JOURNAL OF QUATERNARY SCIENCE (2014) 29(3) 249–260
ISSN 0267-8179. DOI: 10.1002/jqs.2697
Evaluating diatom-derived Holocene pH
reconstructions for Arctic lakes using an expanded
171-lake training set
SARAH A. FINKELSTEIN,1* JOAN BUNBURY,1† KONRAD GAJEWSKI,2 ALEXANDER P. WOLFE,3 JENNIFER K. ADAMS1
and JANE E. DEVLIN1
1
Department of Geography and Department of Earth Sciences, University of Toronto, 22 Russell Street, Toronto, Canada M5S
3B1
2
Department of Geography, University of Ottawa, Ottawa, Canada
3
Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Canada
Received 14 October 2013; Revised 11 February 2014; Accepted 14 February 2014
ABSTRACT: Inference models from diatoms preserved in lake sediments can be used to reconstruct long-term pH
changes to better understand the process of lake ontogeny. An expanded diatom training set was developed using
taxonomically harmonized modern assemblages in surface sediments of 171 lakes spanning a variety of geological
and climatic settings across the Canadian Arctic. Lake-water pH emerged as a significant variable and the most
influential in structuring diatom assemblages. The resulting two-component weighted-averaging partial least
squares pH inference model performs strongly, even after identifying effects of spatial autocorrelation at distances
<20 km. The model was then applied to three dated Holocene diatom stratigraphies from Arctic regions of
contrasting bedrock geology and buffering capacity, and the significance of the pH reconstructions was assessed.
At Lake CF3 in a poorly buffered catchment, a gradual but significant pH decline begins 5000 years after lake
inception, coincident with regional Late Holocene cooling. Reconstructions for two well-buffered, more alkaline
sites were not significant, probably due to poor analogues and other factors driving changes in diatom assemblages.
Due to sparse soil and vegetation in these and other Arctic basins, bedrock composition is the most important
regulator of Holocene pH, and only in poorly buffered lakes does pH primarily represent a climate signal.
Copyright # 2014 John Wiley & Sons, Ltd.
KEYWORDS: diatoms; freshwater; paleolimnology; spatial autocorrelation; transfer function.
Introduction
Lake-water pH is a major factor explaining the distribution of
freshwater diatom taxa (Meriläinen, 1967; Battarbee
et al., 2010). One of the earliest ecological classifications of
diatoms was that of Hustedt, who categorized diatoms into
five pH tolerance groupings (Hustedt, 1937–1939; Battarbee
et al., 1986). Because of the close relationship between
diatom assemblages and pH, diatoms preserved in sediment
records are robust indicators for pH over time. This relationship proved crucial in providing evidence for the role of
human activity in 20th-century lake acidification (Battarbee,
1984). Subsequently, canonical ordinations of diatom relative
frequencies and environmental variables from a variety of
Arctic regions have shown that, while other factors may be
significant, pH is fundamental in explaining the distribution
of diatom taxa in high-latitude lakes (Joynt and Wolfe, 2001;
Antoniades et al., 2005; Keatley et al., 2008; Hadley et al.,
2013). The mechanisms by which diatom taxa differ
regarding pH optima probably originate from the effects of
external pH on diatom physiology, including cell membrane
permeability, intracellular pH homeostasis, rates of metal,
silicic acid and nutrient uptake, carbon fixation and silicon
bio-mineralization (Hervé et al., 2012; Lavoie et al., 2012).
Quantitative pH reconstructions from diatoms in sediment
cores are generated using numerical models which relate
modern diatom abundances to corresponding lake-water pH
(Birks et al., 1990; Birks and Simpson, 2013). Comparisons of
diatom-inferred pH with the outputs of geochemical models
predicting lake-water pH from physical factors often show
Correspondence: S. A. Finkelstein, as above.
E-mail: finkelstein@es.utoronto.ca
†
Present address: Department of Geography and Earth Science, University of
Wisconsin, La Crosse, La Crosse, WI 54601, USA.
Copyright # 2014 John Wiley & Sons, Ltd.
good agreement (Erlandsson et al., 2008), providing further
support for the utility of diatoms in tracking pH changes over
time. While quantitative reconstructions of some environmental variables using diatom assemblages are problematic when
the variables in question are not the primary factors directly
affecting diatom physiology, close scrutiny of diatom–pH
relationships across a variety of regions continues to show
that pH is a consistently important primary control on diatom
assemblages (Juggins, 2013).
Building upon the diatom–pH calibration models developed during the 1980s in response to the acid rain controversy, diatom-based paleolimnology has been used to test
hypotheses for the drivers of natural change in lake-water pH,
including post-glacial landscape evolution, lake ontogeny or
climate (Battarbee et al., 2010). Many of the available
diatom-inferred high-resolution Holocene pH reconstructions
for poorly buffered lakes in forested catchments show a
gradual decrease in pH from the Early Holocene to present
day (Renberg and Hellberg, 1982; Renberg, 1990). This longterm decline may be related to increased fluxes of humic
acids to the lake with vegetation succession and soil development (Whitehead et al., 1989), as well as the depletion over
time of base cations due to chemical weathering of catchment soils (Boyle, 2007; Boyle et al., 2013). Climate is also
hypothesized to drive long-term pH change, both indirectly
through changes in vegetation cover and rates of weathering,
and through direct effects on physical limnology and on
metabolic processes of lake biota (Psenner and Schmidt,
1992; Sommaruga-Wögrath et al., 1997).
While many diatom-inferred pH reconstructions have been
produced from poorly buffered lakes in the forested temperate
zone, fewer pH reconstructions exist from Arctic lakes
spanning varying regimes with respect to buffering capacity.
In tundra and polar desert regions with sparse to minimal
250
JOURNAL OF QUATERNARY SCIENCE
vegetation cover, limited soil development and low rates of
chemical weathering, climate may play an enhanced role in
the regulation of lake-water pH, through the dynamics of lake
ice and primary production and their effects on pCO2.
Prolonged ice cover prevents mixing with the atmosphere
and promotes super-saturation of CO2 in the water column,
and at the same time lowers primary production, both
promoting a decrease in pH (Koinig et al., 1998; Wolfe,
2002).
The goal of the present study was to generate Holocene pH
reconstructions from three Arctic lakes using a newly compiled training set containing 171 lakes spanning a wide pH
gradient. The expanded training set contains a greater range
of taxa relative to local training sets, and thus a larger
proportion of the measured pH gradient is occupied by the
taxa included, allowing for refined estimates of optima and
tolerances with respect to pH (among other significant
variables), and ultimately more skillful pH reconstructions. In
addition to the conventional measures of bootstrapped r2 and
root mean squared error of prediction (RMSEP), the strength
of the diatom–pH relationship across the training set is also
assessed using criteria for determining the effect of spatial
autocorrelation, which may undermine the validity of inference models (Telford and Birks, 2005, 2009).
The three lakes selected for pH reconstruction have
contrasting geological settings, and consequently vary regarding their present-day pH and buffering capacity, resulting in
markedly different diatom assemblages. The statistical significance of the resulting Holocene pH reconstructions is
evaluated, following Telford and Birks (2011), followed by
interpretations of the reconstructions in terms of Holocene
paleoenvironments. Comparing reconstructions from three
dissimilar lakes provides insight into whether Holocene pH
changes are caused by millennial-scale climate evolution, or
whether lake pH is determined more proximally by edaphic,
watershed-scale processes.
Materials and methods
Expanded modern diatom pH training set
A variety of regional pH training sets for freshwater diatoms
in the Canadian Arctic exist (Table 1), but difficulties
surrounding taxonomic harmonization have prevented the
collation of a single database to examine diatom–pH relationships across the full length of the Arctic pH gradient. To
mitigate the impact of taxonomic irregularity in the collation
of a larger training set, we selected only diatom assemblage
data and environmental variables from sites analysed within
our research groups, resulting in a series of 171 lake sites
across the Canadian Arctic (Fig. 1). Data from 62 lakes across
the Arctic islands (Bouchard et al., 2004), 61 lakes on Baffin
Island (Joynt and Wolfe, 2001), 20 lakes on Bylot Island and
the Qorbignaluk Headland on northern Baffin Island (Devlin
and Finkelstein, 2011), two lakes on Melville Peninsula
(Adams and Finkelstein, 2010) and one lake on Melville
Island (Stewart and Lamoureux, 2012) have been previously
published. The remaining 25 unpublished records were
analysed using identical protocols. The lake population spans
broad temperature and precipitation gradients, and has lakewater pH values ranging between 4.7 and 8.6 (Figs 1 and 2).
This pH gradient captures the range of lake-water pH values
across the Canadian Arctic (Hamilton et al., 2001).
For taxonomic consistency, taxon lists and authorities were
carefully vetted, aided by light micrographs and notes. Only
identified taxa were included in the compilation. Similarly,
undifferentiated species within a named genus (e.g. Achnanthes sp.) were not included. Only diatom taxa present in
three or more lakes and with an abundance 1% in at least
one lake were included in statistical analysis; these eliminations reduced the number of taxa from 483 to 261 (supplementary Table S1). Taxonomic harmonization involved crosschecking for synonymies and where necessary updating
diatom nomenclature. The genus Aulacoseira is abundant
and diverse in many lakes, and at times species are difficult
to differentiate under light microscopy. Joynt and Wolfe
(2001) differentiated A. distans and A. lirata (and fo. biseratia)
but other analysts did not do so consistently, so these taxa
were grouped as ‘A. distans-type’. Although A. distans was
described as a pre-Quaternary fossil that may not be present
in North America using the formal concept of the species
(Crawford and Likhoshway, 1999), virtually identical forms
are widely reported from dilute lakes in North America
(Camburn and Charles, 2000), including sites in Arctic
Canada (Wilson et al., 2012). We recognize four Aulacoseira
taxa in the dataset (A. alpigena, A. distans, A. italica, A.
perglabra; Table S1) but acknowledge the need for further
taxonomic work using scanning electron microscopy. Several
additional problematic taxa, such as Sellaphora pupula and
Fragilaria capucina, were identified by some analysts as
having one or more sub-specific forms. However, other
analysts did not consistently differentiate these forms and
varieties, and therefore they were merged into respective
species complexes. Taxon names were verified using the
Index Nominum Algarum (Silva, 2012) and the California
Academy of Sciences Catalogue of Diatom Names (Fourtanier
and Kociolek, 2012).
To check for possible taxonomic inconsistencies between
the five analysts, we compared both the maximum abundances of the most common diatom taxa in terms of number of
occurrences in the 171-lake population, and the most
abundant in terms of maximum abundance at any one site
(Table 2). The distribution of diatom taxa is not expected to
be the same between each analyst due to regional limnological differences. The sites analysed by Bouchard et al. (2004)
and Adams and Finkelstein (2010 and unpublished) are
Table 1. Comparative performance of high-latitude diatom-based freshwater pH inference models: cross-validated r2, root mean squared error of
prediction (RMSEP) and number of lakes for this study and available Arctic diatom pH training sets.
r2
0.78
0.77
0.77
0.69
0.61
0.59
0.28
RMSEP
No. of lakes
Region
Reference
0.35
0.19
0.40
0.40
0.30
0.43
0.46
168
99
64
26
53
61
30
Canadian Arctic
Swedish Lapland
Canadian High Arctic
Canadian High Arctic
Northern Sweden
Baffin Island
Canadian High Arctic
This study
Bigler and Hall (2002)
Antoniades et al. (2005)
Antoniades et al. (2004)
Rosén et al. (2000)
Joynt and Wolfe (2001)
Michelutti et al. (2006)
Copyright # 2014 John Wiley & Sons, Ltd.
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
DIATOM-DERIVED HOLOCENE PH RECONSTRUCTIONS
251
Figure 1. (a) Locations of the three lakes for
which Holocene pH reconstructions were
undertaken, shown in relation to mean July
air temperature. (b) Locations of lakes for the
diatom–pH training set in relation to mean
annual air temperature. (c) pH of training set
lakes in relation to total annual precipitation.
Climate data (30-year means) were extracted
from New et al. (2002). This figure is
available in colour online at wileyonlinelibrary.com.
Copyright # 2014 John Wiley & Sons, Ltd.
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
252
JOURNAL OF QUATERNARY SCIENCE
25
No. of Lakes
20
15
10
5
0
4.5
5.5
6.5
7.5
8.5
pH
Figure 2. Frequency distribution of lake-water pH in the 171-lake
training set.
mostly alkaline lakes in the central Arctic (Table S2), and
alkaliphilous taxa, such as Amphora pediculus or Denticula
kuetzingii, are best represented at those sites. Sites analysed
by Joynt and Wolfe (2001) and Devlin and Finkelstein (2011)
are mostly dilute, more acidic lakes on Baffin and Bylot
Islands, with high percentages of Aulacoseira distans-type.
Keeping in mind these differences, there is broad representation of the most common taxa among analysts. For example,
Achnanthidium minutissimum appears with maximum abundance of 20–40%, and Staurosira construens var. venter
reaches >50% abundance at sites analysed by four of the
analysts. An ordination of the diatom data with detrended
correspondence analysis (DCA) was also performed to show
the distribution of sites according to analyst (Fig. 3). The
ordination differentiates sites primarily by environmental
differences, namely well-buffered versus poorly buffered
waters as a reflection of bedrock geology, and regional
climate, and portrays a certain degree of overlap between
analysts, suggesting taxonomic integrity of the amalgamated
counts.
The 171-lake database includes eight consistently available
environmental variables (Table 3; Table S2). Using a Geographic Information System (GIS), mean July air temperature
(TJul), mean annual air temperature (MAAT), and total annual
precipitation (PAnn) were extracted for each site from a
gridded global climate dataset of 30-year means (New
et al., 2002) (Fig. 1). If not previously calculated, distance
from coast (DFC) and surface area (SfcA) of the lakes were
obtained from 1 : 50 000 or 1 : 250 000 topographic maps.
In situ field measurements of water depth, pH and conductivity are included; in the few cases (n ¼ 2 for pH; n ¼ 4 for
conductivity) when field water quality measurements were
not available, laboratory measurements were used. Although
nutrient, dissolved organic carbon (DOC) and ion concentrations are also important in characterizing diatom distribution, these variables were not consistently available for all
sites, or they were measured at different times of the year,
sometimes before the start of the growing season. Earlyseason sampling can result in values at or below detection
that are not reflective of the main growth phase of diatoms.
Given the overarching significance of pH in explaining
diatom distributions (Hustedt, 1937–1939; Meriläinen, 1967;
Battarbee et al., 1986, 2010), the diatom–pH relationship
remains the focus of the present study; the collation of the
database with eight environmental variables, however,
presents an opportunity for future analyses of diatom–climate
relationships and diatom biogeography.
Model development for pH inference
Relative abundances of diatoms were square-root transformed
to stabilize variances and rare species were down-weighted
in the ordinations. DCA (Hill and Gauch, 1980) was used to
ordinate the samples as a function of species, and to
determine the maximum amount of variation in the diatom
data, represented by gradient lengths. Canonical correspondence analysis (CCA, ter Braak, 1986) was used to show the
relationship between diatom assemblages and available
environmental variables. Exploratory ordinations were also
run using non-metric multidimensional scaling (NMDS) to
ensure that DCA and CCA ordinations were not overly
sensitive to rare species. Because NMDS ordinations showed
similar results, they will not be discussed further. Each
variable was assessed individually in the CCA to determine if
it was significant in explaining the diatom distributions,
followed by forward selection to corroborate our hypothesis
that pH is an important variable in explaining the distribution
of diatom species. Partial CCA, constrained to pH, was used
to show the independent effect of pH on explaining the
variability in diatom assemblages.
Table 2. Maximum relative frequencies (%) reported for the 14 most abundant taxa in the training set, for five different analysts.
Analyst (and number of sites)
Taxon
Achnanthidium minutissimum
Amphora copulata
Amphora pediculus
Aulacoseira distans-type
Cyclotella pseudostelligera
Denticula kuetzingii
Nitzschia perminuta
Psammothidium helveticum
Psammothidium marginulatum
Pseudostaurosira brevistriata
Pseudostaurosira pseudoconstruens
Staurosirella pinnata
Staurosira construens var. venter
Tabellaria flocculosa
GB (62)
EJ (61)
JA (27)
JD (20)
KS (1)
20.2
55.1
23.7
0.2
0.0
46.5
8.1
2.4
33.2
51.7
49.8
69.4
58.0
3.8
23.8
0.0
0.0
92.3
0.0
0.0
5.2
19.4
12.1
10.6
30.7
8.1
54.3
8.8
41.7
16.1
64.8
48.0
20.2
51.1
20.1
3.6
10.4
6.7
14.5
31.6
46.8
6.2
36.4
0.7
0.0
57.0
0.0
0.0
4.9
11.2
56.3
0.2
0.0
85.5
76.6
14.7
2.9
0.0
0.0
0.0
66.2
0.0
1.0
0.0
0.7
0.0
0.0
0.2
0.0
0.0
GB, Giselle Bouchard (Bouchard et al., 2004); EJ, Ernest Joynt III (Joynt and Wolfe, 2001); JA, Jennifer Adams, 25 unpublished, two reported in
Adams and Finkelstein (2010); JD, Jane Devlin (Devlin and Finkelstein, 2011); KS, Kailey Stewart (Stewart and Lamoureux, 2012).
Copyright # 2014 John Wiley & Sons, Ltd.
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
253
3
DIATOM-DERIVED HOLOCENE PH RECONSTRUCTIONS
a)
0
5
6
9
18
20
21
22
23
24
25
27
40
51
52
53
73
79
82
84
104
126
127
128
132
133
143
146
148
149
151
154
155
156
b)
159
163
168
172
178
181
184
192
214
217
226
236
247
249
252
253
255
131 167
15
78
105
115
130
157
166
175
183
221
227
229
232
240
114 118
203
30 56
34
208
112 117
177 1
5 124
176
129 244 223
109 225
228 89 220
138
209
3
55
211 31
16
169
145 63
206
160 216
96
107
251
261
101
108
14
95 85 196
182
222
219 152
173
141
42
32
10
59
12 210
224
121
170 198
26
29
204 94
111
140
161 246
213
190 11 80
38
242
39
250
13 134
93
171 58 158
36
86
44103
147
122
165 17 200
67 218 199 254
113
195
98
81 248
47
49
188
150
164 189 235
19
69
87186
233
191
35
2 243
33
99
257
64
174 116 54
187
75
237 8
77
215
62
28 76
90
238 125
234
201
65
43 144
83
197
162 102 92
179 88
45 13737
119
61205
139
50 180
153 245
239
66 185
194 60 241
71 100
212
258 120
68 193
97
48
70 230
46 4
7
74 231
57
259
142
256
72
91
135 41
136
202
11.9%
260
-2
5.1%
Figure 3. Detrended correspondence analysis
(DCA) plots for the 171 samples displayed by
(a) analyst and (b) the 261 diatom taxa. Taxon
names correspond to the numbering scheme
in Table S1. Loops are used as an aid to
labelling the data points.
0
5
DCA axis 2 λ = 0.30
Adams and Finkelstein
Bouchard et al.
Devlin and Finkelstein
Joynt and Wolfe
Stewart and Lamoureux
-2
6
DCA axis 1 λ = 0.71
As these analyses show that, of the available environmental
variables, pH explains the greatest amount of variance in
diatom distributions (see below), a pH inference model was
developed. Preliminary analyses aided in identifying three
sites as potential outliers regarding pH: JW58, JW60 and
PW01 (Fig. 2). These sites were retained in the ordination as
the analyses were unchanged with their inclusion. However,
they were removed before model development as they had a
Table 3. Summary of the eight environmental variables available for 171 Arctic lakes. CCA correlation coefficients are given for any
environmental variables with correlations >0.3 with CCA axis 1 or axis 2 (denoted with ).
Variable
Abbr.
Units
Mean
Median
Min.
Max.
SD
Mean July air temperature
Mean annual air temperature
Total annual precipitation
Distance from coast
Lake surface area
Water depth
Conductivity
pH
JulyT
MAAT
PAnn
DFC
SfcA
Depth
Cond
pH
˚C
˚C
mm
m
km2
m
mS cm1
4.7
–15.4
221
15 443
0.4014
9.7
73
7.2
4.6
–15.9
188
6500
0.15
6.5
25
7.1
1.2
–22.8
70
300
0.0001
0.5
0
4.7
7.4
–7.0
531
125 000
8.16
55.7
998
8.6
1.1
3.2
113
20 916
0.869
9.2
133
0.8
†
Transform
Log þ 1
Log þ 1
Log þ 1†
Log þ 1
Log þ 1
Correl.
coeff. (CCA)
0.68
0.80
0.39
0.35
0.75
0.86
Distribution remained skewed after transformation.
Copyright # 2014 John Wiley & Sons, Ltd.
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
254
JOURNAL OF QUATERNARY SCIENCE
clear influence on the residuals of preliminary models.
Weighted-averaging regression and calibration (WA) with
cross-validation (1000 iterations) was used to generate the
bootstrapped optimum and tolerance of each of the 261
species to pH. Weighted-averaging partial least-squares (WAPLS) models were assessed using bootstrap cross-validation
and compared based on the performance statistics.
To assess the effects of spatial autocorrelation on model
development, we used a variogram approach (Telford and
Birks, 2009). Coordinates of each lake were projected onto a
plane using an Albers Equal Area Conic projection to avoid
skewing associated with convergence of longitude lines at the
poles. We then evaluated the spatial structure in the environmental data using an empirical variogram (a graphical
representation of the variance of the differences in the
measured environmental variable among pairs of points, as a
function of the distance between the points), followed by
fitting the appropriate theoretical variogram model. The final
step in the autocorrelation analysis was to generate a spatially
structured variable with the same structure as the data in the
empirical variogram using unconditional Gaussian simulations
(1000 iterations) to determine if the pH data were spatially
autocorrelated and to what extent this structure may influence
transfer function development. WA and WA-PLS models were
developed using the package rioja (Juggins, 2012), statistical
significance of the reconstructions was determined using
palaeoSig (Telford, 2012) and spatial autocorrelation was
evaluated using gstat (Pebesma, 2004), all run in the R
environment (R Development Core Team, 2011).
Diatom-based Holocene pH reconstruction
To estimate pH changes over the Holocene with the expanded training set, we selected fossil diatom records spanning
the Holocene from three lakes from the Arctic but located in
different climatic and geochemical environments (Fig. 1a).
Lake CF3 is a poorly buffered lake (pH ¼ 5.9) on Baffin Island;
the lake is a bedrock basin within Archean crystalline rocks
of the Canadian Shield. Lake PW02 on Russell Island in the
Prince of Wales Island group is an alkaline lake (pH ¼ 7.9)
situated within the Paleozoic sedimentary bedrock of the Peel
Sound Formation, consisting locally of conglomerates of
mixed lithology, including sandstone, limestone, dolomite
and gneissic inclusions (Blackadar and Christie, 1963; Dyke
et al., 1992). Lake KR02 on Victoria Island is intermediate
between CF3 and PW02 regarding acidity, with a modern pH
of 7.4. Because it is located at the junction of different
geological formations, the local geology consists of both
clastic and carbonate sediments as well as basalt (Frisch and
Trettin, 1991). The cores recovered from Baffin, Russell and
Victoria Islands span the past 11 200, 9800 and 9900 years,
respectively (Lake CF3: Briner et al., 2006; Lake KR02:
Podritske and Gajewski, 2007; Lake PW02: Finkelstein and
Gajewski, 2008); the original age models were retained. The
newly developed two-component WA-PLS pH inference
model was applied to the fossil data to reconstruct pH at
each site. Sample-specific standard errors were generated via
bootstrap cross-validation procedures. Each of the reconstructions was then tested for statistical significance using the
method of Telford and Birks (2011).
Results
Modern diatom–pH relationships across the study
region
The 261 taxa included in the analyses vary in terms of
numbers of occurrences, maximum abundances and pH
Copyright # 2014 John Wiley & Sons, Ltd.
Table 4. Summary of DCA results of 261 species from 171 lakes,
and CCA results for the same array constrained alternately to eight
environmental variables and solely to lake-water pH.
Axis
Ordination results
DCA – 261 species
Eigenvalues (l)
Gradient length
Cumulative % variance of species data
CCA – 8 significant environmental variables
Eigenvalues (l)
Species–environment correlations
Cumulative % variance of species data
CCA – constrained to pH
Eigenvalue (l)
Species–environment correlation
Cumulative % variance of species data
1
2
3
0.71
4.7
11.9
0.30
2.9
17.0
0.22
3.5
20.6
0.61
0.93
10.1
0.18
0.77
13.1
0.13
0.70
15.4
0.53
0.89
9.0
0.39
optima and tolerances (Table S1). Nearly 40% of diatom taxa
in the pH training set occur in 10 or fewer lakes and 67%
occur in 20 or fewer lakes. Taxa encountered in >40% of the
lakes include Achnanthidium minutissimum (109 lakes),
Staurosirella pinnata (105), Staurosira construens var. venter
(102), Psammothidium helveticum (100), Nitzschia perminuta
(88), Tabellaria flocculosa (88), Aulacoseira distans-type (85),
Pseudostaurosira pseudoconstruens (75) and Psammothidium
marginulatum (73). DCA of the diatom data reveals gradient
lengths of 4.7 and 2.9 standard deviation units on the first
and second axes, respectively (Fig. 3; Table 4). Longer
gradient lengths indicate both high species turnover between
sites and that species responses to the environmental variables are unimodal (Leps and Smilauer, 2003). Eigenvalues of
DCA axes 1, 2 and 3 were 0.71, 0.30 and 0.22, respectively,
and the first three axes together accounted for 20.6% of
variance in the species data. Species located on the left side
of the DCA plot are mostly found in lakes with pH > 7.1;
these lakes are located in the western, central and northern
Arctic Archipelago (Fig. 3). The clustering of sites indicates
similar species compositions in these lakes, whereas the
spread of both taxa and lakes on the right side of the plot
shows that lakes with lower pH values contain assemblages
that are less comparable between sites, and differ greatly
from lakes situated on the left side of the plot (Fig. 3).
Generally, lakes with lower pH (7.0) have considerably
more species turnover on the first and second DCA axes.
The first three CCA axes account for 15.4% of the variance
in the species data (Table 4). While all eight environmental
variables were individually significant in explaining the
dispersion of the species data (Table 4), the variable with the
highest correlation to CCA axis 1 is pH (r ¼ 0.86). CCA
constrained solely to pH indicates that this variable explains
a high proportion of the variance in the diatom data (9.0%)
compared with the total variance explained on the first axis
of the analysis incorporating eight statistically significant
variables (10.1%; Table 4), and together with the high
eigenvalue ratio in the constrained analysis of 1.36 (l1/
l2 ¼ 0.53/0.34), warranted the development of a pH transfer
function.
Lakes on the right side of the CCA biplot are located on
Baffin Island where temperature and precipitation are higher,
and pH and conductivity are lower (Figs 1 and 4). Lakes on
the left side of the biplot are located in the central and
western Arctic, and are generally more alkaline with higher
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
5
DIATOM-DERIVED HOLOCENE PH RECONSTRUCTIONS
255
a)
MAAT
JulyT
PAnn
Cond
pH
-5
Axel Heiberg
Baffin
Banks
Bathurst
Boothia Peninsula
Bylot
Cornwallis
Devon
Ellesmere
Melville
Melville Peninsula
Prince of Wales
Somerset
Victoria
Depth
SfcA
-5
3
CCA axis 2 λ = 0.18
DFC
5
b)
71
100
197
142
86
75 144
170
43
6
9
19
20
21
22
24
25
26
27
40
44
3.0%
178
181
184
192
190
196
199
204
210
214
216
217
218
219
230
231
236
242
246
247
249
252
253
255
10.1%
-3
Figure 4. Canonical correspondence analysis
(CCA) biplots of (a) 171 lakes displayed by
region and constrained to the eight significant
environmental variables (pH; conductivity,
Cond; mean July air temperature, JulyT; mean
annual air temperature, MAAT; total annual
precipitation, PAnn; distance from coast,
DFC; lake surface area, SfcA; water depth,
Depth) and (b) the 261 diatom taxa. Taxon
names correspond to the numbering scheme
in Table S1. Loops are used as an aid to
labelling the data points.
51
52
53
58
69
73
78
79
81
82
84
93
94
95
96
104
126
131
132
133
143
149
146
151
152
154
155
156
159
160
161
163
167
168
171
172
-3
conductivity. Acidophilous taxa such as Aulacoseira distanstype, A. perglabra, Eunotia rhomboidea, Fragilariforma virescens and Achnanthes altaica ordinated on the right side of
the species plot, whereas alkaliphilous taxa on the left side of
the graph included Staurosira construens var. venter, AmphoCopyright # 2014 John Wiley & Sons, Ltd.
39
41
28 66 70
35 125
65 99180
37 137
188 193
45 139
68
238
80
48 147
185 74 191
72
49
136
256
55
54 186
76
239
98
90
83 201
91
64
88
162
42
215
87 212
202
130
102 23597
134 226 254
46
260
113 243
57
92
150 241
119 258
195 205
23 50 18
2
62
120 259
17 248 115
233 194
187
121
67 36
103 141
213
209
179 123 207
250 198 237 234 240 38 47604
7 33 140 122 106
224 222
63 85 175 200 89 153
61
135
110
32
189
10 245
174
182
101
173 8
29
229 227
261 116
14
108
251 228
206
111
105
16
12
107
124 138
211
225
169
232
221
157 220
166
127 15 257 164 3
59
77 11
158
128 165 13
145
129
183
5
148
31 109
208
176 244
203
223
177 30
114
34
117
1 112
56
118
CCA axis 1 λ = 0.61
3
ra pediculus, A. copulata, Denticula kuetzingii and Staurosirella pinnata. Taxa such as Achnanthidium minutissimum,
Nitzschia perminuta and Rossithidium pusillum plot close to
the origin of the CCA space and are generally found in
circum-neutral lakes.
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
256
r 2=0.87
RMSE=0.26
0.4
7.5
7.0
0.2
0.0
-0.2
6.5
-0.4
6.0
-0.6
6.0
8.5
6.5
7.0
7.5
8.0
8.5
Max bias=0.33
6.0
r 2=0.78
RMSEP=0.35
6.5
7.0
7.5
8.0
8.5
Max bias=0.47
0.5
8.0
Residual pH
Predicted pH
0.6
8.0
Residual pH
Estimated pH
8.5
JOURNAL OF QUATERNARY SCIENCE
7.5
7.0
0.0
-0.5
6.5
Discussion
6.0
6.0
6.5
7.0
7.5
8.0
8.5
6.0
6.5
7.0
7.5
8.0
8.5
Observed pH
Figure 5. Estimated (apparent) and predicted (bootstrapped) diatominferred lake-water pH based on the two-component weightedaveraging partial least-squares (WA-PLS) regression model developed
using 261 taxa from 168 lakes (three outliers removed) in the Canadian
Arctic, and corresponding residual plots.
Development of a pH inference model
A WA-PLS two-component inference model using 261
species from 168 lakes (following removal of three outliers)
had a bootstrapped r2 of 0.78 and RMSEP of 0.35 (Fig. 5;
Table 1). The optima and tolerances of each of the 261
species used in the WA-PLS two-component model were
generated by WA (Table S1). Taxon-specific pH optima
ranged between 6.2 and 8.5 and are distributed relatively
evenly across this gradient, with modes of taxa having pH
optima of 6.6 and 8.0.
A spherical variogram model was fitted to the pH data to
assess the effects of spatial structure on model development
(Fig. 6a). The nugget variance is small (0.06), which indicates
that geographically nearby samples tend to be similar, or that
spatial autocorrelation is only significant at very localized
scales (i.e. 20 km; Wackernagel, 2003). The partial sill (0.41)
represents the spatially structured component of pH data, and
the wide scatter among the points that fall within the partial
sill indicate that this dataset has a wide range of values at a
smaller spatial scale (100–400 km). The fit improves at
distances >540 km, where spatial autocorrelation becomes
small. These results (Fig. 6a) are comparable to the freshwater
diatom example of Telford and Birks (2009), where spatial
autocorrelation was not strong enough to affect the development of an inference model. The predictive power of the
transfer function was tested using 1000 unconditional Gaussian simulations to derive a new variable with the same spatial
structure as the observed pH data (Telford and Birks, 2009).
The r2 between observed and predicted pH values was greater
than the r2 of all simulations, indicating that the pH transfer
function was statistically significant and robust enough for
reconstructing lake-water pH (Fig. 6b). In general, WA-PLS
models have been shown previously to be resistant to the
effects of spatial autocorrelation (Telford and Birks, 2005).
Holocene pH reconstructions
The two-component WA-PLS model was applied to the fossil
assemblages from Lake PW02 on Russell Island (Finkelstein
and Gajewski, 2008), Lake KR02 on Victoria Island (Podritske
and Gajewski, 2007) and Lake CF3 on Baffin Island (Briner
et al., 2006) (Fig. 1a). Diatom-inferred pH estimates from Lake
PW02 varied little during the Holocene (range ¼ 7.3–7.6;
Copyright # 2014 John Wiley & Sons, Ltd.
Fig. 7). At Lake KR02, reconstructed pH was variable, ranging
between 6.5 and 7.8, with minimum pH values reconstructed
at 6000, 5000 and 2000 cal a BP. At Lake CF3, the range of
inferred pH was somewhat less (6.5–7.5) than at Lake KR02
and, with the exception of one value at the bottom of the CF3
core, the record exhibits a long-term directional decline in pH
during the Holocene. Of these three reconstructions, only that
from Lake CF3 is statistically significant using the method of
Telford and Birks (2011; Fig. 7; P < 0.01). The reconstructions
for lakes PW02 and KR02 were non-significant at P ¼ 0.954
and P ¼ 0.185, respectively.
Analysis of the inference model and Holocene pH
reconstructions
Ordination by DCA of modern diatom assemblages in the
171-lake training set shows considerable diversity in the
diatom flora across the Canadian Arctic, whereas constrained
ordination by CCA confirms the importance of pH in affecting
the distribution of diatom taxa regionally. The resulting pH
inference model produces reasonable performance statistics
(Table 1; Fig. 5). Spatial structure was encountered in the pH
data for sites in close proximity (<20 km; Fig. 6). This result
emerges in part due to the geographical structure of the
dataset, which is imposed by logistical constraints of working
in remote regions. In some sampling programs for example,
an area is chosen with several lakes in close proximity that
are either sampled on foot or by aircraft, resulting in densely
spaced ‘clumps’ of sites (e.g. QB series, BY series, HB series,
WB series; Fig. 1b) collected from a relatively uniform
environment.
Regardless of spatial structures that exist inherently in the
dataset over short distances, the inference model is effective
because of its large spatial extent, large number of sites and
the long pH gradient (Figs 2 and 5). The discrete nature of
lakes makes limnological variables such as pH less prone to
strong spatial autocorrelation compared with, for example,
air temperature, where advection induces spatial autocorrelation in any measurement series (Telford and Birks, 2009). For
example, lakes on the north coast of Baffin Island and on
Bylot Island are separated by <50 km and share similar
climate, but differ by >2 pH units due to variations in bedrock
geology and surficial materials. While the analysis of spatial
autocorrelation effects provides added confidence in the
robustness of Holocene pH reconstructions, it also emphasizes
that the optimal design of diatom training sets should include
samples from beyond a 20-km radius, in addition to samples
within the 20-km radius encompassing as broad a range as
possible for the target environmental variables.
When compared with other pH inference models derived
from diatom datasets in Arctic regions, the two-component
WA-PLS model presented here has comparable performance
(Fig. 5; Table 1). By way of comparison to temperate regions,
a pH transfer function developed from 494 lakes in northeastern North America produced an r2 of 0.89 and an
average RMSEP of 0.45 (Ginn et al., 2007). Reavie and
Juggins (2011) and Reavie and Edlund (2013) show that
training sets containing >70 samples, spanning broad environmental gradients and containing a variety of diatom taxa
result in the most effective transfer functions, all conditions
which are met in the new model developed here. The
Holocene pH reconstruction presented here for Lake CF3
using the larger training set (n ¼ 171 lakes) spans 1.0 pH unit
(6.5–7.5), compared with the original reconstruction spanning 1.5 units (6.5–8.0), which was developed with a
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
DIATOM-DERIVED HOLOCENE PH RECONSTRUCTIONS
a)
b)
150
0.5
257
Observed r 2(0.78)
100
50
0.2
Frequency
0.3
Sill (0.47)
Partial sill (0.41)
Semivariance
0.4
0.1
Range (539 km)
200
0
Nugget (0.06)
400
600
800
1000
Distance (km)
0.0
0.2
Simulated r
0.4
0.6
0.8
1.0
2
Figure 6. (a) Variogram of the pH data from 168 retained lakes (three outliers removed) showing the relationship between the average variance
between pH at different points in space and their distance apart. The small nugget variance indicates that nearby samples tend to be similar; the sill
represents the variance in the pH variable; the partial sill is the spatially structured component of the pH data; and the range indicates the distance
at which the data are autocorrelated. (b) Histogram of the r2 of 1000 random trials and the observed and inferred r2 (0.78), indicating that the
transfer function is statistically significant and that pH can be reconstructed using this training set.
regional training set containing 61 lakes (Briner et al., 2006).
The larger training set also provides a reconstructed surface
sample pH value closer to the modern value, indicating that
the larger training set spanning longer environmental gradients and containing more taxa provides more accurate
Holocene estimates.
Significance testing of the pH reconstructions using the
methodology of Telford and Birks (2011) indicates that
the pH reconstruction for Lake CF3 is significant, confirming
the emphasis on pH in the original analyses of the record
(Briner et al., 2006). The reconstruction for Lake KR02 is not
significant. However, there are clear and substantial shifts in
diatom assemblage composition and sediment characteristics
in the KR02 record (Podritske and Gajewski, 2007); the nonsignificant pH reconstruction implies that other variables
explain those shifts. The reconstruction for Lake PW02 is not
significant, and there are only subtle stratigraphic changes
in the Holocene record from that site (Finkelstein and
Gajewski, 2008). Non-significance of these two reconstructions may relate in part to the at times low similarity between
some modern and fossil diatom datasets. For example, there
are no lakes in the modern dataset with percentages of small
Fragilarioid taxa as high as those reported for some fossil
samples from Lakes KR02 and PW02 (>90% abundance). It is
has been noted that some modern Arctic diatom assemblages
may have already responded markedly to anthropogenic
environmental change since the onset of industrialization, and
are thus too dissimilar from any fossil assemblages to adequately
quantify past environmental conditions (Smol et al., 2005; Smol
and Douglas, 2007; Adams and Finkelstein, 2010).
pH as a driver of Holocene shifts in diatom
communities in Arctic lakes
A long-term decline of 1 pH unit took place over the past
11 200 years at Lake CF3. This result is comparable to trends
exhibited in subarctic lakes in Fennoscandia and Russia
(Korhola and Weckström, 2004). Holocene pH declines at
Copyright # 2014 John Wiley & Sons, Ltd.
these sites are interpreted to have resulted from catchment
processes, including paludification, vegetation change and
weathering of felsic bedrock lithologies. As Lake CF3 is
situated in a bedrock basin of gneissic rocks, and soils in the
surrounding watershed are thin and poorly developed,
accumulation of acidic run-off in the lake basin may be
contributing to the decline in pH and this conclusion is
supported by the long-term directional trend in pH.
Direct effects of climate on lake ice cover and pCO2 may
also contribute to the Holocene pH trends at Lake CF3
(Wolfe, 2002). In contrast to the Fennoscandian sites reported
on by Korhola and Weckström (2004) where acidification
takes place relatively soon after lake inception, the rate of pH
decline at CF3 is initially low, only increasing after 6000 cal
a BP. This pattern offers some insight into separating edaphic
versus climatic causes of long-term pH decline at this site.
The intensification of pH decline at Lake CF3 after 6000 cal a
BP is coincident with a 4 ˚C decline in chironomid-inferred
July air temperatures at Lake CF3 (Briner et al., 2006) and
with a series of regional records documenting onset of postHolocene Thermal Maximum cooling at this time (Kaufman
et al., 2004). The correlation between cooling climate and
diatom-inferred pH supports the idea that increasing lakewater acidity due to prolonged ice cover is a primary factor
affecting pH at sites such as CF3 where edaphic buffering
capacity is minimal (Koinig et al., 1998).
Lake KR02 provides a complex Holocene record, with
changes through time in the diatom influx and preservation,
and the presence of three diatom dissolution zones (8500–
8200, 7600–7000 and 5600–5000 cal a BP). Diatom zonation
is determined mainly by changes in the relative proportions
of Fragilarioid taxa which dominate the record, and which in
the modern dataset vary in pH optimum by 0.5 units
between pH values of 7.5 and 8.0 (Table S1), along with
lower abundances of benthic, epiphytic and planktonic
diatoms. Major shifts in the diatom assemblages include a
zone defined by up to 15% abundances of the planktonic
diatom genus Cyclotella in the Early Holocene (9600–
8000 cal a BP), which may reflect a longer open water season
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
258
JOURNAL OF QUATERNARY SCIENCE
a)
b)
PW02
c)
KR02
CF3
0
1000
2000
3000
Age (cal a BP)
4000
5000
6000
7000
8000
9000
10000
11000
8.0
6.4
pH
pH
P=0.185
P=0.001
0
0
20
20
40
40
60
40
0
20
Frequency
60
60
80
80
P=0.954
8.0
7.2
pH
100
pH
100
6.4
8.0
7.2
pH
pH
7.2
80
6.4
0.0
0.1
0.2
0.3
0.4
0.0
0.1
0.2
0.3
0.4
0.5
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Proportion variance explained
Figure 7. Lake-water pH reconstructions for (a) Lake PW02, (b) Lake KR02 and (c) Lake CF3 inferred using a two-component weighted-averaging
partial least-squares regression (WA-PLS) model and corresponding histogram of the significance tests.
The histogram represents the proportion of variance in the diatom records explained by 1000 random WA-PLS two-component transfer functions.
The solid line represents the pH reconstruction and the dashed line represents the proportion of variance explained by the first axis of a principal
components analysis of the fossil data (Telford and Birks, 2011). Horizontal grey lines on the inferred pH curves represent the sample-specific
standard errors generated using 1000 bootstrap iterations.
during at the time of the Holocene Thermal Maximum. A
second shift in diatoms is defined by an increase in the
epiphytic diatom Amphora copulata between 5000 and
3000 cal a BP, which may reflect shifts in littoral zone habitat
(Podritske and Gajewski, 2007). Thus, the non-significant pH
reconstruction supports climate and habitat availability as
major drivers of Holocene change in diatom assemblages at
this site, and indicates effective local buffering.
At Lake PW02, there were few changes in the diatom
assemblages and the pH reconstruction is also non-significant. Because of alkaline bedrock, the lake is well buffered,
with a modern pH of 7.9. Fragilarioid diatoms overwhelmingly dominate fossil assemblages throughout the Holocene,
indicating that geological buffering acts to maintain pH in the
range 7.5–8.0 optimal for the Fragilarioid taxa. While nonCopyright # 2014 John Wiley & Sons, Ltd.
varying pH in itself does necessarily mean a reconstruction
will be non-significant, the lack of suitable fossil-modern
analogues may be a factor at this site.
The larger 171-lake training set produced more robust
reconstructions for the Lake CF3 record relative to a smaller,
regional training set. Taken as a whole, our analysis suggests
that larger training sets are in general beneficial for producing
significant and accurate reconstructions, with the conditions
that: (i) the larger training must have well-harmonized taxonomy; (ii) the reconstructions are not performed at the very edge
of environmental gradients (Velle et al., 2011); (iii) the models
are not applied to regions outside the training set’s climate
regime (Self et al., 2011); (iv) the effects of spatial autocorrelation are accounted for (Telford and Birks, 2009); and (v)
modern analogs exist for all fossil assemblages.
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
DIATOM-DERIVED HOLOCENE PH RECONSTRUCTIONS
Conclusions
An expanded freshwater diatom training set for the Canadian
Arctic is an improvement over existing regional training sets,
as it spans broader taxonomic and environmental gradients,
and is thus more likely to find acceptable analogs for diverse
fossil assemblages. This emerges as a potentially important
feature, given that the current rate and magnitude of changes
in diatom assemblages in Arctic lakes mandate training sets
that are as taxonomically diverse as possible. The expanded
training set allows for improved reconstructions of significant
parameters in the model, in particular lake-water pH, given
the improved performance statistics for this variable. Application of the new pH model to three diatom sequences shows
the disparate roles for and drivers of pH change over the
Holocene, illustrating the complex relationships between
geology, catchment and lake processes, and climate in Arctic
lakes.
Supporting Information
Additional supporting information can be found in the online
version of this article:
Table S1. Numbers, abbreviations, name, authority and family
for taxa presented in this study.
Table S2. Analyst, reference, latitude, longitude, island/
peninsula, mean July air temperature, mean annual air
temperature, total annual precipitation, distance from coast,
lake surface area, water depth, lake-water conductivity and
lake-water pH for 171 lakes in the training set.
Acknowledgements. Our Arctic field research is funded by the
Natural Sciences and Engineering Research Council of Canada, the
Government of Canada Fund for International Polar Year, the Parks
Canada Agency, the Polar Continental Shelf Program and the Northern
Scientific Training Program. This research was also supported by a
post-doctoral fellowship to J.B. from the Centre for Global Change
Science at the University of Toronto. We thank Jiye (Grace) Jeon for
help in the laboratory, Kailey Stewart and Neal Michelutti for
providing diatom data from West Lake and Lake CF3, respectively,
and Richard Telford and Andrew Medeiros for helpful comments.
Finally, we are very grateful for the ongoing support of the Nunavut
Research Institute (Nunavummi Qaujisaqtulirijikkut). The authors
declare no conflict of interest.
Abbreviations. CCA, canonical correspondence analysis; DCA, detrended correspondence analysis; NMDS, non-metric multidimensional scaling; RMSEP, root mean squared error of prediction; WA,
weighted-averaging; WA-PLS, weighted-averaging partial least-squares
References
Adams JK, Finkelstein SA. 2010. Watershed-scale reconstruction of
middle and late Holocene paleoenvironmental changes on Melville
Peninsula, Nunavut, Canada. Quaternary Science Reviews 29:
2302–2314.
Antoniades D, Douglas MSV, Smol JP. 2004. Diatom species–
environment relationships and inference models from Isachsen, Ellef
Ringnes Island, Canadian High Arctic. Hydrobiologia 529: 1–18.
Antoniades D, Douglas MSV, Smol JP. 2005. Benthic diatom
autecology and inference model development from the Canadian
High Arctic Archipelago. Journal of Phycology 41: 30–45.
Battarbee RW. 1984. Diatom analysis and the acidification of lakes.
Philosophical Transactions of the Royal Society of London Series B
305: 451–477.
Battarbee RW, Charles DF, Bigler C, et al. 2010. Diatoms as indicators
of surface-water acidity. In The Diatoms: Applications for the
Environmental and Earth Sciences. 2nd edn, Smol JP, Stoermer EF
(eds). Cambridge University Press: Cambridge; 98–121.
Copyright # 2014 John Wiley & Sons, Ltd.
259
Battarbee RW, Smol JP, Meriläinen J. 1986. Diatoms as indicators of
pH: an historical review. In Diatoms and Lake Acidity, Smol JP,
Battarbee RW, Davis RB, Meriläinen J (eds). Dr W Junk: Dordrecht;
5–14.
Bigler C, Hall RI. 2002. Diatoms as indicators of climatic and
limnological change in Swedish Lapland: a 100-lake calibration set
and its validation for paleoecological reconstructions. Journal of
Paleolimnology 27: 97–115.
Birks HJB, Line JM, Juggins S, et al. 1990. Diatoms and pH
reconstruction. Philosophical Transactions of the Royal Society of
London Series B Biological Sciences 327: 263–278.
Birks HJB, Simpson GL. 2013. ‘Diatoms and pH reconstruction.
(1990) revisited. Journal of Paleolimnology 49: 363–371.
Blackadar RG, Christie RL. 1963. Geological Reconnaissance: Boothia Peninsula, and Somerset, King William, and Prince of Wales
Islands, District of Franklin, Geological Survey of Canada Paper
69-13: Ottawa: Canada.
Bouchard G, Gajewski K, Hamilton PB. 2004. Freshwater diatom
biogeography in the Canadian Arctic Archipelago. Journal of
Biogeography 31: 1955–1973.
Boyle J, Chiverrell R, Plater A, et al. 2013. Soil mineral depletion
drives Early Holocene lake acidification. Geology 41: 415–418.
Boyle JF. 2007. Loss of apatite caused irreversible early-Holocene
lake acidification. Holocene 17: 543–547.
Briner JP, Michelutti N, Francis DR, et al. 2006. A multi-proxy
lacustrine record of Holocene climatic change on northeastern
Baffin Island, Arctic Canada. Quaternary Research 65: 431–442.
Camburn KE, Charles DF. 2000. Diatoms of Low-Alkalinity Lakes in
the Northeastern United States, Special Publ.18. Academy of
Natural Sciences of Philadelphia: Philadelphia.
Crawford RM, Likhoshway Y. 1999. The frustule structure of original
material of Aulacoseira distans (Ehrenberg) Simonsen. Diatom
Research 14: 239–250.
Devlin JE, Finkelstein SA. 2011. Recent ecological changes in aquatic
ecosystems in two distinct physiographic regions of Sirmilik
National Park, Nunavut, Canada. Journal of Paleolimnology 45:
23–39.
Dyke AS, Morris TF, Green DEC, et al. 1992. Quaternary Geology of
Prince of Wales Island, Arctic Canada, Memoir 433. Geological
Survey of Canada: Ottawa.
Erlandsson M, Bishop K, Fölster J, et al. 2008. A comparison of
MAGIC and paleolimnological predictions of preindustrial pH for 55
Swedish lakes. Environmental Science and Technology 42: 43–48.
Finkelstein SA, Gajewski K. 2008. Responses of Fragilarioid-dominated diatom assemblages in a small Arctic lake to Holocene climatic
changes, Russell Island, Nunavut, Canada. Journal of Paleolimnology 40: 1079–1095.
Fourtanier E, Kociolek JP. 2012. Catalogue of Diatom Names.
California Academy of Sciences: San Francisco.
Frisch T, Trettin HP. 1991. Precambrian successions in the northernmost part of the Canadian Shield. In Geology of the Innuitian
Orogen and Arctic Platform of Canada and Greenland, Trettin HP
(ed.). Geological Survey of Canada: Ottawa; 109–115.
Ginn BK, Cumming BF, Smol JP. 2007. Diatom-based environmental
inferences and model comparisons from 494 northeastern North
American lakes. Journal of Phycology 43: 647–661.
Hadley KR, Douglas MSV, Lim D, et al. 2013. Diatom assemblages
and limnological variables from 40 lakes and ponds on Bathurst
Island and neighboring high Arctic islands. International Review of
Hydrobiology 98: 44–59.
Hamilton PB, Gajewski K, Atkinson DE, et al. 2001. Physical and
chemical limnology of 204 lakes from the Canadian Arctic
Archipelago. Hydrobiologia 457: 133–148.
Hervé V, Derr J, Douady S, et al. 2012. Multiparametric analyses
reveal the pH-dependence of silicon biomineralization in diatoms.
PLOS ONE 7: 1371/journal.pone.0046722.
Hill MO, Gauch HG. 1980. Detrended Correspondence-Analysis an improved ordination technique. Vegetatio 42: 47–58.
Hustedt F. 1937–1939. Systematische und ökologische Untersuchungen über die Diatomeen-Flora von Java, Bali, Sumatra. Archiv für
Hydrobiologie (Suppl.): 15–16.
Joynt EH III, Wolfe AP. 2001. Paleoenvironmental inference models
from sediment diatom assemblages in Baffin Island lakes
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)
260
JOURNAL OF QUATERNARY SCIENCE
(Nunavut, Canada) and reconstruction of summer water temperature. Canadian Journal of Fisheries and Aquatic Sciences 58: 1222–
1243.
Juggins S. 2012. rioja: an R package for the analysis of Quaternary
science data, version 0.7-3.
Juggins S. 2013. Quantitative reconstructions in paleolimnology: new
paradigm or sick science. Quaternary Science Reviews 64: 20–32.
Kaufman DS, Ager TA, Anderson NJ, et al. 2004. Holocene thermal
maximum in the western Arctic (0-180 degrees W). Quaternary
Science Reviews 23: 529–560.
Keatley BE, Douglas MSV, Smol JP. 2008. Evaluating the influence of
environmental and spatial variables on diatom species distributions
from Melville Island (Canadian High Arctic). Botany 86: 76–90.
Koinig KA, Schmidt R, Sommaruga-Wögrath S, et al. 1998. Climate
change as the primary cause for pH shifts in a high alpine lake.
Water, Air, and Soil Pollution 104: 167–180.
Korhola A, Weckström J. 2004. Paleolimnological studies in Arctic
Fennoscandia and the Kola Peninsula (Russia). In Long-Term
Environmental Change in Arctic and Antarctic Lakes, Pienitz R,
Douglas MSV, Smol JP (eds). Springer: Dordrecht; 381–418.
Lavoie M, Le Faucheur S, Boullemant A, et al. 2012. The influence
of pH on algal cell membrane permeability and its implications for
the uptake of lipophilic metal complexes. Journal of Phycology 48:
293–302.
Leps J, Smilauer P. 2003. Multivariate Analysis of Ecological Data
Using CANOCO. Cambridge University Press: Cambridge.
Meriläinen J. 1967. The diatom flora and the hydrogen-ion concentration of water. Annales Botanici Fennici 4: 51–58.
Michelutti N, Smol JP, Douglas MSV. 2006. Ecological characteristics
of modern diatom assemblages from Axel Heiberg Island (High
Arctic Canada) and their application to paleolimnological inference
models. Canadian Journal of Botany 84: 1695–1713.
New M, Lister D, Hulme M, et al. 2002. A high-resolution data set of
surface climate over global land areas. Climate Research 21: 1–25.
Pebesma EJ. 2004. Multivariate geostatistics in S: the gstat package.
Computers and Geosciences 30: 683–691.
Podritske B, Gajewski K. 2007. Diatom community response to
multiple scales of Holocene climate variability in a small lake on
Victoria Island, NWT, Canada. Quaternary Science Reviews 26:
3179–3196.
Psenner R, Schmidt R. 1992. Climate-driven pH control of remote
alpine lakes and effects of acid deposition. Nature 356: 781–783.
R Development Core Team. 2011. R: A Language and Environment
for Statistical Computing. R Foundation for Statistical Computing:
Vienna; www.r-project.org.
Reavie ED, Edlund MB. 2013. Assessing the performance of a diatom
transfer function on four Minnesota lake sediment cores: effects of
training set size and sample age. Journal of Paleolimnology 50: 87–
104.
Reavie ED, Juggins S. 2011. Exploration of sample size and diatombased indicator performance in three North American phosphorus
training sets. Aquatic Ecology 45: 529–538.
Renberg I. 1990. A 12600 year perspective of the acidification of LillaOresjon, Southwest Sweden. Philosophical Transactions of the Royal
Society of London Series B Biological Sciences 327: 357–361.
Copyright # 2014 John Wiley & Sons, Ltd.
Renberg I, Hellberg T. 1982. The pH history of lakes in southwestern
Sweden, as calculated from the subfossil diatom flora of the
sediments. Ambio 11: 30–33.
Rosén P, Hall R, Korsman T, et al. 2000. Diatom transfer-functions
for quantifying past air temperature, pH and total organic carbon
concentration from lakes in northern Sweden. Journal of Paleolimnology 24: 109–123.
Self AE, Brooks SJ, Birks HJB, et al. 2011. The distribution and
abundance of chironomids in high-latitude Eurasian lakes with
respect to temperature and continentality: development and application of new chironomid-based climate-inference models in
northern Russia. Quaternary Science Reviews 30: 1122–1141.
Silva P. 2012. Index Nominum Algarum. University Herbarium
University of California: Berkeley.
Smol JP, Douglas MSV. 2007. From controversy to consensus: making
the case for recent climate change in the Arctic using lake
sediments. Frontiers in Ecology and the Environment 5: 466–474.
Smol JP, Wolfe AP, Birks HJ, et al. 2005. Climate-driven regime shifts
in the biological communities of Arctic lakes. Proceedings of the
National Academy of Sciences of the United States of America
102: 4397–4402.
Sommaruga-Wögrath S, Koinig KA, Schmidt R, et al. 1997. Temperature effects on the acidity of remote alpine lakes. Nature 387: 64–67.
Stewart KA, Lamoureux SF. 2012. Seasonal and microhabitat influences on diatom assemblages and their representation in sediment
traps and surface sediments from adjacent High Arctic lakes: Cape
Bounty, Melville Island, Nunavut. Hydrobiologia 683: 265–286.
Telford RJ. 2012. palaeoSig: significance tests for palaeoenvironmental reconstructions, version 1.0.
Telford RJ, Birks HJB. 2005. The secret assumption of transfer
functions: problems with spatial autocorrelation in evaluating
model performance. Quaternary Science Reviews 24: 2173–2179.
Telford RJ, Birks HJB. 2009. Evaluation of transfer functions in
spatially structured environments. Quaternary Science Reviews 28:
1309–1316.
Telford RJ, Birks HJB. 2011. A novel method for assessing the
statistical significance of quantitative reconstructions inferred from
biotic assemblages. Quaternary Science Reviews 30: 1272–1278.
ter Braak CJF. 1986. Canonical correspondence analysis: a new
eigenvector technique for multivariate direct gradient analysis.
Ecology 67: 1167–1179.
Velle G, Kongshavn K, Birks HJB. 2011. Minimizing the edge-effect
in environmental reconstructions by trimming the calibration set:
Chironomid-inferred temperatures from Spitsbergen. Holocene 21:
417–430.
Wackernagel H. 2003. Multivariate Geostatistics: an Introduction
with Applications. 3rd edn. Springer: Berlin.
Whitehead DR, Charles DF, Jackson ST, et al. 1989. The developmental history of Adirondack (N.Y.) lakes. Journal of Paleolimnology 2: 185–206.
Wilson CR, Michelutti N, Cooke CA, et al. 2012. Arctic lake
ontogeny across multiple interglaciations. Quaternary Science
Reviews 31: 112–126.
Wolfe AP. 2002. Climate modulates the acidity of Arctic lakes on
millennial time scales. Geology 30: 215–218.
J. Quaternary Sci., Vol. 29(3) 249–260 (2014)