig. 8). Eight analyses where determined from this 11.5 meter thick unit. The
lower and upper contact of this unit is defined by the concentration of S. The upper contact occurs
within the upper diamicton of the Dashwood Drift. Compared to unit 3, unit 4 displays a marked
decrease in sand and a corresponding increase in both silt and clay. At the contact between unit 4 and 5
there is a further drop in sand content. This unit contains S concentrations with a mean of 1053 ppm
compared to a range of mean values from 486- 248 ppm for the other units (Appendix A). Unit 4
16
deviates from the units below by increases in the concentration of As, Ca, Cr, Cu, Fe, Mn, S, Ti, V, and
Zn. There are also decreases in the concentration of Ba, K, Rb, Sr and to minor degree Zr. The change
from unit 3 to 4 is reflected in a shift form positive to negative scores for PC1.
Unit 5
Seven analyses were determined within this 4 meter thick unit. The unit is characterized by an
upwards decreasing concentration of Ca from 264 ppm ant the base to 187 ppm at the top of the unit. S
concentrations decrease to below detection limits. Both units 4 and 5 display an upward increase in the
concentration of Cr, Cu, Ti, and V, and a decrease in concentration of K and Sr (Appendix B). Zr
concentrations remain the same throughout both units 4 and 5. The upper contact of unit 5 corresponds
to a shift in trend from negative to more positive PC1 scores.
Figure 8. At a depth between 101 and 106 meters there is an increase in the sand and silt
content and a corresponding decrease in Cu, Fe, and Zn concentrations.
4.1.2 Cowichan Head Formation
The Cowichan Head Formation is 2 meters thick. Two samples were collected for pXRF spectrometry.
For most elements in the Cochrane borehole the concentration of these samples represents a transition
from the underlying Dashwood Drift and the overlying Quadra Sand (Appendix B). The lower sample
often is similar in elemental concentrations to that of the underlying Dashwood Drift unit 5 values and
the upper sample is similar in elemental concentrations to the overlying Quadra Sand (eg: Cr, Cu, Fe and
V). This is also displayed in the large shift from negative PC1 scores in the underlying Dashwood Drift
unit 5 to positive PC1 scores in the overlying Quadra Sand (Fig. 5). Previous work by Alley (1979)
indicates a lag deposit in some localities between the Dashood Drift and the Cowichan Head Formation.
This lag deposit was not observed in either the Cochrane or Spider boreholes.
4.1.3 Quadra Sand
For the 78 m thick Quadra Sand 48 analyses were analysed. The sand displays a marked decrease in
elemental abundances compared to both the underlying and overlying sediments for As, Cu, Fe, (S), Ti,
V, and Zn whereas there is a marked increase in elemental abundances for Ca and Sr (Fig. 9). These
trends in elemental concentrations correlate with a decrease in silt and clay and an increase in sand (Fig.
9). For Sr the mean of the concentration for the Dashwood Drift is 285 ppm, for the Cowichan Head
Formation, 201 ppm and for the Vashon Drift 262 ppm, compared to the Quadra Sand with a mean
concentration of 511 ppm. PC1 displays a strong shift to positive scores in the lower portion of the
Quadra Sand followed by a slight shift towards negative scores in the mid portion of the sands (Fig. 5).
This is followed by a further slight shift towards more positive scores at a depth of 40 meters before a
17
shift to strong negative scores to the top of the borehole. PC 2 displays several shifts (~5) in the trend of
scores throughout the sands with a strong inflection at the contact with the overlying sandy Vashon Drift
(Fig. 5).
4.1.4 Vashon Drift
For the 7 m thick Vashon Drift 8 analyses where determined. Vashon Drift can be differentiated into an
upper and lower unit based on variations in elemental abundances and changes in both the silt and clay
content. Ca (Fig. 10), Cu, and Sr (Fig. 9) display a much lower concentration in the basal 4 samples and
sharp increase in concentration for the upper 4 samples. Furthermore, Rb (Fig. 10), As, K, and Zn
(Appendix B) display the opposite trend with higher concentrations in the lower four samples compared
to the upper four. Elements such as Fe (Fig. 10), Ba, Mn, and especially Ti, and V and to a lesser
degree Zr display no change with depth (Appendix B). For both the Vashon Drift and Capilano
Sediments PC1 continues the trend towards negative scores that was established in the upper 10 meters
of the Quadra Sand (Fig. 5). PC 2 displays a sharp inflection in a trend to negative scores to a trend
towards positive scores at the contact between the Quadra Sand and the Vashon Drift. This trend
continues to the top of the borehole (Fig. 5).
Figure 9. Quadra Sand is recognized by an increase in sand content and a decrease in clay
content. Geochemically several elements including Cu and Ti display a decrease while Sr
displays an increase in concentration’s from both the underlying and overlying sediments.
4.1.5 Capilano Sediments
For the 3 m thick Capilano Sediments 4 analyses were determined. With such a low number of samples
it is difficult to ascertain geochemical trends with any degree of certainty however there is a decrease in
concentration from the base of the Capilano Sediments upwards for Ba, Ca, K, Rb, and Zn. Similarly
there is an increase in the concentration for Cu, Fe, Mn, and V (Fig. 10). The grain size for this interval
decreases in clay and silt with a corresponding increase in sand however there is no direct relationship
between the trends in concentration and the variations in grain size.
18
Figure 10. Vashon Drift is divided into 2 units, differentiated by minor changes in sand and silt
content and distinctive change in some element concentrations such as Ca and Rb. Note that
some elements such as Fe display little variation between the upper and lower portion of the
Vashon Drift.
4.2 Chemostratigraphic Trends Spider borehole
Basal sediments of the Spider borehole are assigned to the Dashwood Drift and consist of 3 diamicton
horizons separated by a mud/clay horizon. These sediments are overlain by a coarsening upwards
sequence (mud to coarse grained sand) similar to the Cochrane borehole. At Spider these sediments are
overlain by a second coarsening upwards sequence from fine sand to coarse sand and gravel. In turn
these sediments are overlain by a fining upwards sequence to a several meter thick silt horizon. These
coarsening and fining upwards sequences are assigned to the Cowichan Head Formation. A sharp
increase in sand content marks the contact with the overlying Quadra Sand. Opposite to the Cochrane
borehole the lower half of the Quadra Sand consists predominantly of fine grained sand while the upper
half consists primarily of fine and medium grained sand. These sediments are overlain by the Vashon
Drift, a sandy diamicton that is thinner in comparison to the Cochrane borehole. Uppermost sediments
consist of fine-grained sand assigned to the Capilano Sediments.
Individual elemental concentrations and grain size data are plotted with depth adjacent to the
stratigraphic section for the Spider borehole and displayed in Appendix B. Centered log ratio
transformed geochemical data for principal components (PC) 1, 2 and 3 are plotted adjacent to the
stratigraphic section of the Spider borehole and displayed in Figure 11. Elements of similar relationship
are plotted on Figure 12 and 13 for PC1 vs PC2 and PC 2 vs PC3. For these figures elements with a
similar relationship display a positive correlation while elements with opposite relationship display a
negative correlation. The inset plot on Figure 12 displays the variation of the lengths of axes of the
principal components. In this inset diagram little variation occurs from PC4 onwards, thus only PC1, 2
and 3 are considered to be significant.
PC1 is dominated by a positive co-relationship for Sr, K, Ba, and Rb and negative scores for Fe, Mn, Ni,
Ti, and V (Fig. 12).
PC2 is dominated by a positive co-relationship of scores for Fe, Mn, Ti, and V and a negative
relationship for As, Cr, and Cu (Fig. 12, 13).
PC3 is dominated by a positive co-relationship of scores for Rb, K, Zn, and As and a negative
relationship for Ca (Fig. 13).
19
Figure 11. Centered log ratio transformed geochemical data for PC1, PC2 and PC3 plotted
adjacent to the Spider borehole stratigraphic section.
4.2.1 Dashwood Drift
For the 41m thick Dashwood Drift 50 analyses were determined from the Spider borehole. These
sediments can be subdivided into five units based on changes in sediment character, elemental
concentrations and grain size (Appendix B). Some elements such as Cu, Ti (Fig. 14), Cr, Fe, V, and Zn
(Appendix B), display an overall increase in concentration or in the case of Sr (Fig. 14) little change in
concentration from the base of the formation upwards through the Dashwood Drift and approximately
20 m into the over lying Cowichan Head Formation, even though the silt and clay content decreases at
the contact. The lower half of the diamicton sequence displays S concentrations that vary from below
detection limit (275 ppm) to a maximum value 720 ppm (AppendixB).
20
Figure 12. Principal component loading and scores PC1 and PC2 for the Spider data. Elements
of similar relationship are plotted as a positive correlation while elements with an opposite
relationship plot as a negative correlation. The inset diagram displays the eigenvalues and the
number of eigenvectors.
Unit 1
The lower unit comprises a diamicton that is 13 meters thick in core with the basal contact
defined by the end of the core. Analyses were determined for 15 samples. The basal portion of this unit
consists of a thin horizon of coarse sand and gravel that is reflected with changes in the concentration of
Ca, K, Rb, Ti, and V.
Unit 2
The contact between the underlying Unit 1 and Unit 2 occurs at a depth of 104 meters
corresponding to an increase in the sand content and a decrease in the clay content. This unit is 7 meters
thick. Analyses were determined for 7 samples. Results display a marked decrease in K from a mean of
7899 ppm for unit 1 to mean of 7125 ppm for unit 2, and a slight increase in Ca from a mean of 42899
ppm compared to a mean of 46696 ppm for Unit 2 (Fig. 15). Rb, and Zn also display a decrease in
concentrations compared to both the underlying and overlying sediments. The contact between unit 1
21
and 2 is also defined by a shift in trend for the co-relationship of elements and an inflection point at a
depth of 104 meters (dashed line on Fig. 11) that most likely reflects the increased loading of Ca,
possibly indicting a change in provenance for sediment between these 2 till units.
Figure 13. Principal component loading and scores PC2 and PC3 for the Spider data. Elements
of similar relationship are plotted as a positive correlation while elements with an opposite
relationship plot as a negative correlation. The inset diagram displays the eigenvalues and the
number of eigenvectors.
22
Figure 14. Although there is a decrease in silt and clay content at the Dashwood Drift
Cowichan Head Formation contact some elements including Cu, Ti and Sr continue in
concentration trends from the bottom of the Dashwood Drift into the overlying Cowichan Head
Formation sediments.
Unit 3
Six analyses were determined within this 4 m thick mud sequence that contains marine shells. An
increase in silt and clay content from 93-97 meters in depth is reflected in an increase in concentration of
As, Cr K, (S), Ti and Zn and a decrease in concentration of Ca (Fig. 15) and Sr. Both Ti and Sr display
a significant change in concentrations from sediments belonging to units 1 and 2 compared to units 3-5.
PC1 and PC2 display a change in the relationship of elements at the contact with the overlying unit 4
sediments (Fig. 11).
Figure 15. Unit 2 of the Dashwood Drift is defined by an increase in sand content and Ca
concentration’s and, a decrease in clay content and K and Rb concentrations.
Unit 4
Unit 4 is composed of a diamicton that is capped by a 1 meter thick mud/clay horizon containing marine
shells. Ten analyses where determined from this 7 meter thick unit. The unit displays a marked increase
in sand and a corresponding decrease in both silt and clay (Apeendix B). Ca and Sr display an increase
in concentrations from both the underlying and overlying sediments. The upper 1 meter thick clay
horizon displays a spike in concentrations of K, Zn (Fig. 16), Ba, Rb (Appendix B) and a negative
spike in Ca, Fe, Mn, Ti, and V (Appendix B). For some elements such as Ba, Ca, there is considerably
more variation in concentrations of samples collected from unit 4 compared to those collected from unit
5. The unit 4/5 contact is also visible in the S and Sr concentration changes.
23
Figure 16. Changes in sand and clay content for unit 5 of the Dashwood Drift are visible in
elemental concentration changes for K and Zn while Sr displays no change in concentration.
Unit 5
Twelve analyses were determined within this 8 meter thick unit. The uppermost sample was collected
5cm below the contact with the overlying Cowichan Head Formation. Unit 5 displays sand contents that
are higher in the lower half than the upper half with corresponding changes in silt and clay contents.
These changes are visible in both K and Zn concentrations (Fig. 16) but are not detectable for most
elements. For example Sr shows no variation through the unit including the 1 meter thick mud/clay at
the horizon at the contact with the overlying Cowichan Head Formation Sr is slightly diminished in
concentration compared to the underlying unit 4 sediments. The diamicton of unit 5 displays S values
with a mean concentration of 688 ppm compared to values mainly below detection limit for both the
underlying and overlying sediments.
At a depth of 83 meters PC1, 2, and 3 all display an inflection point (Fig. 11) indicating that there is
change in the covariance of multiple element chemistry that is not visible in any single element and is
not related to a change in grain size. The contact between the Dashwood Drift and the Cowichan Head
Formation is visible with PC2 (Fig. 11). Sediment observations at this contact suggest a sharp
boundary, however for elements associated with PC1 and PC3 there is little to no change in the
concentration from the Dashwood Drift to the sand of the Cowichan Head Formation suggesting that the
sediment provenance does not change.
4.2.2 Cowichan Head Formation
Eighteen samples were collected from the 18.5 m thick Cowichan Head for pXRF spectrometry. The
formation can be subdivided into a lower oxidized mud rich horizon coarsening-upwards to a coarse-
sand. These sediments are overlain by a second coarsening-upwards sequence grading from fine sand to
coarse sand with cobbles and gravel lenses. These coarse sands and gravels fine upwards to a medium-
sand that is capped by a 4.5 meter thick clay unit similar to the clay horizon at the top of the Cowichan
Head Formation in the Cochrane borehole. For the lower Cowichan Head Formation both the sand and
silt content vary considerably, however, both an overall lower coarsening-upwards sequence followed
by a fining-upwards sequence is depicted in figure 17. The upper Cowichan Head Formation consists of
a silt-rich horizon (near 95%) and low sand content.
24
Figure 17. Cowichan Head Formation coarsening-and fining-upwards sequences. For the lower
Cowichan Head Formation Ba, Cr, and Sr display geochemical concentrations similar to the
underlying Dashwood Drift with upper Cowichan Head Formation elemental concentrations
similar to the overlying Quadra Sand.
For most elements the lower and middle portion of the Cowichan Head Formation returns concentrations
that are similar to that of the underlying Dashwood Drift. This includes a continual upward increase in
concentration for Cr, Cu, Fe, (K), Mn, Ti, V, and Zn (Appendix B). Elements such as Ba and Sr display
little to no change in concentration through the upper Dashwood Drift and lower Cowichan Head
Formation (Fig. 17). PC3 displays a shift in trend from negative values to positive values at a depth of
74.5 meters corresponding to the top of the coarse-grained sand portion of a coarsening-upwards
sequence (Fig 11). At a depth of 72 meters PC1 has a shift in trend from more negative values to
positive values at the base of the course sand unit while PC2 with a similar trend defines the gravel
horizon (-68.5 m) within the course sand unit (Fig. 11).
Elemental concentrations of the upper silt display similar values to the overlying Quadra Sand (e.g. Ba,
Cr, Cu, Rb, Sr) compared to the underlying lower Cowichan Head Formation coarse sands (Fig. 17).
For some elements (Fe, K, Ti) the upper silt unit is very distinct from both the underlying coarse sands
and the overlying Quadra Sand (Appendix B).
Overall PC1 displays a significant positive shift towards the base of the Quadra Sand where there are
higher concentrations of Sr (Fig. 11).
4.2.3 Quadra Sand
For the 52.5 m thick Quadra Sand 47 analyses were determined. Grain size displays very little variation
within the unit (Fig. 18). For some elements the sand displays a distinct geochemical signature
compared to the underlying and overlying sediments for Cr, K, Mn, Rb, Ti, and V (Appendix B). For
Cu, K and Sr (Fig. 18) and Ba, Ca, and Zn (Appendix B) concentrations are similar to the underlying
Cowichan Head Formation These trends in elemental concentrations correlate with a decrease in silt and
clay content. This is especially notable at 54 m depth where there is a contact between medium to fine
sand and an overlying 2 meter thick silt horizon. At this contact there is a marked decrease in Ba, Ca,
Mn and Sr with an increase in K and a spike in S, the only sample to return a value above detection
limits (Appendix B). PC1 displays little variation throughout the Quadra Sand. From 45 m depth to the
base of the Quadra Sand PC2 and PC3 display opposite relationships (Fig. 11). This is also somewhat
true for a depth of 20 m to 40 m however at the top of the Quadra Sand and both PC2 and PC3 display a
similar trend. The strong change in co-relation of elements associated with PC2 and PC3 at a depth of
25
30 meters suggests that the lower portion of the Quadra Sand represents sediment of a different
lithology, provenance and depositional environment however individual element concentrations do not
display any significant change in concentration levels. Clague (1976) classifies the depositional
environment of the Quadra Sand as a short transport braided river system. Trettin (2004) divides the
sands into 2 members, the lower being associated with a high energy, nearshore delta and the upper
associated with a fluvial braidplain environment. PC2 and PC3 confirm that there is a difference
between the lower and upper portion of the Quadra Sand.
Figure 18. Sand and silt content for the Quadra Sand. Concentrations of Cu and Sr are similar
to the underlying upper Cowichan Head Formation whereas K displays concentration’s similar
to the lower Cowichan Head Formation.
4.2.4 Vashon Drift
For the 1 m thick Vashon drift two analyses where determined. For Cu (Fig. 18), Fe, Mn, and V
(Appendix B), there is a marked decrease in the concentration compared to the underlying Quadra Sand
and an increase in concentrations with the overlying Capilano Sediments. For Ba, Ca and to a lesser
degree K (Fig. 18) there is a marked decrease in concentrations between the Vashon and the overlying
Capilano Sediments (Appendix B). PC1, 2, and 3 each display a tnd towards negative scores for both the
Vashon and Capilano Sediments that may reflect near surface weathering reactions and solubility of
some elements.
4.2.5 Capilano Sediments
For 1.5 m thick Capilano Sediments two analyses where determined. With only 2 analyses it is difficult
to ascertain geochemical trends in the unit however, there is a decrease in concentration from the
underlying Vashon Drift to the Capilano Sediments for Ba, Ca, and Ti. Similarly there is an increase in
the concentration for Cu, Fe, Mn, Sr, and V (Fig. 18). These relationships are reflected in the change in
grain size between sand, silt and clay content.
26
5.0 Summary
Chemostratigraphic data obtained from the use of a portable X-ray fluorescent spectrometer for samples
obtained from a 130 m deep (Cochrane) and the 117 m deep (Spider) borehole can be used to
differentiate the sediments based on changes in elemental concentrations. These data can be plotted and
compared with the visual stratigraphic sediment descriptions and laboratory grain size analyses.
Individual element concentrations can be subjected to multi-element principal component analyses to
reduce the number of variables being examined.
From data analyses the lowermost Dashwood Drift displays local variability with probable
discontinuous gravel lenses in the lower portion and variable thickness marine clays in the upper
portion. Core logging of the till indicates consistency for much of the till; however, changes in both
single elemental abundances and multi-element co-relations indicate local variability within these
sediments.
For both the Cochrane and Spider boreholes the Cowichan Head Formation sediments directly overlying
the Dashwood Drift comprise a few meter thick coarsening-upwards sequence. In the Spider borehole
lower Cowichan Head Formation sediments consist of an additional 10 meter thick coarsening-upwards
sequence that is not present in the Cochrane borehole. These sands can be differentiated from the
overlying upper Cowichan Head Formation sediments and the Quadra Sand by the change in
concentration of Cr, Cu, Mn, and Sr (Fig. 19). Hicock and Armstrong (1983) describe two distinct
patterns of sedimentation that represent a regional glacial and meltwater system transporting sediments
from valleys in the Coast Mountains and more localized fluvial system depositing sediments source
from nearby mountain valleys. Data collected here indicates that the source sand/silt/clay of the
Dashwood Drift and the lower Cowichan Head Formation are similar in elemental concentration trends,
while the upper Cowichan Head Formation and the overlying Quadra Sand display a different trend
suggesting that the source for these sediments changed.
For the Quadra Sand the most distinguishing feature is the consistency of Sr concentrations throughout
the unit (including the upper Cowichan Head Formation) although there is a decrease in Sr
concentrations in the top few meters of the Quadra Sand in the Spider borehole. Sr concentrations are
slightly higher in the upper Cowichan Head Formation of the Spider borehole compared to the Quadra
Sand. Fe and Ti concentrations are lower than expected given that Crow et al., (2014) report elevated
and highly varying magnetic susceptibility values indicating a moderate to significant magnetic mineral
content. As suggested by Trettin (2004) and observed here through principal component analyses there
is a difference in multiple elemental associations and abundances between the lower and upper Quadra
Sand.
The Vashon Drift and Capilano Sediments can be differentiated from each other, and the underlying
Quadra Sand, by variations in the concentration of elements such as Ba, Fe, and Mn.
To our knowledge these results are the first systematic geochemical characterization of late Pleistocene
succession in the Nanaimo Lowlands. The Cochrane and Spider borehole data demonstrates that pXRF
geochemistry is valuable in differentiating stratigraphic units and can provide insight into the
provenance of these sediments.
27
Figure 19. For the Spider borehole Cr and Sr concentrations for the Dashwood Drift and Lower
Cowichan Formation display similar trends that are markedly different from the Upper
Cowichan Formation and Quadra Sand.
28
6.0 Acknowledgments
Special thanks to Barbara Medioli who helped with the graphical core, sample depth calculations and for
her review of the text along with Jan Bednarski. Thanks to Don Cummings for logging the core.
Support from drill site geologists Daniel Paradis and Jan Bednarski is much appreciated. This data was
collected as part of the Nanaimo Lowlands Aquifer Activity, a collaborative project of the Regional
District of Nanaimo and the Aquifer Assessments and Support to Mapping - Groundwater Inventory
Project of the Groundwater Geoscience Program, Geological Survey of Canada, Natural Resources
Canada.
7.0 Referencess
Alley, N.F., 1979. Middle Wisconsin Sstratigraphy and climatic reconstruction, souther Vancouver
Island, British Columbia. Quaternary Research, v. 11, p. 213-237.
Armstrong, J.E. and Clague, J.J., 1977. Two major Wisconsin lithostratigraphic units in
southwestern British Columbia. Canadian Journal of Earth Sciences, v. 14, n. 7, p. 1471-1480.
Bednarski, J., 2014. Surficial geology and Pleistocence stratigraphy from Deep Bay to Naoose
Harbour, Vancouver Island, British Columbia, Geological Survey of Canada, Open File 7681.
Clague, J.J., 1977. Quadra Sand: a study of Late Pleistocene geology and geomorphic history of
coastal southwest British Columbia. Geological Survey of Canada, Paper 77-17, 24 p.
Clague, J.J., 1976. Quadra Sand and its relation to the late Wisconsin glaciation of southwest
British Columbia. Canadian Journal of Earth Sciences, v. 13, p. 803-815.
Clague, J.J., Froese, D., Hutchinson, I., James, T.S., and Simon, K.M., 2005. Early growth of the
last Cordilleran ice sheet deduced from glacio-isostatic depression in southwest British Columbia,
Canada. Quaternary Research, v. 63, p. 5359.
Crow, H.L., Knight, R.D., Russell, H.A.J., Pugin, A.J.-M., and Cartwright, T.J., 2014. Downhole
geophysical data from five boreholes in the Nanaimo Lowlands, British Columbia; Geological
Survey of Canada, Open File 7567, 1 zip file. doi:10.4095/294925.
Crow, H L; Knight, R D; Medioli, B E; Hinton, M J; Plourde, A; Pugin, A J -M; Brewer, K D;
Russell, H A J; Sharpe, D R; Geological Survey of Canada, Open File 7079, 2012; 31 pages.
doi:10.4095/291486
Dawson, G.M., 1877-1878 . Report of explorations in the southern portion of British Columbia.
Report of the Geological Survey of Canada, Part B, p. 105-133.
Fyles, J.G., 1963. Surficial geology of Horne Lake and Parksville map-areas, Vancouver Island,
British Columbia. Geological Survey of Canada, Memior 318. 142 p.
29
Gabrielse, H. and Yorath C.J.,1992. Introduction, Chapter 1in Geology of the Cordilleran Orogen
in Canada, Gabrielse, H. and Yorath C.J. (editors), Geological Survey of Canada, n. 4, p. 3-11.
Grunsky, E.C., 2010. The interpretation of geochemical survey data. Geochemistry:Exploration,
Environment, Analysis, v.10, p. 27-74.
Grunsky, E.C., 2001. A program for computing rq-mode principal components analysis for S-Plus
and R. Computers & Geosciences, v. 27, p. 229235.
Hastie, T. J., 1992. Generalized additive models. Chapter 7 of Statistical Models in S eds J. M.
Chambers and T. J. Hastie, Wadsworth & Brooks/Cole.
Hicock, S.R. and Armstrong, J.E. 1985.Vashon Drift: definition for the formation in the
GeorgiaDepression, southwest British Columbia. Canadian Journal of Earth Sciences,v. 22, p. 748-
757.
Hicock, S.R. and Armstrong, J.E. 1983. Four Pleistocene formations in southwest British
Columbia: their implications for patterns of sedimentation of possible Sangamonian to early
Wisconsinan age. Canadian Journal of Earth Sciences, v. 20, p. 1232-1247.
Hicock, S.R. and Armstrong, J.E. 1981. Coquitlam Drift: a pre-Vashon Fraser glacial formation in
the Fraser Lowland, British Columbia. Canadian Journal of Earth Sciences, v. 18, p. 1443-1451.
Knight, R.D., Moroz, M., and Russell, H.J.A. 2012. Geochemistry of a Champlain Sea aquitard,
Kinburn, Ontario: portable XRF analysis and fusion chemistry; Geological Survey of Canada, Open
File 7085, 1 CD-ROM doi:10.4095/290969
Knight, R.D., Kjarsgaard, B.A., Plourde, A.P., and Moroz, M., 2013. Portable XRF spectrometry of
standard reference materials with respect to precision, accuracy, instrument drift, dwell time
optimization, and calibration; Geological Survey of Canada, Open File 7358. doi:10.4095/292677
McMartin, I and McClenaghan, M.B., 2001. Till geochemistry and sampling techniques in glaciated
shield terrain: a review. In Drift Exploration in Glaciated Terrains, ed. by McClenaghan, M.B.,
Bobrowsky, P.T., Hall, G.E.M., and Cook, S.J. Geological Society of London, Special publication
185, p. 19-43.
Morris, P.A., 2009. Field-portable X-ray fluorescence analysis and its application in QSWA.
Geological Survey of Western Australia, Record 2009/7, 23 p.
Mustard, P.S., 1994. The Upper Cretaceous Nanaimo Group, Georgia Basin, in J.W.H. Monger, ed.,
Geological Hazards of the Vancouver Region, Southwestern British Columbia: Geological Survey
of Canada, Bulletin 481, p. 27-95.
30
Plourde, A.P., Knight, R.D., and Russell, H.A.J., 2012. Portable XRF spectrometry of insitu and
processed glacial sediment from a borehole within the Spiritwood buried valley, southwest
Manitoba; Geological Survey of Canada, Open File 7262. doi:10.4095/291922
Pugin, A. J.-M., Pullan, S. E., and Hunter, J. A., 2009. Multi-component high resolution seismic
reflection profiling; The Leading Edge, v. 28, p. 1248-1261936-945
Radu, T. and Diamond, D., 2009. Comparison of soil pollution concentrations determined using
AAS and portable XRF techniques. Journal of Hazardous Materials, v. 171, p. 1168-1171.
Rock, N.M.S., 1988. Numerical geology. Springer Verlag, Berlin, p. 427.
Stanley, G., Gallagher, V., NiMhairtin, F., Brogan, J., Lally, P., Doyle, E., and Farrell, L., 2009.
Historic mines site inventory and risk classification. Geochemical characterization and
environmental matters, Geological Survey of Ireland v.1, Appendix 4, p.40.
Thermo Scientific, 2008. Elemental limits of detection in Si02 and SRM matrices using soil
analysis. Niton XL3t GOLDD + Series x-ray fluorescence analyzer spec sheet.
Trettin, H.P., 2004. Wisconsinan stratigraphy at northern margin of Strait of Georgia, southern
Cortes Island and vicinity, British Columbia; Geological Survey of Canada, Current Research 2004-
A4, 11 p.
Whitbread, M.A.I., 2004. Lithogeochemical alteration around the Century and elura Zn-Pb-Ag
deposits: Detecting alteration expressions in deep and near surface environemnets. PhD. Thesis,
University of Canberra, Australia, p.410.
Zhou, D., Chang, T., and Davis, J.C., 1983. Dual extraction of R-mode and Q-mode factor
solutions. Mathematical Geology, v. 15, p. 581606.
Zhu, Y., Weindorf, D.C., and Zhang, W., 2011. Characterizing soils using a portable X-ray
fluorescence spectrometer: 1. Soil texture. Geoderma, v. 167-168, p. 167-177.
... The gravel is a mixture of plutonic rocks derived from the Coast Mountains and volcanic and sedimentary rocks from the east coast of Vancouver Island. The geochemical signature of Dashwood has an elevated concentration of metals compared to Quadra Sand (Knight et al., 2015). Dashwood Drift was deposited during the penultimate glaciation early Wisconsinan, as deduced from the presence of marine shells which date beyond the range of radiocarbon in overlying sediments (Clague, 1980). ...
... It consists of horizontally and crossstratified, well-sorted sand with minor silt and gravel with wood and peat lenses in its lower portions. Geochemically the Quadra is distinguished from other units by the Sr concentrations throughout the unit (including the upper Cowichan Head Formation; Knight et al., 2015). Quadra is interpreted to be outwash deposited during the transition from non-glacial to glacial conditions at the onset of the Fraser Glaciation (Armstrong and Clague, 1977;Clague, 1977Clague, , 1976. ...
... Vashon Drift also includes sand and gravel in eskers, kame terraces, and ice-contact fans and deltas deposited by meltwater streams during early deglaciation. The geochemical signature of Vashon has a number of complicated trends over the unit thickness related to various element groupings (Knight et al., 2015). The streams initially would have flowed along the margins of the retreating Cordilleran and local glaciers, with some deltas or terraces building into short-lived ice-dammed lakes. ...
... Mining Mode uses Fundamental Parameters which is recommended for elements expected to exceed >1% concentration. In order to honor the protocol used for previous borehole studies (Knight et al., 2015a, Knight et al., 2015b, a dwell time of 60 seconds was used for each filter (Main, Low, and High) for a total of 180 seconds per analysis for soil mode. For mining mode a dwell time of 45 seconds was used for each filter (Main, Low, High, and Light) for a total of 180 seconds per analysis. ...
Book
This book describes an array of power tools for data analysis that are based on nonparametric regression and smoothing techniques. These methods relax the linear assumption of many standard models and allow analysts to uncover structure in the data that might otherwise have been missed. While McCullagh and Nelder's Generalized Linear Models shows how to extend the usual linear methodology to cover analysis of a range of data types, Generalized Additive Models enhances this methodology even further by incorporating the flexibility of nonparametric regression. Clear prose, exercises in each chapter, and case studies enhance this popular text.
Article
The development of new technologies for use in field soil survey has produced powerful new quantitative tools for assessing soil physicochemical properties in-situ. One such technology, portable X-ray fluorescence (PXRF) spectrometry, has shown considerable promise in evaluating elemental concentrations in soils for a wide variety of applications. Less research is available on how PXRF can be applied to quantify soil physical properties (e.g. soil texture). This study evaluated the feasibility of predicting soil clay and sand contents from PXRF data on 584 soil samples collected from highly diverse regions of Louisiana and northeastern New Mexico (Capulin Volcano National Monument), USA. An Innov-X Delta Premium PXRF was used to sequentially scan soil samples under both field and laboratory conditions and assess 15 elements (K, Ca, Ti, Cr, Mn, Fe, Co, Cu, Zn, As, Rb, Sr, Zr, Ba, and Pb). Elemental concentrations were then related to soil textural data processed through traditional laboratory methods with multiple linear regression. Among all elements evaluated, Fe and Rb showed particular significance for soil textural prediction. The regression models for sand and clay contents of both study sites were strongly correlated to soil textural separates with PXRF-soil texture R2 of 0.862, 0.959, 0.892, and 0.780 for Louisiana sand and clay, and Capulin sand and clay, respectively. Independent validation sub-datasets confirmed that PXRF readings can be used to estimate soil textural parameters with high R2 values of 0.854, 0.682, 0.975, 0.891, 0.875, and 0.876, and low RMSE values of 5.53%, 5.92%, 2.68%, 6.26%, 5.43%, and 2.66%, for Louisiana sand, silt, and clay and Capulin sand, silt, and clay, respectively. The RMSE values of clay are substantially lower than those reported in previous studies using other proximal sensing techniques. While regional differences may require localized standardization of the PXRF with a unique combination of elements germane to the study area, PXRF shows considerable promise as a technique for rapidly assessing soil textural separates in-situ.
Article
It is mathematically possible to extract both R-mode and Q-mode factors simultaneously (RQ-mode factor analysis)by invoking the Eckhart-Young theorem. The resulting factors will be expressed in measures determined by the form of the scalings that have been applied to the original data matrix. Unless the measures for both solutions are meaningful for the problem at hand, the factor results may be misleading or uninterpretable. Correspondence analysis uses a symmetrical scaling of both rows and columns to achieve measures of proportional similarity between objects and variables. In the literature, the resulting similarity is a χ 2distance appropriate for analysis of enumerated data, the original application of correspondence analysis. Justification for the use of this measure with interval or ratio data is unconvincing, but a minor modification of the scaling procedure yields the profile similarity, which is an appropriate measure. Symmetrical scaling of rows and columns is unnecessary for RQ-mode factor analysis. If the data are scaled so the minor product W'Wis the correlation matrix, the major product WW'is expressed in the Euclidean distances between objects. Therefore, RQ-mode factor analysis can be performed so that the Rmode is a principal components solution and the Qmode is a principal coordinates solution. For applications where the magnitudes of differences are important, this approach will yield more interpretable results than will correspondence analysis.
Article
Geochemical data are generally derived from government and industry geochemical surveys that cover areas at various spatial resolutions. These survey data are difficult to assemble and integrate due to their heterogeneous mixture of media, size fractions, methods of digestion and analytical instrumentation. These assembled sets of data often contain thousands of observations with as many as 50 or more elements. Although the assembly of these data is a challenge, the resulting integrated datasets provide an opportunity to discover a wide range of geochemical processes that are associated with underlying geology, alteration, landscape modification, weathering and mineralization. The use of data analysis and statistical visualization methods, combined with geographical information systems, provides an effective environment for process identification and pattern discovery in these large sets of data. Modern methods of evaluating data for associations, structures and patterns are grouped under the term 'data mining'. Mining data includes the application of multivariate data analysis and statistical techniques, combined with geographical information systems, and can significantly assist the task of data interpretation and subsequent model building. Geochemical data require special handling when measures of association are required. Because of its compositional nature logratios are required to eliminate the effects of closure on geochemical data. Exploratory multivariate methods include: scatterplot matrices (SPLOM), adjusting for censored and missing data, detecting atypical observations, computing robust means, correlations and covariances, principal component analysis, cluster analysis and knowledge based indices of association. Modelled multivariate methods include discriminant analysis, analysis of variance, classification and regression trees neural networks and related techniques. Many of these topics are covered with examples to demonstrate their application.
Article
Two lithostratigraphic units, Quadra Sand and the Cowichan Head Formation, are overlain by Vashon till and associated glacial sediments and underlain by Dashwood and Semiahmoo drift deposits in coastal southwest British Columbia. Each unit is formally described and stratotypes are presented. Quadra Sand consists of cross-stratified, well-sorted sand, minor gravel, and silt deposited as outwash in front of glaciers advancing into the Georgia Depression at the beginning of the Fraser Glaciation. It is diachronous, deposition having commenced earlier than 29,000 years BP at the north end of the Georgia Depression but not until after 15,000 years BP at the south end of the Puget Lowland. The Cowichan Head Formation, deposited during the Olympia nonglacial interval, underlies Quadra Sand and consists of parallel-bedded silt, sand, and gravel, in part plant-bearing. The unit is divisible into a lower marine member and an upper fluvial and estuarine member.
Article
Multicomponent seismic reflection methods are a new tool for oil and gas exploration and reservoir monitoring (Miles 1988), but such technologies have not yet been extensively exploited for near-surface exploration related to hydrogeological and/or geotechnical investigations. With the advantage of relatively inexpensive recording systems for near-surface applications, we show that the use of multicomponent high-resolution seismic reflection methods has great potential as a new means of observing and characterizing the physical parameters of the shallow subsurface, and in particular of groundwater reservoirs.
Article
Vashon Drift was deposited during the Fraser Glaciation (late Wisconsinan) at the time of maximum expansion of the southwestern part of the Cordilleran ice sheet when it filled the Georgia Depression about 14 500 years ago. The drift is present throughout the depression and comprises till and glaciofluvial and glaciolacustrine sediments derived from source areas surrounding the coastal trough. It is overlain by Capilano Sediments and underlain by Quadra Sand, also of Fraser age. Drift deposition was diachronous and complex, probably caused by alpine glaciers coalescing in the trough with the ice margin repeatedly grounding and floating in seawater. Studies of bedrock striae, till fabrics, and clast provenance reveal that Vashon ice movement was generally southward, although locally controlled by topography. A time–space diagram is presented that confirms the long-held hypothesis that advance and decay of Vashon ice were rapid.
Article
Semiahmoo and Dashwood drifts were deposited during the penultimate glaciation, and Highbury Sediments and Muir Point Formation during the last interglacial. These sediments are defined and described, and stratotypes are established for them.Based on arboreal pollen assemblages, lithologic similarity, and relative stratigraphic position, Semiahmoo Drift of the Fraser Lowland is tentatively correlated with Dashwood Drift of Vancouver Island. Stone provenance, till fabrics, glaciotectonic structures, cross-bedding, and stone imbrication indicate that regional ice and meltwater flowed out of major fiords in the Coast Mountains and into the ancestral Strait of Georgia during the penultimate glaciation. Ice flowed southeastward against the east coast of Vancouver Island, and down the axis of the strait. Pollen and molluscs in Semiahmoo and Dashwood fossiliferous muds give evidence of transition from glaciomarine to marine conditions at the close of the penultimate glaciation.Highbury Sediments of the Fraser Lowland perhaps correlate with the Muir Point Formation of Vancouver Island, based on lithologic similarity and relative stratigraphic position. Stone provenance and paleocurrent data indicate that stream flow during the last interglacial issued from nearby mountains and onto coastal floodplains that shelved into the ancestral Strait of Georgia and Juan de Fuca Strait. Abundant Pseudotsuga and Tsuga heterophylla pollen indicate that climate during that time was at least as warm as the present.
Article
Coquitlam Drift is formally defined and stratotypes established for it in the Coquitlam – Port Moody area, B.C. It is a Pleistocene formation consisting of till, glaciofluvial, ice-contact, and glaciomarine sediments deposited between 21 700 and 18 700 years BP, during the Fraser Glaciation (late Wisconsin) and prior to the main Vashon glacial maximum at about 14 500 years BP. The drift was deposited in short pulses by valley and piedmont glaciers fluctuating into the Fraser Lowland from the Coast Mountains to the north and Cascade Mountains to the east.