Journal of Great Lakes Research 41 (2015) 20–29
Contents lists available at ScienceDirect
Journal of Great Lakes Research
journal homepage: www.elsevier.com/locate/jglr
Anthropogenic influences on the sedimentary geochemical record in
western Lake Superior (1800–present)
Molly D. O'Beirne a,⁎, Ladislaus J. Strzok a,b, Josef P. Werne a,c, Thomas C. Johnson a,d, Robert E. Hecky a,e
a
Large Lakes Observatory, University of Minnesota Duluth, Duluth, MN 55812, USA
Minnesota Pollution Control Agency, St. Paul, MN 55155, USA
Department of Geology and Planetary Science, University of Pittsburgh, Pittsburgh, PA 15260, USA
d
Department of Earth and Environmental Sciences, University of Minnesota Duluth, Duluth, MN 55812, USA
e
Department of Biology, University of Minnesota Duluth, Duluth, MN 55812, USA
b
c
a r t i c l e
i n f o
Article history:
Received 13 September 2013
Accepted 23 September 2014
Available online 6 December 2014
Communicated by Gerald Matisoff
Index words:
Sedimentary organic carbon
Carbon isotopes
n-Alkane biomarkers
Lake Superior
Productivity
Anthropogenic impacts
a b s t r a c t
The sediments of western Lake Superior hold a record of environmental changes that have accompanied the settlement and urbanization of the surrounding watershed. Organic carbon concentrations are low (1.5%) with little
variation in stable isotope composition (−26.5 ± 0.5‰) prior to 1900. Organic carbon and nitrogen concentrations begin to rise after 1900, as increased anthropogenic disturbance led to enhanced inputs of terrigenous
matter as well as nutrients (i.e., nitrogen and phosphorus) from the watershed. An episode of enhanced aquatic
productivity from 1900 to 1970 is recorded in the sediments by the 13C-enrichment of bulk organic carbon as well
as the observed correlation between the bulk and aquatic molecular δ13C records, coinciding with the major developmental period of the Duluth-Superior harbor region. Decreasing organic carbon accumulation after 1925,
prior to regulatory implementation of municipal discharges to the lake, is likely due to the construction of hydropower dams along the St. Louis River and a decrease in forest harvest within the immediate watershed.
Recentshort-lived decreases in the accumulation rate of organic matter can be attributed to the implementation
and operation of water treatment plants, but the 13C-enrichment observed in the last ~ decade remains enigmatic, though we hypothesize that it may be attributable to climate change impacts on primary production.
© 2014 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.
Introduction
Lakes are intimately connected with their surrounding environment,
making them sensitive indicators of perturbations within their watersheds. Organic geochemical proxies preserved in lake sediments can provide a history of ecosystem change (e.g., carbon cycling). These proxies
enable us to gauge the historical impacts of regional development and
subsequent regulations within lake systems while providing insight to
guide possible management action to address any emerging issues.
The stable carbon isotope composition of sedimentary organic carbon (δ13Corg), a proxy for lacustrine productivity (Meyers, 1994), is
often used to track the response of lakes to historic anthropogenic activities (i.e., nutrient loading) and subsequent regulation. This approach is
based on the premise that the δ13C of organic carbon produced photosynthetically responds to the rate of primary productivity in the water
column (Schelske and Hodell, 1991; Meyers, 2003). In general, increases
in lacustrine productivity should lead to increases in the deposition of
total organic carbon (TOC) and total nitrogen (TN) concentrations
⁎ Corresponding author at: Department of Geology and Planetary Science, University of
Pittsburgh, 200 Space Research Coordination Center, 4107 O'Hara Street Pittsburgh, PA
15260-3332. Tel.: +1 412 624 8780.
E-mail address: [email protected] (M.D. O'Beirne).
along with changes in their stable isotope compositions (δ13Corg
and δ15N, respectively). δ13Corg and δ15N are typically 13C- and
15
N-enriched (depleted) with increasing (decreasing) in-lake primary
production (Mckenzie, 1985; Meyers, 1997, 2003). Thus, the bulk isotope compositions as well as accumulation rates of Corg and TN can be
used as a proxy for lacustrine productivity.
One caveat to this general approach is that sedimentary C and N are
derived from both terrigenous and aquatic sources, thus increases or decreases cannot be unequivocally allocated to one source or the other.
Similarly, the isotopic composition of sedimentary TOC is influenced
by both terrigenous and aquatic sources, making it useful to look at molecular biomarkers (e.g., n-alkanes) to separate contributions from
these sources. The two principal sources of hydrocarbons to lake sediments are those of photosynthetic algae and bacteria, dominated by
the odd-numbered short chain n-alkanes(n-C17 + n-C19 + n-C21)
(Cranwell et al., 1987; Giger et al., 1980) and those contributed from
vascular land plants, which contain large proportions of the oddnumbered long chain n-alkanes(n-C27 + n-C29 + n-C31) that are produced in their epicuticular waxy coatings (Cranwell, 1981; Eglinton
and Hamilton, 1963, 1967; Rieley et al., 1991). In order to circumvent
the issue of source, the stable carbon isotope composition of n-alkanes
(δ13Cn-alkane), when compared to the bulk δ13Corg record, can provide
an indication of the relative contributions of terrigenous and aquatic
http://dx.doi.org/10.1016/j.jglr.2014.11.005
0380-1330/© 2014 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.
M.D. O'Beirne et al. / Journal of Great Lakes Research 41 (2015) 20–29
sourced organic carbon to the bulk δ13Corg record (Bourbonniere and
Meyers, 1996; Meyers, 2003).
The Laurentian Great Lakes have experienced dramatic changes in response to anthropogenic loadings and management activities. Studies of
organic geochemical proxies preserved within sediment cores from
Lakes Erie and Ontario have shown periods of increased lacustrine productivity in correspondence with watershed/regional development.
Schelske and Hodell (1995) and more recently, Lu et al. (2010), found
that increases in lacustrine productivity (increases in the δ13Corg) of Lake
Erie were related to historic changes in anthropogenic phosphorus loading during the development of the watershed and subsequent decreases
were in response to the implementation of phosphorus abatement programs in the mid-1970s. A study by Hodell and Schelske (1998) in Lake
Ontario shows similar correspondence between productivity and anthropogenic activity as inferred from the sedimentary geochemical record.
One study from Lake Superior, measuring sedimentary biogenic silica
(BSi), provides evidence of low-level eutrophication associated with the
late 1800s European habitation and subsequent development of the watershed (Schelske et al., 2006). To date, these findings based on BSi in
Lake Superior have not been corroborated by other proxies.
Here, we report temporal trends in productivity over the last 150
years, through the use of bulk organic and stable isotope compositions
of carbon and nitrogen as well as n-alkane molecular biomarker abundances and their stable carbon isotope compositions, as contained in
the sediments of the western arm of Lake Superior. We compare two
cores from the same location taken 6 years apart which allow us to determine not only the reproducibility of the sedimentary record but also
evaluate the effect of early sedimentary diagenesis on the preservation
of biomarker signals through time.
Regional setting and background
Lying along boundary of the United States (US) and Canada, Lake
Superior (Fig. 1), the largest of the Laurentian Great Lakes and the
21
world's largest freshwater lake by surface area, is often regaled as the
most “pristine” of the Great Lakes. This is due, in part, to the small
amount of surface runoff Lake Superior receives in relation to its
volume, as well as to the fact that most of the 127,700 km2 watershed
(considered small, in comparison to its surface area with a terrigenous
catchment area to lake area ratio of ~1.5:1) is sparsely populated and
heavily forested with little agriculture. As a result, Lake Superior has
not suffered from the high nutrient loadings or industrial pollution to
the same extent as other Great Lakes (e.g., Lake Erie; Bourbonniere
and Meyers, 1996; Schelske et al., 2006).
Although Lake Superior is considered the least altered of the Laurentian Great Lakes, has been classified as ultra-oligotrophic and mostly retains its pristine water quality conditions (Munawar and Munawar,
1978; Weiler, 1978), changes in its physical, biological and chemical
processes have been documented. Observations of rising nitrate concentrations since the 1900s (Sterner et al., 2007; Weiler, 1978), as
well as increasing water temperatures over the past two centuries
(Austin and Colman, 2007, 2008) and changes in primary productivity
(Urban et al., 2005; Urban, 2007; Vollenweider et al., 1974) indicate
that anthropogenic activities can impact Lake Superior.
The highest population density within the Lake Superior watershed
is located in the St. Louis River drainage basin, which is also the second
largest inflow to Lake Superior. The St. Louis River basin has been heavily impacted by anthropogenic changes and has been designated as one
of 43 areas of concern (having impaired beneficial uses due to pollution)
within the Laurentian Great Lakes by the International Joint Commission (IJC). The area of concern was later extended to include the
Duluth-Superior Harbor region (Fig. 1).
The objective of this study is to define changes in the primary productivity of the western end of Lake Superior (Fig. 1) over the past
150 years in order to ascertain what the important drivers of change
may have been and to assess the efficacy of regulatory action to address
those changes. If anthropogenically induced changes in the trophic state
of western Lake Superior have occurred, we should be able to track
Fig. 1. (A) Map of Laurentian Great Lakes, (B) map of western Lake Superior with core locations, IJC designated Area of Concern (shaded area) and the St. Louis river with dam locations (left
to right, Scanlon Dam, Thomson Dam, and Fond du Lac Dam).
22
M.D. O'Beirne et al. / Journal of Great Lakes Research 41 (2015) 20–29
these changes in the sedimentary geochemical record, as has been done
successfully in lakes Erie and Ontario (Hodell and Schelske, 1998; Lu
and Meyers, 2009; Lu et al., 2010; Schelske and Hodell, 1995). Therefore,
we hypothesized that increases in productivity should coincide with
increases in the δ13C of both bulk and molecular aquatic biomarker
proxies and such increases should be expected to occur within the
timeframe of regional development (1900–present). Conversely, decreases in productivity would be expected to be observed after regulatory actions were put in place in the mid-1970s.
Methods
Sediment cores
Two sediment multicores, designed to capture a well-preserved sediment water interface, were collected aboard the R/V Blue Heron from
the westernmost arm of Lake Superior (Fig. 1), one in 2003 (BH03-1;
46° 48.51′N, 91° 51.95′W; water depth 44 m) and one in 2009 (BH095; 46° 49.00′N, 91° 51.9′W; water depth 48 m). Core sites were identified
using a Knudsen 12 kHz Hi-Res and CHIRP seismic reflection profilers
onboard the R/V Blue Heron. The coring sites were chosen based on evidence of uninterrupted stratigraphy indicating continuous sedimentation characteristic of similar locations in western Lake Superior. While
storm wave activity can periodically disrupt sedimentation in the relatively shallow depths of these core sites, the sedimentation rate in this
part of Lake Superior is unusually high and we assume that it greatly exceeds the occasional erosional effect of a passing storm. This assumption
is confirmed by 210Pb results, discussed below. Core BH09-5MC was selected to give an indication of the reproducibility of results, acting as a
sister location to core BH03-1MC as well as to evaluate possible diagenetic effects on the developing record and recent trends.
Initial core description and splitting were performed at the National
Lacustrine Core Repository (LacCore). Subsamples were taken from
each core at 0.5 cm intervals to 10 cm and every 1 cm thereafter to
the base of the core. Sediment samples were subsequently frozen and
freeze-dried for geochemical analyses.
210
Pb geochronology
A total of 24 sub-samples from core BH03-1, and 20 sub-samples
from BH09-5 were taken for 210Pb geochronology. Intervals at greatest
sample depth were included to determine the background (supported)
210
Pb activity. Unsupported 210Pb activity is expected to be negligible at
these depths, where sediment age is likely to exceed 150 years (N 6
half-lives).
Bulk and stable isotopic analysis of carbon and nitrogen
Bulk sediment samples were analyzed for weight percent organic C
(TOC) and total N (TN) concurrently with isotope ratio determinations.
All concentrations were determined on acid fumigated (decarbonated)
samples, eliminating variability in total C content caused by any carbonate within samples. N2 and CO2 peak areas (Isodat v3.0) were converted
to weight percent compositions using response factors generated from
standards of known composition (acetanilide, caffeine, B-2153, B-2159
and urea), which were run between every ten samples. Analytical precision, based on replicate standard runs for bulk measurements, was typically better than ±0.80% (1σ, n = 24) for carbon and nitrogen weight %
values and better than ± 0.20‰ for carbon and ± 0.25‰ for nitrogen
isotope values. Reproducibility between duplicate samples was better
than ± 0.05% for both carbon and nitrogen weight % values; isotopic
values of carbon and nitrogen between duplicate samples were better
than ±0.05‰ and ±0.09‰, respectively. Analyses were performed at
the Large Lakes Observatory (LLO) Stable Isotope Lab using a Costech
Elemental Analyzer coupled with a ThermoFinnigan DeltaPlusXP stable
isotope ratio monitoring mass spectrometer (EA-IRMS).
Lipid analysis for molecular biomarkers
The sediment sampling frequency for lipid analysis was half that for
bulk analysis (i.e., ten samples from each core were analyzed for molecular biomarkers). The total lipid extract (TLE) was obtained from sediments using a DIONEX Accelerated Solvent Extractor (ASE) 350, with
a solvent ratio of 9:1 dichloromethane/methanol (DCM/MeOH). TLEs
from core BH03-1 were separated into sub-fractions using Alltech®
aminopropyl bond elute columns. The neutral fraction, containing the
n-alkane biomarkers of interest, was eluted first with 9:1 DCM/MeOH.
After sulfur removal by activated (reduced) copper beads, TLEs from
core BH09-5 (not exceeding 5 mg in weight) and neutral fractions
from core BH03-1 were separated by activated (150 °C for 2 h)
alumina column chromatography using solvents of increasing polarity
(9:1 hexane/DCM, 2:1 DCM/MeOH and MeOH, for apolar, aromaticand
polar compounds, respectively). The n-alkanes were present in the
first (apolar) fraction; this fraction was next passed through a Ag+ impregnated silica pipette column to separate saturated and unsaturated
hydrocarbons. The saturated fraction was eluted first with n-hexane.
Each fraction was dried under a stream of N2 gas.
The saturated fraction was quantified using an Agilent 6890 GC
equipped with a flame ionization detector (FID). An internal standard
(androstane) was added to each sample prior to injection for quantification of individual n-alkanes. Identification of n-alkanes was accomplished using an Agilent 6890A Series GC System equipped with an
Agilent 5973 single quadrupole mass spectrometer (GC/MS).
Stable carbon isotope compositions of individual n-alkanes were determined with a ThermoFinnigan DeltaPlusXP isotope ratio mass spectrometer, using a modified GCC III interface (GC-C-IRMS). Calibration
was performed using a mixture of n-alkanes (C16 to C30, Schimmelman
“Mix A”) of known isotopic value, which was analyzed multiple times
daily. Based on these replicate measurements, typical precision of
individual δ13Cn-alkane measurements was better than ± 0.50‰ (1σ,
n = 24). Squalene was co-injected with each sample as an internal isotopic standard. Samples were run in duplicate, with reproducibility better than ± 0.50‰. In all samples the C17n-alkane was not present in
adequate abundance to obtain an accurate isotopic measurement. In
core BH09-5, the C19n-alkane in the first two samples co-eluted with another compound and those isotopic values have been excluded from the
data set. Values are expressed in conventional delta notation as per mil
(‰) deviations from the Vienna Pee Dee Belemnite (VPDB) standard.
Correction for the Suess Effect
To account for the change in the δ13C of atmospheric CO2 from anthropogenic fossil fuel burning (the Suess effect) over the past two centuries, it is necessary to correct bulk and molecular δ13C values.
According to Verburg (2007), lakes that are in isotopic equilibrium
with the atmosphere (e.g., lakes N 1000 km2Bade et al. (2004); Lake
Superior is ~ 82,000 km2) will reflect the isotopic signature in atmospheric CO2 and the Suess effect. Verburg (2007) additionally suggests
that the Suess correction be applied with caution in extremely heterotrophic lakes where the source of sedimentary organic matter is primarily allochthonous. Atilla et al. (2011) have shown that Lake Superior is in
equilibrium with CO2atm and that the lake, overall, is not respiring large
amounts of terrigenous CO2 (terrigenous carbon inputs do not exceed
17% of the total carbon budget; Zigah et al., 2011). Zigah et al. (2011)
also report bulk Δ14CDIC measurements in Lake Superior track
atmospheric radiocarbon levels, with a calculated residence time for
dissolved inorganic carbon (DIC) of ~ 3 years. These observations substantiate our application of the Suess correction to bulk and molecular
δ13C values in Lake Superior sediments. Therefore, all δ13C values were
corrected using the equation of Verburg (2007), as it encompasses the
entire time period of the two cores (2009 to 1700 AD) covered by our
age model.
M.D. O'Beirne et al. / Journal of Great Lakes Research 41 (2015) 20–29
The Pearson test (two-tailed) was conducted using SigmaPlot v. 11.0
to examine bivariate correlations among different paleoenvironmental
proxies. The α-level was set at 0.05.
23
decrease in rates around 1950. The MAR for the first 12 cm (2002–
1948 A.D.) in BH03-1 is 0.086 g/cm2/yr and 0.094 g/cm2/yr below
that. Core BH09-5 shows a mass sedimentation rate of 0.090 g/cm2/yr
for the top 14 cm (2008–1950 A.D.) and 0.110 g/cm2/yr below.
Results
Sedimentation rate and age model
Bulk geochemical data
The age–depth relationship of the two cores was estimated from
semi-log plots of excess 210Pb activity versus accumulated sediment
mass (Fig. 2) using methods described in Johnson et al. (2012). The
slopes of the straight-line segments below the uppermost centimeter
in the core are proportional to sediment mass accumulation rates
(MARs), which are applied to arrive at sediment age at each horizon.
A bioturbation zone is apparent in the uppermost cm, exhibits the
steepest slope observed (Edgington and Robbins, 1990; Robbins and
Edgington, 1975) and precludes application of a Constant Rate of Supply
(CRS) model to derive sediment age profiles in the core. The MAR in the
bioturbated zone is assumed to be constant and equal to that derived
from the slope of the line segment immediately below it. The 210Pb
data revealed a bioturbated zone of approximately one cm in each
core, which is equivalent to ~3 years of sedimentation.
The profiles of excess 210Pb vs. accumulated sediment mass are remarkably similar in the two cores (Fig. 2). The MARs in Lake Superior
have not been constant over the periods of depositional history recorded in these two cores. We approximate the age models by selecting two
intervals of constant MAR in each core, determined by a least squares
linear fit through the data in each interval (Fig. 2). While the profiles
could have been split into more than 2 straight-line segments to achieve
more refined age models, the improved temporal resolution, if any
more accurate, would not impact the conclusions of our study. The calculated mass accumulation rates in both cores are analogous, showing a
Patterns of accumulation of the bulk proxy data are similar in both
BH03-1 and BH09-5. Most notable are the increases in organic carbon
mass accumulation rate (TOCMAR) and total nitrogen mass accumulation rate (TNMAR) beginning in 1850, reaching a maximum in 1930
(Fig. 3). Values tend to decrease after 1930 showing an additional increase beginning after 1990. The Corg:N record tracks changes in both
the TOCMAR and TNMAR records, with the highest values occurring
when TOCMAR (Pearson r = 0.850, p b 0.0001, n = 28) and TNMAR (Pearson r = 0.458, p b 0.01, n = 28) are the highest, before 1940. Corg:N remain high after TOCMAR and TNMAR decrease after 1940, before
beginning a continuous steady decline since 1970.
The general pattern observed in bulk sediment carbon isotopic
values (δ13Corg) for both cores is a period of 13C-enrichment from
1900 to 1960 followed by a leveling off and slight decline of δ13Corg
values in the 1970s (Fig. 3). The most striking increase is observed between 1900 and 1960, where δ13Corg values increase almost 2‰, from
b−26‰ to −24.5‰. A second period of increasing δ13Corg begins post
1990, with the most enriched δ13C signatures occurring at the top of
the cores. Unlike δ13Corg values, δ15N values start their increase later,
ca. 1925, with values increasing from 1.5‰ to 3‰ by 1975. δ15N values
also decrease more significantly from 1980 to 2000. Similar to the earlier record, the δ15N values again increase but later than δ13C beginning in
2000 and increasing to the top of the core, following the observed
trends in TOC and TN MARs.
Fig. 2. Left: excess 210Pb vs. accumulated sediment mass in each core. The sediment mass accumulation rates (MARs) presented correspond to the linear segments that were fitted to the
data illustrated in the linear portions of the two cores from western Lake Superior. Right: age vs. depth plots with associated error.
24
M.D. O'Beirne et al. / Journal of Great Lakes Research 41 (2015) 20–29
Fig. 3. Temporal variation in bulk organic geochemical proxies within two corresponding cores from western Lake Superior. Shaded areas indicate periods of inferred enhanced productivity based on the multiple proxies.
Molecular biomarkers
The flux of both aquatic (C17 + C19 + C21) and terrigenous (C27 +
C29 + C31) n-alkane biomarkers show an overall increasing trend up
to 1925, after which values decrease until 1975 and level off thereafter;
with a minor increase around 2000 (Fig. 5). Variations in the weighted
isotopic averages of both aquatic (i.e., δ13C19 and δ13C21 expressed as
δ13Caquatic) and terrigenous (i.e., δ13C27, δ13C29 and δ13C31, expressed as
δ13Cterrigenous) n-alkane biomarkers are used to examine changes in
primary productivity as observed in the bulk δ13Corg record (Fig. 5).
δ13Caquatic values display an increase (−34‰ to −30‰) from 1825 to
1960, decreasing slightly thereafter, before increasing again in the mid1990s. δ13Cterrigenous values show a decreasing trend between 1825 and
1925, after which values remain steady. Generally, aquatic biomarker
δ13C values fluctuate between −35‰ and −30‰, while terrigenous biomarker δ13C values fluctuate between −33‰ and −30‰ (Fig. 6).
Discussion
Effects of diagenesis on the sedimentary geochemical record
It is important when using organic geochemical indicators to reconstruct the recent past (the period in which degradation rates are the
greatest) to consider the role of diagenesis in altering the sedimentary
geochemical record. Organic carbon in excess of that which can be attributed to either steady-state degradation (Li et al., 2012) or changing
sediment MARs is present in the two cores deposited since 1900. This
conclusion is supported by downcore increases in TOCMAR profiles and
elemental Corg:N ratios in sections of the two cores (Fig. 3), which are
not typical of diagenetic profiles present in Lake Superior sediments
(cf. Li et al., 2012). Our data indicate that TOCMAR is a reliable indicator
of productivity changes in these cores, especially when used in conjunction with other proxies (i.e., δ13Corg, δ15N and Corg:N).
Period of 13C-enrichment (1900–1970)
The Lake Superior watershed witnessed its first European settlements in the early 1800s, with the beginning of the fur trade.
Subsequent population increases occurred with the further developments of logging, mining, shippingand processing of raw materials,
until the Duluth-Superior Harbor region reached its historic maximum
of 100,000 people in 1922. A series of water quality studies, conducted
by the United States Geological Survey (USGS) and Minnesota Board of
Health in the 20s, 40sand then 60s concluded that the water quality
within the St. Louis River had steadily declined with regional development (U.S. Department of the Interior, 1969 and references therein).
The cumulative impacts of this cultural and economic development
eventually resulted in the International Joint Commission (IJC) identifying the region as an Area of Concern under the Great Lakes Water Quality
Agreement of 1987 because of multiple beneficial use impairments.
We use the time period, 1800–1900, as a baseline on which to compare later changes in the bulk proxy data. Prior to 1900, we expect little
anthropogenic disturbance in the watershed as European settlement
and development of the region was slow. Prior to 1900, both cores exhibit δ13Corg, δ15Nand Corg:N values (Figs. 3 and 4) that indicate organic
M.D. O'Beirne et al. / Journal of Great Lakes Research 41 (2015) 20–29
25
matter derived primarily from aquatic algae growing within a cold, oligotrophic environment (Meyers and Ishiwatari, 1993), such as Lake
Superior—although a slight increase in TOCMAR and Corg:N could be related to logging in the region. The mass accumulation rates of TOC and
TN differ little between the two cores and appear to be highly correlated
temporally, suggesting dominant controls from in-lake primary production. This inference is supported by Corg:N ratios, with values falling
mostly near the lacustrine algae range rather than within C3 or C4 land
plants (Meyers, 1994; Fig. 4), although an increasing contribution
from terrigenous C3 land plants to the sediment organic matter pool is
observed between 1900 and 1970.
We interpret the observed increases in δ13Corg and δ13Caquatic values
between 1900 and 1960 to reflect a period of enhanced lacustrine productivity (Figs. 3 and 5). Increasing δ13Corg values are consistent with increasing lacustrine primary production from 1900 to 1960 and coincide
with increased anthropogenic disturbances during the development of
the St. Louis River watershed. The observed 13C-enrichment of sedimentary organic matter could also be due to the preferential mineralization
of isotopically light carbon compounds during early diagenesis. This
possibility seems unlikely; in low organic content (b2–3%) sediments
such as those in Lake Superior, isotopic discrimination associated with
mineralization appears to have a minor influence on the δ13Corg(Meyers,
1994). Additionally, Hodell and Schelske (1998) found that in Lake Ontario post-burial diagenesis reduced the mass of organic carbon buried
within sediment cores taken 6 years apart but did not change the
Fig. 5. Temporal variation in the flux of both aquatic and terrigenous n-alkane biomarkers
within two cores from western Lake Superior. Shaded areas indicate periods of inferred
enhanced productivity based on the multiple proxies.
Fig. 4. Cross-plot of atomic Corg:N and δ13Corg from two corresponding multicores. Data
from the two cores demonstrate the relative source contributions to bulk sediment organic matter in western Lake Superior. Top: time periods of differing source contributions are
noted; magnified version of bottom plot. Bottom: generalized atomic Corg:N ratios and
δ13Corg values of major sources of organic matter to lake sediments (after Meyers, 1994).
δ13Corg values; a condition which we also observe in the two cores
from western Lake Superior presented in this study (Fig. 3).
Carbon isotopic values of algal-derived molecules (δ13Caquatic)
should not be influenced by terrigenous carbon inputs, and therefore
they can provide a better indication of trophic state than that of bulk
δ13Corg and n-alkane abundances (Meyers, 2003). The trends in the
δ13Caquatic signature agree well with the bulk organic carbon
δ13Corgsignature, suggesting that aquatic productivity drives the isotopic composition of the sedimentary geochemical record (Fig. 6). Absolute n-alkane abundances for both aquatic and terrigenous sources
show similar trends down core. Increased fluxes from 1920 to 1950 correspond with increases in aquatic productivity and transport of terrigenous material to the lake, consistent with the development of the
watershed (Fig. 5).
The variability in the δ15N record and the temporal shifts in both directions—especially evident in core BH03-1(Fig. 3)—indicate that multiple sources of N with differing isotope compositions are influencing the
26
M.D. O'Beirne et al. / Journal of Great Lakes Research 41 (2015) 20–29
Fig. 6. Comparison of temporal variations in molecular δ13Caquatic, δ13Cterrigenous, and bulk δ13Corg values between two cores from western Lake Superior. Shaded areas indicate periods of
inferred enhanced productivity based on the multiple proxies.
final δ15N signature preserved in the sediment core records. The more
rapid temporal variations between 1900 and 1970 are likely due to watershed disturbances causing the export of varying amounts of soil N to
this part of the lake (soil NO−
3 : 3–12‰, Hodell and Schelske, 1998). This
interpretation is supported by increases in the Corg:N ratio between
1900 and 1970 (Fig. 4). However, the most plausible explanation for
the overall long term trend of increasing δ15N between 1900 and 1970
in the two cores is the gradual increase in primary production, as evidenced by increasing δ13Corg and δ13Caquatic values during the same
time period (Fig. 6).
The observation that δ15N starts increasing after the observed increase in productivity (increases in TOCMAR, TNMAR and δ13Corg) indicates that nitrogen limitation is not the factor determining the amount
of primary production (i.e., δ13Corg begins increasing around 1900,
whereas δ15N increases starting at ~1920; Fig. 3). Thus, some other limiting nutrient must be responsible for the increases in primary production from 1900 to 1960. Historical measurements of phosphorus within
Lake Superior are nonexistent until the 1970s; therefore, we can only
surmise that the nutrient responsible for the historical increase in productivity is phosphorus.
P-availability is known to control productivity within presentday Lake Superior with Fe playing a co-limiting role at least during
artificially stimulated primary production (Ivanikova et al., 2007;
Sterner et al., 2007). Schelske et al. (2006) offer evidence of anthropogenically driven eutrophication in western Lake Superior,
reporting enhanced biogenic silica accumulation between 1900 and
1970 from which they infer increased historic phosphorus supply
(diatom growth in Lake Superior is phosphorous limited; Schelske
et al., 1986). This is consistent with data presented by Munawar
and Munawar (1978) who report the dominant phytoplankton species in the western arm of Lake Superior during 1973 was the diatom
species Melosira granulata (now identified as Aulacoseira granulata),
which is often associated with eutrophication/perturbed environments (Holland, 1968). The observed increases in productivity presented in the current study are consistent with increased
anthropogenic phosphorus loading beginning in 1900, which also
correlates well with the increases in productivity, in response to
P loading, reported in Lakes Erie and Ontario during the same time
period (Hodell and Schelske, 1998; Lu and Meyers, 2009; Lu et al.,
2010; Schelske and Hodell, 1995).
M.D. O'Beirne et al. / Journal of Great Lakes Research 41 (2015) 20–29
Lake Superior is the largest managed freshwater body in the world. It
has been regulated (levels and flow) through the International Joint
Commission (IJC) since the signing of the Boundary Waters Treaty in
1909. It is clear from observations of the sedimentary geochemical record that anthropogenic disturbances correlated with increasing
human populations led to increased lacustrine productivity, most likely
driven by P loading, during the major developmental period of the watershed, at least in the Duluth-Superior harbor region.
Declines in proxy data beginning in 1925
Decreases in MARs of organic C (with no accompanying change in
δ13Corg) and total N and fluxes of both aquatic and terrigenous n-alkane
biomarkers began post 1940, years before the strict regulatory P control
actions of the mid-1970s as required after the signing of the Great Lakes
Water Quality Agreement in 1972. Three large hydropower dams were
built on the St. Louis River between 1900 and 1925 (Fig. 1); the Thomson Water Project (a series of dams) in 1907, the Scanlon Dam in 1922
and the Fond du Lac Dam in 1924. During this same time period, the
lumber industry began its decline due to depletion of high value lumber
products from the forests and increased competition; only one of
fifteen original lumber mills was still operating in 1925 (White and
Mladendoff, 2004; Minnesota Historical Society). The construction of
these dams and the decrease in lumbering are the most plausible causes
of the observed declines in TOCMAR and TNMAR in both cores, as well as
the delivery of aquatic and terrigenous n-alkane biomarkers in BH09-5
and to a lesser extent in BH03-1(Figs. 3 and 5). Corg:N ratios show that
terrigenous contribution to the sedimentary record is still influential
in this time period as values trend towards C3 land plants between
1900 and 1970 (Fig. 4).
An additional, although comparatively short, decrease in TOCMAR
and TNMAR is observed between 1975 and 1990 (Fig. 3). In 1971,Minnesota State Legislature approved the creation of the Western
Lake Superior Sanitary District (WLSSD), which then became operational in 1978. This could account for the observed decrease in sediment
organic matter, as the tertiary plant combined virtually all of the
Minnesota industrial and municipal discharges in the area and reduced
anthropogenic P loading to the western end of the lake. Decreases in
Corg:N ratios (Fig. 4) indicate that algal derived organic matter becomes
a larger component of sediment organic matter around 1975, even
though δ13Corg shows little variation (Fig. 3).
During this same time period (1975–1990), δ15N values decrease
while TNMAR remains relatively stable (Fig. 3), which is counter to
expectation if aquatic productivity alone is driving the observed δ15N
signal. Therefore, this period of 15N-depletion is a likely reflection
of continually increasing atmospheric deposition of isotopically light
N to the lake from anthropogenic fossil fuel combustion and synthetic
fertilizers that is subsequently incorporated during primary production.
The contribution of fertilizers to the depletion of δ15N is unlikely to be of
significance in Lake Superior, given the small watershed with little
agricultural development. Large increases in NOx emissions were
observed in the United States (EPA, 2000) and Canada(Chen et al.,
2000) between 1940 and 1980, after which emissions remain relatively
constant. Increasing δ15N values after 1990 in the cores may be a
response to constant NOx emissions, whereby δ15N values begin to reflect the influence of primary production rather than anthropogenic
light N inputs.
The latest increase in productivity (1990–present)
The latest increase in δ13Corg values, beginning near 1990, is enigmatic but may be attributable to increasing water temperatures and
length of the stratified season. Magnuson et al. (1997) predicted that
increased epilimnetic temperatures are unlikely to affect rates of phytoplankton primary production directly, but increases in annual production should be expected within the Laurentian Great Lakes because of
27
longer open water growing seasons. This phenomenon has been observed in the history of Lake Baikal, a high-latitude large lake similar
to Lake Superior (Hampton et al., 2008; Moore et al., 2009). In contrast,
Lehman (2002) predicted a longer stratified season causing nutrient
limitation, resulting in a reduction in rates of productivity within Lake
Superior, which is already the most oligotrophic of the Laurentian
Great Lakes. Lake Tanganyika in East Africa has experienced an enhanced thermal gradient between the epilimnion and hypolimnion,
resulting in inefficient mixing (nutrient turnover) of the two layers
and a reduction in primary production (Verburg et al., 2003; O'Reilly
et al., 2003; Tierney et al., 2010). The length of ice-free season has
been predicted to be proportional to phytoplankton production in the
Canadian Shield Lakes (Fee et al., 1992). Lake Superior was noted as
an exception to this due to its large areas of open water during most
winters. However, Lake Superior is not completely ice-free for 12
months of the year, and a shift to longer ice-free and summer stratified
periods has been noted in the past 20 years (Austin and Colman, 2007,
2008). Quay et al. (1986) showed that the δ13C of dissolved inorganic
C increased as the stratified season lengthened in Lake Washington
which would be reflected in the δ13Corg photosynthetically produced
from the DIC. A recent three-dimensional model of Lake Superior, including dynamic and thermodynamic ice cover, shows an increasing
trend in total annual primary production (for the simulation period of
1985–2008) that correlates with decreasing winter ice cover, as well
as increasing annual average surface water temperatures (White et al.,
2012). Our results, showing increasing bulk organic C and N content
within sediments as well as enriched δ13Corg and δ13Caquatic values, are
consistent with the modeled results and the results of Quay et al.
(1986).
Implications
Although our data clearly indicate that anthropogenic disturbances
in the St. Louis River watershed impacted lacustrine productivity within
the western arm of Lake Superior, we do not assume this to be the case
for the whole of Lake Superior due to its large size and increasing depth
off shore. In general, we would expect impacts such as we observe to be
limited to relatively shallow, near shore areas in large lakes, where the
loading of nutrients and productivity would be highest. Thus, these
are constrained to the western arm of Lake Superior, the area most
heavily influenced by disturbances in the watershed of the second largest inflow (i.e., the St. Louis River) to Lake Superior. However, we note
that the human impacts described could potentially influence the rest
of Lake Superior, or any lake system, given a large enough anthropogenic input relative to lake size.
On the other hand, the influence of increased atmospheric deposition on the nitrogen record (i.e., 15N-depletion) is likely to be a lake
wide occurrence. Lake Superior has the largest surface area for any
lake in the world, and due to the fact that atmospheric deposition
isnot localized, we would expect lake wide incorporation of atmospherically deposited, isotopically light N during primary production. The observation of 15N-depletion, as recorded in the sediments, and attribution
to atmospheric sources is one of the first of its kind for Lake Superior and
large lakes in general, but has been previously observed in many small,
remote lakes (Holtgrieve et al., 2011). While atmospheric deposition of
isotopically light N is the most plausible scenario relating to the N record, more detailed studies are needed in order to confirm our interpretation. Our observations regarding the most recent increases in
productivity post 1990 are the first direct sedimentary evidence of a recent trend of increasing productivity within Lake Superior. This trend is
consistent with increasing temperatures stimulating primary productivity through a positive feedback related to the length of the ice-free
period, although further investigation should be undertaken to confirm
this inference and whether or not it is a lake wide or localized
phenomenon.
28
M.D. O'Beirne et al. / Journal of Great Lakes Research 41 (2015) 20–29
Conclusions
Sediments of western Lake Superior provide a historic record of
changes in the source and flux of sedimentary organic matter; these,
in turn, provide a record of historical ecosystem disturbances/changes
within the western arm of the lake and its watershed. Early settlement
and subsequent industrial development produced an increase in aquatic
productivity, most likely driven by P loading associated with urbanization of the watershed. Increases in aquatic primary productivity are
marked by enhanced organic carbon delivery to the sediment as well
as the enrichment of δ13Corg values of sedimentary organic matter.
Further evidence of aquatic productivity being the major influence driving the changes seen in δ13Corg is observed in the increase of aquatic biomarker δ13Caquatic values during the same time period. The most recent
decrease in bulk δ13Corg, δ15N and Corg:N values may have begun as early
as 1925 and were largely accomplished prior to establishment of
WLSSD. Nevertheless, δ13Corg and δ15N show their strongest decrease
after 1970, which indicates that improved waste collection and treatment did reduce anthropogenic P loading, thereby reducing algal productivity. Observations of recent increased primary productivity are
enigmatic but may be attributable to increasing water temperatures
stimulating primary productivity through a positive feedback related
to the length of the ice-free period.
Acknowledgements
Support for this research was provided by Minnesota Sea Grant to
RLH, TCJ and JPW. The 210Pb analyses were carried out by alpha spectrometry in the Department of Soil Science, University of Manitoba,
under the direction of Dr. Paul Wilkinson. The authors give thanks to
the following: the crew of the R/V Blue Heron for core collection,
Sarah Grosshuesch and Koushik Dutta for their laboratory expertise,
Julia Halbur for help with sample preparation and Jillian Votava. The authors also thank the associate editor and two anonymous reviewers
whose valuable comments considerably improved this manuscript.
References
Atilla, N., McKinley, G.A., Bennington, V., Baehr, M., Urban, N., DeGrandpre, M., Desai, A.R.,
Wu, C., 2011. Observed variability of Lake Superior pCO2. Limnol. Oceanogr. 56,
775–786.
Austin, J.A., Colman, S., 2007. Lake Superior summer water temperatures are increasing
more rapidly than regional air temperatures: a positive ice-albedo feedback. J.
Geophys. Lett. 34.
Austin, J.A., Colman, S., 2008. A century of temperature variability in Lake Superior.
Limnol. Oceanogr. 53, 2724–2730.
Bade, D.L., Carpenter, S.R., Cole, J.J., Hanson, P.C., Hesslein, R.H., 2004. Controls of δ13C-DIC
in lakes: geochemistry, lake metabolism, and morphometry. Limnol. Oceanogr. 49,
1160–1172.
Bourbonniere, J.A., Meyers, P.A., 1996. Sedimentary geolipid records of historical changes
in watersheds and productivities in Lakes Ontario and Erie. Limnol. Oceanogr. 41,
352–359.
Chen, J., Chen, W., Liu, J., Cihlar, J., Gray, S., 2000. Annual Carbon Balance of Canada's
forests during 1895–1996. Glob. Biogeochem. Cycles 14, 839–850.
Cranwell, P.A., 1981. Diagenesis of free and bound lipids in terrestrial detritus deposited
in a lacustrine sediment. Org. Geochem. 3, 79–89.
Cranwell, P.A., Eglinton, G., Robinson, N., 1987. Lipids of aquatic organisms as potential
contributors to lacustrine sediments-II. Org. Geochem. 11, 513–527.
Edgington, D.N., Robbins, J.A., 1990. Time scales of sediment focusing in large lakes as revealed by measurement of fallout Cs-137. In: Tilzer, M., Serruya, S. (Eds.), Large Lakes:
Ecological Structure and Function. Brock/Springer Series in Contemporary Bioscience.
Springer Verlag, Berlin, pp. 210–223.
Eglinton, G., Hamilton, R.J., 1963. The distribution of alkanes. In: Swain, T. (Ed.), Chemical
Plant Taxonomy. Academic Press.
Eglinton, G., Hamilton, R.J., 1967. Leaf epicuticular waxes. Science 156, 1322–1335.
Environmental Protection Agency (EPA), 2000. National air pollution emission trends,
1900–1998. Rep. EPA-454/R-00-002, (238 pp., Washington D.C.).
Fee, E.J., Shearer, J.A., DeBruyn, E.R., Schindler, E.U., 1992. Effects of lake size on phytoplankton photosynthesis. Can. J. Fish. Aquat. Sci. 49, 2445–2459.
Giger, W., Schaffner, C., Wakeham, S.G., 1980. Aliphatic and olefinic hydrocarbons in recent sediments of Greifensee, Switzerland. Geochim. Cosmochim. Acta 44, 119–129.
Hampton, S.E., Izmest'eva, L.R., Moore, M.V., Katz, S.L., Dennis, B., Silow, E.A., 2008. Sixty
years of environmental change in the world's largest freshwater lake—Lake Baikal,
Siberia. Glob. Chang. Biol. 14, 1–12.
Hodell, D.A., Schelske, C.L., 1998. Production, sedimentation, and isotopic composition of
organic matter in Lake Ontario. Limnol. Oceanogr. 43, 200–214.
Holland, R.E., 1968. Correlation of Melosira species with trophic conditions in Lake
Michigan. Limnol. Oceanogr. 13, 555–557.
Holtgrieve, G.W., Schindler, D.E., Hobbs, W.O., Leavitt, P.R., Ward, E.J., Bunting, L., Chen, G.,
Finney, B.P., Gregory-Eaves, I., Holmgren, S., Lisac, M.J., Lisi, P.J., Nydick, K., Rogers, L.A.,
Saros, J.E., Selbie, D.T., Shapley, M.D., Walsh, P.B., Wolfe, A.P., 2011. A coherent signature of anthropogenic nitrogen deposition to remote watersheds of the Northern
Hemisphere. Science 334, 1545–1548.
Ivanikova, N.V., McKay, R.M.L., Bullerjahn, G.S., Sterner, R.W., 2007. Nitrate utilization by phytoplankton in Lake Superior is impaired by low nutrient (P, Fe) availability and seasonal light limitation—a cyanobacterial bioreporter study. J.
Phycol. 43, 475–484.
Johnson, T.C., Van Alstine, J.D., Rolfhus, K.R., Colman, S.M., Wattrus, N.J., 2012. A high resolution study of spatial and temporal variability of natural and anthropogenic compounds in offshore Lake Superior sediments. J. Great Lakes Res. 38, 673–685.
Lehman, J.T., 2002. Mixing patterns and plankton biomass of the St. Lawrence Great Lakes
under climate change scenarios. J. Great Lakes Res. 28, 213–250.
Li, J., Crowe, S.A., Miklesh, D., Kistner, M., Canfield, D.E., Katsev, S., 2012. Carbon mineralization and oxygen dynamics in sediments with deep oxygen penetration, Lake
Superior. Limnol. Oceanogr. 57, 1634–1650.
Lu, Y., Meyers, P.A., 2009. Sediment lipid biomarkers as recorders of the contamination
and cultural eutrophication of Lake Erie, 1909–2003. Org. Geochem. 912–921.
Lu, Y., Meyers, P.A., Eadie, B.J., Robbins, J.A., 2010. Carbon cycling in Lake Erie during cultural eutrophication over the last century inferred from the stable carbon isotope
composition of sediments. J. Paleolimnol. 43, 261–272.
Magnuson, J.J., Webster, K.E., Assel, R.A., Bowser, C.J., Dillon, P.J., Eaton, J.G., Evans, H.E.,
Fee, E.J., Hall, R.I., Mortsch, L.R., Schindler, D.W., Quinn, F.H., 1997. Potential effects
of climate changes on aquatic systems: Laurentian Great Lakes and precambrian
shield region. Hydrol. Process. 11, 825–871.
McKenzie, A.J., 1985. Carbon Isotopes and Productivity in the Lacustrine and Marine
Environment. In: Stumm, Werner (Ed.), Chemical Processes in Lakes. WileyInterscience, New York, pp. 99–118.
Meyers, P.A., 1994. Preservation of elemental and isotopic source identification of sedimentary organic matter during and after deposition. Chem. Geol. 144, 289–302.
Meyers, P.A., 1997. Organic geochemical proxies of paleoocenagraphic, paleolimnologic,
and paleoclimatic processes. Org. Geochem. 27, 213–250.
Meyers, P.A., 2003. Applications of organic geochemistry to paleolimnological
reconstructions: a summary of examples from the Laurentian Great Lakes. Org.
Geochem. 34, 261–289.
Meyers, P.A., Ishiwatari, R., 1993. Lacustrine organic geochemistry—an overview of indicators
of organic matter sources and diagenesis in lake sediments. Org. Geochem. 20, 867–900.
Moore, M.V., Hampton, S.E., Izmest'eva, L.R., Silow, E.A., Peshkova, E.V., Pavlov, B.K., 2009. Climate change and the World's “Sacred Sea”—Lake Baikal, Siberia. Bioscience 59, 405–417.
Munawar, M., Munawar, I.F., 1978. Phytoplankton of Lake Superior 1973. J. Great Lakes
Res. 4, 415–442.
O'Reilly, C.M., Alin, S.R., Plisnier, P.D., Cohen, A.S., Mckee, B.A., 2003. Climate change decreases
aquatic ecosystem productivity of Lake Tanganyika, Africa. Nature 424, 766–768.
Quay, P.D., Emerson, S.R., Quay, B.M., Devol, A.H., 1986. The carbon cycle for Lake
Washington-A stable isotope study. Limnol. Oceanogr. 31, 596–611.
Rieley, G., Collier, R.J., Jones, D.M., Eglinton, G., Eakin, P.A., Fallick, P.A., 1991. Sources of
sedimentary lipids deduced from stable isotope analyses of individual compounds.
Nature 352, 425–427.
Robbins, J.A., Edgington, D.N., 1975. Determination of recent sedimentation rates in Lake
Michigan using Pb-210 and Cs-137. Geochim. Cosmochim. Acta 39, 285–304.
Schelske, C.L., Hodell, D.A., 1991. Recent changes in productivity and climate of Lake
Ontario detected by isotopic analysis of sediments. Limnol. Oceanogr. 36, 961–975.
Schelske, C.L., Hodell, D.A., 1995. Using carbon isotopes of bulk sedimentary organic matter to reconstruct the history of nutrient loading and eutrophication in Lake Erie.
Limnol. Oceanogr. 40, 918–929.
Schelske, C.L., Fahnenstiel, G.L., Haibach, M., 1986. Phosphorus enrichment, silica
utilization, and biogeochemical silica depletion in the Great Lakes. Can. J. Fish.
Aquat. Sci. 43, 407–415.
Schelske, C.L., Stoermer, E.F., Kenney, W.F., 2006. Historic low-level phosphorous enrichment in the Great Lakes inferred from biogenic silica accumulation in sediments.
Limnol. Oceanogr. 51, 728–748.
Sterner, R.W., Anagnostou, E., Brovold, S., Bullerjahn, G.S., Findlay, J.C., Kumar, S., McKay,
R.M.L., Sherrel, R.M., 2007. Increasing stoichiometric imbalance in North America's
largest lake: nitrification in Lake Superior. Geophys. Res. Lett. 34.
Tierney, J.E., Mayes, M.T., Meyer, N., Johnson, C., Swarzenski, P.W., Cohen, A.S., Russell, J.M.,
2010. Late-twentieth-century warming in Lake Tanganyika unprecedented since AD
500. Nat. Geosci. 3, 422–425.
U.S. Department of the Interior, 1969. Federal Water Pollution Control Administration—Great Lakes Region. An Appraisal of Water Pollution in the Lake Superior
Basin.
Urban, N.R., 2007. Nutrient cycling in Lake Superior: a retrospective and update. In:
Munawar, M., Heath, R. (Eds.), The State of Lake Superior. Ecovision World Monograph Series.
Urban, N.R., Auer, M.T., Green, S.A., Lu, X., Apul, D.S., Powell, K.D., Bub, L., 2005. Carbon
cycling in Lake Superior. J. Geophys. Res. 110, C06S90. http://dx.doi.org/10.1029/
2003JC002230.
Verburg, P., 2007. The need to correct for the Suess effect in the application of d13C in sediment in autotrophic Lake Tanganyika, as a productivity proxy in the anthropocene. J.
Paleolimnol. 37, 591–602.
Verburg, P., Hecky, R.E., Klins, H., 2003. Ecological consequences of a century of warming
in Lake Tanganyika. Science 301, 505–507.
M.D. O'Beirne et al. / Journal of Great Lakes Research 41 (2015) 20–29
Vollenweider, R.A., Munawar, M., Stadelmann, P., 1974. A comparative review of phytoplankton and primary production in the Laurentian Great Lakes. J. Fish. Res. Board
Can. 31, 739–762.
Weiler, R.R., 1978. The chemistry of Lake Superior. J. Great Lakes Res. 4, 370–385.
White, M.A., Mladendoff, D.J., 2004. Old-growth forest landscape transitions from preEuropean settlement to present. Landsc. Ecol. 9, 191–205.
29
White, B., Austin, J., Matsumoto, K., 2012. A three-dimensional model of Lake Superior
with ice and biogeochemistry. J. Great Lakes Res. 38, 61–67.
Zigah, P.K., Minor, E.C., Werne, J.P., McCallister, S.L., 2011. Radiocarbon and stable isotopic
insights into provenance and cycling of carbon in Lake Superior. Limnol. Oceanogr.
56, 867–886.