Geochimica et Cosmochimica Acta, Vol. 68, No. 3, pp. 477– 489, 2004
Copyright © 2004 Elsevier Ltd
Printed in the USA. All rights reserved
0016-7037/04 $30.00 ⫹ .00
Pergamon
doi:10.1016/S0016-7037(03)00458-7
Increased concentrations of dissolved trace metals and organic carbon during snowmelt in
rivers of the Alaskan Arctic
ROBERT D. REMBER* and JOHN H. TREFRY
Department of Marine and Environmental Systems, Florida Institute of Technology, Melbourne, FL 32901, USA
(Received December 20, 2002; accepted in revised form July 1, 2003)
Abstract—Arctic rivers typically transport more than half of their annual amounts of water and suspended
sediments during spring floods. In this study, the Sagavanirktok, Kuparuk and Colville rivers in the Alaskan
Arctic were sampled during the spring floods of 2001 to determine levels of total suspended solids (TSS) and
dissolved and particulate metals and organic carbon. Concentrations of dissolved organic carbon (DOC)
increased from 167 to 742 ␮mol/L during peak discharge in the Sagavanirktok River, at about the same time
that river flow increased to maximum levels. Concentrations of dissolved Cu, Pb, Zn and Fe in the
Sagavanirktok River followed trends observed for DOC with 3- to 25-fold higher levels at peak flow than
during off-peak discharge. Similar patterns were found for the Kuparuk and Colville rivers, where average
concentrations of dissolved trace metals and DOC were even higher. These observations are linked to a large
pulse of DOC and dissolved metals incorporated into snowmelt from thawing ponds and upper soil layers. In
contrast with Cu, Fe, Pb and Zn, concentrations of dissolved Ba did not increase in response to increased
discharge of water, TSS and DOC. Concentrations of particulate Cu, Fe, Pb and Zn were more uniform than
observed for their respective dissolved species and correlated well with the Al content of the suspended
particles. However, concentrations of particulate Al were poorly correlated with particulate organic carbon.
Results from this study show that ⬎80% of the suspended sediment and more than one-third of the annual
inputs of dissolved Cu, Fe, Pb, Zn and DOC were carried to the coastal Beaufort Sea in 3 and 12 d,
respectively, by the Kuparuk and Sagavanirktok rivers. Copyright © 2004 Elsevier Ltd
Russian Arctic are 1.5 to 3 times lower than concentrations in
other major, non-arctic rivers (Martin et al., 1993). Such results
suggest that rivers in the Russian Arctic are less influenced by
chemical weathering or anthropogenic contamination.
In contrast to the Russian Arctic, very few studies of trace
metals have been carried out for rivers in the Alaskan Arctic.
Instead, geochemical research in the Alaskan Arctic has concentrated on TSS, major elements, nutrients and DOC in rivers
as well as streams, lakes and ponds (Brown et al., 1962; Lock
et al., 1989; Telang et al., 1991; Kling et al., 1992; Kriet et al.,
1992). Similar to the studies in Russia, sampling in the Alaskan
Arctic has focused on the open water period from July to
September.
A large gap exists in our knowledge of riverine concentrations and transport of trace metals and organic carbon in the
Alaskan Arctic, especially during peak discharge in the spring.
In this investigation, concentrations of dissolved and particulate
Ba, Cu, Fe, Pb, Zn and organic carbon in the Sagavanirktok,
Kuparuk and Colville rivers of arctic Alaska were determined
throughout the spring thaw and again during the summer (Fig.
1). The primary sites occupied during this investigation were
near river mouths to provide representative values for the
composition of water flowing into the coastal Beaufort Sea.
1. INTRODUCTION
Most rivers that drain into the Arctic Ocean carry 40 to 80%
of their annual volume of water during the spring floods (Arnborg et al., 1967; Gordeev et al., 1996). In addition to water
discharge, Telang (1985) showed that ⬃35% of the annual
discharge of dissolved organic carbon (DOC) into the Canadian
Beaufort Sea from the Mackenzie River occurred during the
June snowmelt. Similar results have been reported for the Lena
River where concentrations of DOC decreased from ⬎1000
␮mol/L during spring floods in June to 600 to 700 ␮mol/L
during September (Cauwet and Sidorov, 1996). Studies in the
Alaskan Arctic have shown that concentrations of total suspended solids (TSS) in the Colville River follow the same trend
as water flow with ⬎70% of the annual discharge of TSS
occurring during June (Arnborg et al., 1967). Large seasonal
discharges of water at high latitudes, linked with increased
concentrations of TSS and DOC, emphasize the importance of
spring floods to the arctic hydrologic cycle.
Few studies in the Arctic have investigated trends in river
transport of trace metals during the spring floods. Data for
dissolved and particulate trace metals in the Arctic are available
for some rivers in Russia; however, these investigations have
focused on the open water period from July to September rather
than the spring floods when water discharge and concentrations
of TSS and DOC peak (Martin et al., 1993; Dai and Martin,
1995; Guieu et al., 1996; Zhulidov et al., 1997; Moran and
Woods, 1997). Reported concentrations of dissolved and particulate trace metals in wetlands, rivers and estuaries of the
* Author to whom
([email protected]).
correspondence
should
be
2. STUDY AREA
The Sagavanirktok, Kuparuk and Colville rivers lie within
the Arctic climatic zone where annual temperatures average
–12°C and mean precipitation is ⬃12 cm/yr⫺1 (Fig. 1) (Telang
et al., 1991). The drainage basins of Arctic rivers in Alaska
include the following three physiographic provinces: the Arctic
Mountain Province, the Arctic Foothills Province, and the
addressed
477
478
R. D. Rember and J. H. Trefry
Table 1. Locations of sampling sites and U.S. Geological Survey
(USGS) gauges.
Site
Sagavanirktok River
(near Prudhoe Bay)
Sagavanirktok River
Sagavanirktok River
(USGS gauge)
Kuparuk River
Kuparuk River
(USGS gauge)
Colville River
Fig. 1. Map showing sampling locations for the Sagavanirktok,
Kuparuk and Colville rivers in the Alaskan Arctic. The northernmost
stations on the Sagavanirktok, Kuparuk and Colville rivers are the
primary sampling sites. The southernmost location marked on the
Sagavanirktok River is at the site of the U.S. Geologic Survey gauging
station. Snow samples were collected from the station located ⬃30 km
upstream from the primary sampling site on the Sagavanirktok River.
Box in inset map identifies the study area in northeastern Alaska.
Arctic Coastal Plain Province (Payne et al., 1951). Based on the
classification scheme proposed by Craig and McCart (1975),
the Sagavanirktok and Colville rivers can be classified as
mountain streams that drain snowfields and glaciers in the
Brooks Range. In contrast, the yellow-colored water of the
Kuparuk River is more representative of a tundra stream (Lock
et al., 1989). All three rivers flow into the coastal Beaufort Sea.
The geology of the region was summarized by Payne et al.
(1951). Marine and non-marine sediments including peat deposits from the Quaternary underlie the coastal plain. Tertiary
deposits from the Sagavanirktok Formation (limestone, chert)
are exposed on the coastal plain in the Kuparuk and Sagavanirktok basins. Shales from the Triassic to Cretaceous are
exposed in the Foothills Province and Brooks Range. The
Lisburne limestone and dolomite group from the Mississippian
and Pennsylvanian ages are concentrated in the eastern Brooks
Range within the drainage basin of the Sagavanirktok River
(Payne et al., 1951). Previous studies have shown that the
Sagavanirktok River drains primarily limestone deposits and
has concentrations of dissolved Ca that are ⬃2 times higher
than in the Kuparuk and Colville rivers (Telang et al., 1991).
Walker and Webber (1979) show that westward of the carbon-
Sample
type
Latitude
(N)
Longitude
(W)
Water
70°15.033⬘
148°18.484⬘
Snow
Flow
70°01.684⬘
69°00.527⬘
148°37.781⬘
148°49.214⬘
Water
Flow
70°19.812⬘
70°16.540⬘
149°00.527⬘
148°57.350⬘
Water
70°09.519⬘
150°56.791⬘
ate-rich Sagavanirktok River, soil pH decreases from ⬎7 to ⬍6
as the tundra shifts from wet alkaline to wet acidic because
soils become rich in organic acids.
The Kuparuk River drains an area of 8140 km2 and is
typically frozen for 7 to 9 months of the year (Hinzman et al.,
1991). Water flow usually reaches the coast in late May or early
June, largely due to discharge of melted snow that has accumulated during the winter months (McNamara et al., 1998).
During the initial surge of snowmelt, water discharge ranges
from 500 to 3500 m3/s at peak flow and then decreases to ⬍100
m3/s within 1 to 2 weeks (U.S. Geological Survey [USGS],
1971–2001). In 2001, the Kuparuk River transported 1.2 km3 of
water with ⬃75% of the discharge in June (USGS, 2001).
During a 3-yr study of the upper drainage basin of the Kuparuk
River, Kriet et al. (1992) determined that concentrations of TSS
ranged from 0.4 to 35 mg/L with sediment yields of 0.5 to
3.5 t/km2/yr (average ⫽ 1.6 t/km2/yr). The entire drainage
basin is underlain by permafrost that ranges in thickness from
250 m near the foothills to ⬎600 m near the coast (Osterkamp
and Payne, 1981). The active soil layer tends to increase in
depth throughout the summer months and typically thaws to
maximum depths of 25 to 40 cm (Hinzman et al., 1991).
The Sagavanirktok River is the second largest river on the
North Slope with a drainage area of ⬃15,000 km2 that extends
287 km from the Brooks Range to the Beaufort Sea (Robinson
and Johnsson, 1997). Similar to the Kuparuk River, spring
discharge of accumulated snowmelt occurs during late May to
early June with a peak discharge that ranges from ⬃300 to
1200 m3/s at the USGS gauge (USGS, 1971–2001, Table 1,
Fig. 1). However, discharge does not decrease as rapidly during
June in the Sagavanirktok River as in the Kuparuk River,
presumably due to the larger drainage basin and additional
sources of water from snowfields and glaciers in the Brooks
Range (McNamara et al., 1998). Water discharge at the Sagavanirktok River gauge (#15908000) totaled 1.3 km3 in 2001
(USGS, 2001). However, the gauge receives water flow from
only ⬃20% of the drainage basin and thus total water discharge
into the coastal Beaufort Sea is conceivably ⬃5 times greater
(USGS, 1971–2001; Robinson and Johnsson, 1997). During
June 2001, water discharge at the Sagavanirktok gauge accounted for 35% of the total annual discharge. Data for concentrations of TSS or sediment yield were not available.
The Colville River, the largest river in the Alaskan Arctic,
extends for 600 km and encompasses ⬃57,000 km2 in all three
Trace metals and organic carbon in Alaskan Arctic rivers
479
Table 2. Results for river water reference materials.
Ba (n ⫽ 3)
Cu (n ⫽ 5)
Fe (n ⫽ 3)
Pb (n ⫽ 5)
Zn (n ⫽ 5)
a
b
SLRS-3 certified
concentrationsa
(nmol/L)
NIST-1640 certified
concentrationsa
(nmol/L)
This studyb
(preconcentration)
(nmol/L)
This studyb
(direct analysis)
(nmol/L)
—
21.3 ⫾ 1.1
1791 ⫾ 35.8
0.42 ⫾ 0.03
15.9 ⫾ 1.4
1078 ⫾ 16
—
—
—
—
—
—
—
0.41 ⫾ 0.03
16.1 ⫾ 0.8
1070 ⫾ 13
21.6 ⫾ 0.6
1808 ⫾ 4
—
—
⫾ 95 confidence limits.
⫾ 1 standard deviation.
physiographic provinces. The USGS no longer maintains a
gauge on the Colville River; however, data from 1976 to 1977
indicate that peak discharge approaches 8500 m3/s (USGS,
1971–2001). Arnborg et al. (1967) calculated water discharge
and suspended sediment transport by the Colville River for
1962 and found that ⬃15 km3 of water and ⬃6 ⫻ 106 tons of
suspended sediment (116 t/km2/yr) are discharged annually.
3. SAMPLING AND ANALYSIS
3.1. Sampling
River sampling during the spring snowmelt was carried out
during June 2001. The Sagavanirktok River was sampled 20
times between June 3 and June 23 from a location near Prudhoe
Bay (Table 1, Fig. 1). Sampling of the Kuparuk River was
carried out nine times between June 10 and June 22, and the
Colville River, not easily accessible, was sampled five times in
June when helicopter support was available (Table 1, Fig. 1).
Water samples also were collected from the Sagavanirktok and
Kuparuk rivers during August 2001 and 2002 and the Colville
River during August 2001. Four snow samples were collected
during May 2002 adjacent to the Sagavanirktok River at a
location ⬃30 km upstream of Prudhoe Bay (Table 1, Fig. 1).
Water samples were collected using 0.5- and 1-L low-density
polyethylene (LDPE) bottles that were carefully washed with
concentrated HNO3 and rinsed twice (24 h per rinse) with
18-megaohm distilled deionized water (DDW) before sampling. Samples from the Sagavanirktok and Kuparuk rivers
were collected below the surface (⬃10 –20 cm) by lowering a
polyvinyl chloride (PVC) bottle holder from a bridge into the
rivers. The PVC sampler weighed ⬃14 kg, thereby allowing
the sample bottle to remain vertical and relatively stationary in
the water column during sampling. Water from the Colville
River was collected by hand using LDPE bottles while standing
in the river and facing or walking slowly upstream. Snow
samples were collected along the riverbanks and above the
frozen river using wide mouth high-density polyethylene
(HDPE) containers that were washed with concentrated HNO3
and rinsed twice (24 h per rinse) with DDW. The surface snow
(⬃5 cm) was discarded and subsurface samples were collected
using an HDPE scoop and the containers were sealed in plastic
bags.
Water samples were collected and filtered in a laminar flow
hood at a laboratory near Prudhoe Bay during the same day,
usually within 1 to 4 h of collection. Snow samples were
thawed at the laboratory and filtered immediately. The filtering
unit was acid washed (5 N HNO3), rinsed with distilled water
before each filtration and the first 20 to 50 mL of filtrate were
discarded as an additional rinse. Levels of TSS were determined by filtering 0.025 to 1 L of water through preweighed
and acid-washed (5 N HNO3) polycarbonate filters (0.40 ␮m,
47 mm) in a laminar-flow hood to minimize contamination. The
filters containing suspended solids were placed in acid-washed
petri dishes and stored in plastic bags. The filtrate was placed
in acid-cleaned polyethylene bottles and acidified to pH 2 with
Ultrex II HNO3 and stored in plastic bags. A separate portion
of river water to be analyzed for particulate organic carbon
(POC) was filtered through precombusted (500°C) Gelman
Type A/E glass-fiber filters (47 mm). The filters were stored in
petri dishes and sealed in plastic bags. The filtrate was collected
for DOC, transferred to glass vials with Teflon-lined caps and
frozen until analysis.
3.2. Analysis
3.2.1. Dissolved metals
Concentrations of dissolved Ba and Cu in the filtered and
acidified water samples were determined by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) using a PerkinElmer ELAN 5000 instrument. Dissolved Fe concentrations
were determined for the acidified samples by graphite furnace
atomic absorption spectroscopy (GFAAS) using a PerkinElmer Model 4000 instrument with an HGA-400 heated graphite atomizer and an AS-40 autosampler. Accuracy was assessed
using the certified reference material (CRM) SLRS-3, from the
National Research Council of Canada (NRC), or the standard
reference material (SRM) NIST-1640, from the U.S. National
Institute of Standards and Technology (NIST), with all values
within the range of certified concentrations (Table 2). Acid
blanks were prepared before analysis and were below detection
limits of 0.7, 0.05, and 0.18 nmol/L, for Ba, Cu and Fe,
respectively. Spiked samples were used to determine recoveries
for dissolved Ba, Cu and Fe that averaged 94% (n ⫽ 4), 97%
(n ⫽ 4) and 96% (n ⫽ 4), respectively. The precision of
replicate analyses (n ⫽ 4) averaged 5% for Ba, 7% for Cu and
9% for Fe.
Concentrations of dissolved Pb and Zn were determined in a
subset of the samples from July and August (Sagavanirktok
River, n ⫽ 14; Kuparuk River, n ⫽ 7; Colville River, n ⫽ 5)
by preconcentrating the metals from filtered water samples
480
R. D. Rember and J. H. Trefry
(⬃400 mL) in a laminar flow hood using the technique described by Nakashima et al. (1988). In this procedure, high
purity Fe, Pd and NaBH4 are used to co-precipitate the dissolved metals at pH 8.5. The precipitate is collected on acidwashed filters and then dissolved in Ultrex II HNO3 and Ultrex
II HCl and diluted to a final volume of ⬃4 mL. The concentration factor is ⬃100-fold using this method. Concentrations
of dissolved Pb and Zn were determined by ICP-MS with
detection limits of 10 pmol/L and 0.15 nmol/L, respectively.
Recoveries of dissolved Pb and Zn from spiked samples averaged 98% (n ⫽ 4) and 92% (n ⫽ 4), respectively. The precision
of replicate analyses averaged 6% for Pb (n ⫽ 4) and 9% for Zn
(n ⫽ 4). A dissolved metal blank was prepared with the
following reagents used in the preconcentration technique: 1
mL Ultrex II NH4OH, 1 mL of each of the Fe and Pd carriers
and 2.5 mL of the NaBH4. The precipitate from the blank was
filtered and dissolved in Ultrex II HNO3 and Ultrex II HCl and
was equivalent to 0.30 ⫾ 0.03 nmol/L for Zn (n ⫽ 4) and 36 ⫾
4 pmol/L for Pb (n ⫽ 4). Concentrations of dissolved Pb and Zn
in this paper are blank corrected. Concentrations of dissolved
Pb and Zn determined for the CRM SLRS-3 during this study
were within the range of certified concentrations (Table 2).
3.2.2. Particulate metals
Filters containing 1 to 15 mg of suspended sediment were
dried in a humidity-controlled (50% relative humidity) clean
room for 24 h and weighed to ⫾1 ␮g. The suspended sediment
was dissolved using the method described by Trefry and Trocine (1991). Briefly, the filters with suspended sediment were
placed in stoppered, 15-mL Teflon test tubes and the suspended
sediment was completely decomposed and dissolved using
Ultrex II HNO3, HF and HCl. The test tubes were heated at
⬃75°C in a water bath, cooled and reheated. The resultant
solutions were transferred to acid-washed LDPE bottles and
diluted to 6 mL using DDW rinses of the test tubes. Small
portions (⬃5–10 mg) of SRM 2704, a marine sediment from
NIST, and filter blanks also were prepared in duplicate with
each set of samples to assess accuracy and precision.
Concentrations of Al, Fe and Zn in suspended particles were
determined by flame atomic absorption spectroscopy using a
Perkin-Elmer 4000 instrument. Concentrations of Cu were
quantified by GFAAS. Values for particulate Ba and Pb were
determined by ICP-MS. Values for filter blanks were below the
method detection limits of 0.15%, 0.15%, 9 ␮g/g, 0.4 ␮g/g, 0.2
␮g/g and 2 ␮g/g for Al, Fe, Ba, Cu, Pb and Zn, respectively.
Concentrations of metals determined for the SRM were within
the range of certified concentrations (Table 3). The precision of
replicate analyses averaged 3% for Al (n ⫽ 4), 4% for Fe (n ⫽
4), 4% for Ba (n ⫽ 4), 6% for Cu (n ⫽ 4), 4% for Pb (n ⫽ 4)
and 7% for Zn (n ⫽ 4).
3.2.3. Organic carbon
Filters for POC were treated with 15% H3PO4 to remove
inorganic carbon phases, rinsed with DDW and dried. The
filters were placed in ceramic boats and combusted at 900°C in
a Shimadzu TOC-5050A carbon system with an SSM-5000A
solid sampling module. The organic carbon content of the
samples was determined using a four-point calibration curve
Table 3. Results for sediment reference materials.
Al (%) (n ⫽ 6)
Ba (␮g/g) (n ⫽ 6)
Cu (␮g/g) (n ⫽ 6)
Fe (%) (n ⫽ 6)
Pb (␮g/g) (n ⫽ 6)
Zn (␮g/g) (n ⫽ 6)
POC (%) (n ⫽ 7)
NIST-2704 certified
concentrationsa
This studyb
6.11 ⫾ 0.16
414 ⫾ 12
98.6 ⫾ 5.0
4.11 ⫾ 0.1
161 ⫾ 17
438 ⫾ 12
2.14 ⫾ 0.03c
6.15 ⫾ 0.09
408 ⫾ 9.0
102 ⫾ 2.1
4.11 ⫾ 0.1
165 ⫾ 16
430 ⫾ 1.3
2.04 ⫾ 0.06
⫾ 95 confidence limits.
⫾ 1 standard deviation.
c
MESS-2 (NRC).
a
b
with pure sucrose as the standard. The calibration curve was
checked every 10 samples by analyzing the CRM MESS-2, a
marine sediment issued by the NRC. Values for the CRM were
within 5% of certified concentrations (Table 3). The precision
of the POC analysis for replicate samples averaged 6.9% (n ⫽
4).
The DOC concentrations of river water were determined by
combustion using a Shimadzu TOC-5050A carbon system. The
DOC concentrations were calculated by subtracting inorganic
carbon (IC) from total carbon (TC). Four-point calibration
curves were prepared using potassium hydrogen phthalate (TC)
and sodium bicarbonate (IC). The calibration curve was
checked every 10 samples by repeating the analysis of a
midrange standard. The precision of replicate analyses averaged 4.1% (n ⫽ 7) for the DOC analyses.
4. RESULTS AND DISCUSSION
4.1. Total Suspended Solids and Water
Concentrations of TSS increased from ⬃40 to ⬎600 mg/L in
the Sagavanirktok River during the first 8 d of water flow (June
4 –12, 2001) and decreased to ⬍50 mg/L within 4 to 5 d
following maximum flow (Fig. 2a). Data for water flow at the
USGS gauge, located ⬃150 km upstream, followed a similar
trend. A 1- to 2-d offset between the peaks in TSS and water
flow is consistent with a lag time for flow between the USGS
gauge and the primary sampling site located ⬃150 km downstream (Fig. 2a). Uncertainty in the water flow data, along with
the upstream location of the gauge, make it difficult to evaluate
possible differences in timing between maximum levels of TSS
and water flow.
The flow gauge in the Sagavanirktok River was not installed
by the USGS until the river was free of ice on June 10, 2001,
⬃7 d after water flow began. Thus, during 2001 and most other
years, the ascending limb of the hydrograph is calculated, not
measured for the Sagavanirktok River. Furthermore, because
the gauge is upstream in the western drainage basin of the
Sagavanirktok River, the gauge covers only ⬃20% of the
watershed and may not represent characteristic flow patterns
for the entire basin. During 2001, the total volume of water
measured at the USGS gauge in the Sagavanirktok River was
⬃1.3 km3, yielding a total extrapolated flow of ⬃6.5 km3/yr
(USGS, 1971–2001).
Concentrations of TSS in the Kuparuk River peaked at 66
Trace metals and organic carbon in Alaskan Arctic rivers
Fig. 2. Concentrations of total suspended solids (●), measured water
flow (E) and calculated water flow (䊐) in the (a) Sagavanirktok and (b)
Kuparuk rivers during June 2001 using water discharge data from U.S.
Geological Survey gauges #15908000 and #1589600, respectively. The
shaded areas indicate the period defined as peak discharge for each
river.
mg/L on June 10 and decreased to ⬍4 mg/L within 7 d (Fig.
2b). These TSS values are low relative to the Sagavanirktok
River and are more typical of an arctic river that has no
significant mountain source of suspended solids (Craig and
McCart, 1975). The gauge on the Kuparuk River is located
within ⬃10 km of the coastal Beaufort Sea and covers ⬎90%
of the drainage basin. In 2001, water discharge by the Kuparuk
River totaled ⬃1.2 km3 (USGS, 1971–2001). The same procedure for gauge installation used for the Sagavanirktok River is
followed for the Kuparuk River. Therefore, the initial water
flow in the Kuparuk River is estimated and not measured (Fig.
2b).
Similar to the Sagavanirktok River, concentrations of TSS
increased to ⬎600 mg/L in the Colville River during peak
discharge. No water discharge data are available for the
Colville River because the USGS no longer maintains a gauge
on the river. Arnborg et al. (1967) estimated annual water
export by the Colville River to be ⬃15 km3.
For the purposes of discussion in this paper, peak discharges
in the Sagavanirktok and Kuparuk rivers are operationally
defined as the periods of June 5 to 16 and June 10 to 12,
respectively. During those periods, the discharge of TSS is ⬎50
mg/L and water flow rises to a maximum and then decreases to
more uniform values during June and throughout the summer
(Fig. 2). Similar trends between water flow and concentrations
of TSS in the Sagavanirktok and Kuparuk rivers suggest that
TSS data from the Colville River can be used to help estimate
peak discharge from about June 6 to 14.
481
Fig. 3. Concentrations of dissolved organic carbon and water discharge (dashed line) in the (a) Sagavanirktok (●) and (b) Kuparuk
rivers (Œ) during June 2001. The shaded areas indicate the period
defined as peak discharge for each river.
4.2. Dissolved Organic Carbon and Trace Metals
During the first 5 d of discharge (June 3–7), concentrations
of DOC in the Sagavanirktok River increased sharply from 167
to 742 ␮mol/L, levels that are up to 6 times greater than during
off-peak discharge later in June (Fig. 3a, Table 4). Similar
trends for concentrations of DOC also were observed in the
Kuparuk and Colville rivers, where peak concentrations were
1320 and 1100 ␮mol/L, respectively (Table 4, Fig. 3b). Higher
concentrations of soluble organic matter in the Kuparuk and
Colville rivers, relative to the Sagavanirktok River, reflect
differences in the nature of the drainage basins as described
later in this paper.
During peak discharge in June 2001, maximum concentrations of DOC in the Sagavanirktok, Kuparuk and Colville rivers
were 2 to 6 times higher than minimum concentrations measured during off-peak discharge in June (Table 4). Increased
concentrations of DOC during peak flow are amplified with
respect to total transport by large increases in water discharge.
However, maximum concentrations of DOC in the Sagavanirktok River occurred before maximum water discharge and the
lag effect only strengthens this observation. Then, levels of
DOC decreased as water flow continued to increase, possibly
due to dilution by snowmelt.
Concentrations of DOC also have been observed to increase
during peak water discharge in many rivers of the world (Cauwet and Meybeck, 1987; Depetris and Paolini, 1991; Cauwet
and Sidorov, 1996; Boyer et al., 1997). This phenomenon is
caused when snowmelt or rising water percolates through lit-
482
R. D. Rember and J. H. Trefry
Table 4. Concentrations of dissolved metals, dissolved organic carbon (DOC) and total suspended solids (TSS) in the Sagavanirktok, Kuparuk and
Colville rivers during peak and off-peak discharge, June 2001.
Ba
(nmol/L)
Cu
(nmol/L)
Fe
(nmol/L)
Pb
(pmmol/L)
Zn
(nmol/L)
DOC
(␮mol/L)
TSS
(mg/L)
Mean
SD
Range
Mean
SD
Range
Mean
SD
Range
229
⫾16
200–256
142
⫾11
131–153
369
⫾26
340–392
13.2
⫾3.3
9.0–15.1
12.6
⫾1.2
11.3–13.5
38.9
⫾3.1
34.3–42.3
755
⫾410
201–1290
3800
⫾484
3250–4170
1450
⫾536
1066–1825
153
⫾55
75–225
256
⫾3.8
253–260
290
⫾45
258–321
3.7
⫾1.8
2.0–6.7
5.8
⫾0.4
5.5–6.3
2.1
⫾0.3
1.7–2.6
480
⫾222
167–742
1170
⫾136
1050–1320
835
⫾160
667–1100
267
⫾171
78–609
63
⫾6
55–66
468
⫾272
333–610
Mean
SD
Range
Mean
SD
Range
Mean
SD
Range
239
⫾13
225–231
171
⫾16
135–194
376
⫾24
359–393
6.5
⫾1.4
4.6–8.8
12.1
⫾0.7
11.2–12.7
24.5
⫾1.8
23.3–26.3
110
⫾48
51–170
1150
⫾382
658–2080
900
⫾770
357–1440
59
⫾24
34–92
196
⫾64
105–288
196
⫾43
133–244
1.8
⫾0.5
1.2–2.4
4.4
⫾0.2
4.1–4.8
4.0
⫾1.5
2.9–5.0
196
⫾65
117–408
728
⫾39
692–788
454
⫾160
342–608
31
⫾12
14–40
4.0
⫾2.4
1.7–7.4
89
⫾41
42–113
River
Peak discharge
Sagavanirktok (n ⫽ 9)
Kuparuk (n ⫽ 3)
Colville (n ⫽ 2)
Off-peak discharge in June
Sagavanirktok (n ⫽ 8)
Kuparuk (n ⫽ 6)
Colville (n ⫽ 3)
terfall and DOC-rich interstitial water in the upper soil horizon
where concentrations of DOC can be ⬎8000 ␮mol/L (Boyer et
al., 1997; Michaelson et al., 1998). Depetris and Paolini (1991)
found that concentrations of DOC increased from ⬃400 to 900
␮mol/L during a flooding event in the Orinoco River. A positive correlation between DOC and water flow also was reported to occur in the Lena River where concentrations of DOC
were at least 30 to 40% higher during the spring floods than
during September (Cauwet and Sidorov, 1996).
Concentrations of dissolved Fe, Cu, Pb and Zn follow trends
shown for DOC (Fig. 4). In the Sagavanirktok River, concentrations of dissolved Fe increased sharply from 57 to 1290
nmol/L in ⬃7 d. Then, as flow decreased, concentrations of
dissolved Fe returned to ⬃50 nmol/L within 8 d of maximum
levels. Concentrations of dissolved Cu, Pb and Zn in the
Sagavanirktok River also increased to maximum levels of 15
nmol/L, 225 pmol/L and 6.7 nmol/L, respectively, during peak
flow (Fig. 4). These maximum values at peak discharge are ⬃3,
5.5 and 6 times greater, respectively, than minimum levels
during off-peak discharge in June and occur during the period
when water flow is highest, indicating that net transport of
dissolved metals, like DOC, is greatly enhanced during peak
discharge.
Strong correlations (r ⬎ 0.82) were found for DOC vs.
dissolved Cu, Fe, Pb and Zn (Table 5). The trends, correlations
and timing for peak levels of trace metals and DOC support the
idea that concentrations of dissolved metals and DOC are
strongly influenced by the discharge of soil interstitial water
and shallow surface water that is diluted by snowmelt and
flushed from surrounding soils into the Sagavanirktok River.
During the short summers, the arctic coastal plain is covered
with pools of standing water and lakes where DOC accumulates and trace metals are leached from soils. The permafrost,
along with a topography that averages ⬍2 m on the coastal
plain, inhibit lateral water flow during the summer months,
thereby extending the residence time of the surface water in the
system (McNamara et al., 1998). Then, after the long frozen
winter (8 –9 months), increased surface runoff in the spring
provides a direct pathway for the release of accumulated and
freshly leached DOC and dissolved metals into the rivers from
the thawing ponds and soils.
On June 3, 2001, the first day of water flow in the Sagavanirktok River, concentrations of dissolved Cu, Fe, Pb and DOC
in the river water were ⬃3 to 25 times lower than at peak flow.
Much of this early runoff preceded complete thawing of soils
and ponds and may have contained a significant snow component. Results for filtered snow samples collected for this study
are as follows: Cu, 0.91 ⫾ 0.47 nmol/L; Fe, 17.8 ⫾ 7.1 nmol/L;
Pb, 30 ⫾ 7 pmol/L, Ba, 5 ⫾ 0.8 nmol/L and DOC, 56 ⫾ 6
␮mol/L. These concentrations of Cu, Fe, Pb, Ba and DOC in
snow are 2 to 25 times lower than during the first day of
discharge in June for the Sagavanirktok River and 7.5 to 86
times lower than maximum levels during peak discharge (Table
4). Thus, in the period before peak flow, concentrations of
dissolved metals and DOC in river water are low, yet somewhat
elevated above values for snow. As melting of snow and soils
progresses, dissolved metals and organic carbon that are released from interstitial water and ponds are carried to the rivers
in increasing amounts during the brief period of accelerated
runoff from the land.
Average concentrations of dissolved Fe, Pb, Zn and organic
carbon in the Kuparuk River during peak and off-peak discharge in June were ⬃2 to 5 times higher than in the Sagavanirktok River (Table 4). Concentrations of dissolved Cu in the
Kuparuk River are not significantly different (t test, ␣ ⫽ 0.05,
p ⫽ 0.05) during peak vs. off-peak flow in June. During peak
discharge, average concentrations of DOC, Cu, Fe and Pb in the
Colville River were 1.5 to 2 times higher than during off-peak
flow (Table 4). In general, concentrations of dissolved metals
and DOC in the Colville River were higher than those found in
Trace metals and organic carbon in Alaskan Arctic rivers
483
Fig. 4. Concentrations of dissolved (a) Fe, (b) Cu, (c) Pb, (d) Zn and (e) Ba in the Sagavanirktok River during June 2001.
The shaded areas indicate the period defined as peak discharge. The solid line on the Ba graph is the mean value and dashed
lines represent ⫾2 standard deviations from the mean.
the Sagavanirktok River, but less than in the Kuparuk River
(Table 4).
Higher concentrations of DOC and dissolved metals in the
Kuparuk and Colville rivers, relative to the Sagavanirktok
River, are most likely related to differences in regional lithol-
ogy and soil pH (organic acids). Parkinson (1977) found an
inverse relationship between calcium carbonate equivalents and
organic matter concentrations in the region. These data suggest
that soils within the Kuparuk and possibly Colville drainage
basins may undergo more intense chemical weathering due to
484
R. D. Rember and J. H. Trefry
Table 5. Pearson’s correlations between dissolved trace metals and
dissolved organic carbon (DOC) in the Sagavanirktok River during
June 2001. All correlations are significant at p ⫽ 0.01.
Cu
Fe
Pb
Zn
DOC
Cu
Fe
Pb
Zn
1
0.86
0.89
0.79
0.82
1
0.93
0.90
0.90
1
0.90
0.91
1
0.87
higher concentrations of organic acids and lower pH, whereas
the carbonate-rich Sagavanirktok River drainage basin may be
well buffered from pH changes (Walker and Webber, 1979).
In contrast with Cu, Fe, Pb and Zn, concentrations of dissolved Ba in the Sagavanirktok River vary by only ⫾7% (CV
⫽ SD/x̄ ⫻ 100) during peak discharge (Table 4, Fig. 4).
Concentrations of dissolved Ba are even less variable (CV
⬍5%) during pre and postpeak discharge in June (Table 4, Fig.
4). Dissolved Ba concentrations were ⬃25% lower during
maximum water flow as a result of dilution by snowmelt. These
results indicate that concentrations of dissolved Ba are not
enhanced by flushing of metals from soils during peak flow and
are consistent with other studies that show weak complexation
between Ba and organic ligands (Dupre et al., 1999; Pokrovsky
and Schott, 2002). Partitioning of Ba between dissolved and
particulate phases also may help control concentrations of
dissolved Ba as described later in this paper.
Concentrations of dissolved Ba in the Sagavanirktok (234 ⫾
14 nmol/L), Kuparuk (171 ⫾ 16 nmol/L) and Colville (376 ⫾
24 nmol/L) rivers (Table 4) are within the range (138 –574
nmol/L) reported by Guay and Falkner (1998) for the Mackenzie River. These data support the conclusion by Guay and
Falkner (1998) that concentrations of dissolved Ba in Arctic
rivers of North America are higher than those determined for
any of the Eurasian Arctic rivers (24 –120 nmol/L). Guay and
Falkner (1998) suggest that differences in concentrations between arctic rivers in North America and Eurasia are due to the
chemical composition and weathering characteristics within
their respective drainage basins.
Concentrations of dissolved Ba, Cu, Pb, Zn and DOC during
August 2001 and August 2002, when water flow was greatly
reduced (⬍50 m3s), are consistent with or lower than concentrations found in other arctic rivers during the months of
August and September (Table 6). For example, concentrations
of dissolved Pb during August (2001, 2002) in the Sagavanirktok, Kuparuk and Colville rivers range from 19 to 56 pmol/L.
These values are within the broad range of concentrations
found in the Yenisey (25–29 pmol/L), the Ob (55– 83 pmol/L)
and the Lena (80 pmol/L) rivers during September. Overall,
concentrations of dissolved metals in arctic rivers of Alaska are
5 to 10 times lower in August than during peak flow in June.
4.3. Particulate Trace Metals and Organic Carbon
Concentrations of particulate Cu, Pb, Zn (␮g/g dry weight),
Fe and OC (percentage dry weight) in the Sagavanirktok River
are more uniform than observed for the dissolved species (Fig.
5). In general, concentrations of Cu (CV ⫽ 8.5%), Pb (CV ⫽
11%), and Zn (CV ⫽5.7%) are less variable than concentrations of Fe (CV ⫽16%) and POC (CV ⫽31%) (Table 7). In
sharp contrast with dissolved Cu, Fe, Pb and Zn, particulate
concentrations of these four metals were poorly correlated with
organic carbon, and relatively well correlated with particulate
Al (Table 8). The poor correlations for POC vs. Cu, Fe, Pb and
Zn suggest that POC does not play a significant role in controlling concentrations of particulate trace metals in the Sagavanirktok River, but that aluminosilicates (clays) are more
important.
The trends in concentrations of particulate Ba in the Sagavanirktok River follow closely with those observed for Fe, Cu,
Pb and Zn, suggesting that concentrations of particulate Ba also
are controlled by the aluminosilicate content of the suspended
solids (Fig. 6a). The correlation coefficient for particulate Al
vs. Ba for the Sagavanirktok River (r ⫽ 0.88) indicates that
variations in particulate Ba concentrations result from shifts in
Al concentrations during peak and off-peak discharge in June.
When concentrations of particulate Ba from the Kuparuk and
Colville rivers are added to data for the Sagavanirktok River
(Fig. 6a), a strong relationship between Al and Ba is still
observed (r ⫽ 0.90). The data in Figure 6a include values from
all three rivers with a range of ⬃2 to ⬎600 mg/L for TSS, and
concentrations of POC ranging from ⬍1 to ⬎7%. Therefore,
variations in concentrations of particulate Ba in the Sagava-
Table 6. Ranges and average concentrations of dissolved metals in arctic rivers during August and September.
River
Sagavanirktok
Kuparuk
Colville
Lena
Lena
Ob-Irtysh
Yenisey
World Average
Date (this
study) or
reference
TSS
(mg/L)
Ba
(nmol/L)
Cu
(nmol/L)
Fe
(nmol/L)
Pb
(pmol/L)
Zn
(nmol/L)
DOC
(␮mol/L)
8/11/01
8/2/02
8/11/01
8/2/02
8/11/01
1, 2, 3
4
2, 5
2, 5
6, 7
1.8
0.70
0.64
0.16
6.5
10–20
5–7
—
—
—
285
288
295
349
540
130
—
100
125
—
4.1
3.5
12
9.8
11
12–15
9.7
29–38
22–30
24
38
—
185
—
68
434–850
410
430–654
251–317
716
27
19
25
29
56
—
80
55–83
25–29
150
0.52
1.4
0.52
0.75
0.49
1.2
5.3
—
—
9.2
92
310
425
325
316
495–661
—
614–829
334
480
(1) Guieu et al. (1996); (2) Guay and Falkner (1998); (3) Cauwet and Sidorov (1996); (4) Martin et al. (1993); (5) Dai and Martin (1995); (6) Martin
and Windom (1991); (7) Meybeck (1982).
Trace metals and organic carbon in Alaskan Arctic rivers
485
Fig. 5. Concentrations of particulate (a) Ba, (b) Cu, (c) Fe, (d) Pb, (e) Zn and (f) POC in the Sagavanirktok River during
June 2001. The solid line shows mean concentrations in each plot and dashed lines represent ⫾2 standard deviations from
the mean.
nirktok, Kuparuk and Colville rivers result from dilution of
fine-grained aluminosilicates by larger-grained sands and other
non-aluminosilicate minerals. Concentrations of dissolved Ba
vary by a factor of 2 to 3 among rivers; however, these
variations are directly proportional to concentrations of particulate Ba. The average ratio for particulate Ba/dissolved Ba is 23
⫾ 5 L/g for all samples (n ⫽ 31). The relatively uniform ratio
supports a direct relationship between dissolved Ba and exchangeable particulate Ba; however, these results cannot be
confirmed by this study (Tables 4 and 7). Similar trends were
not observed for Cu, Fe, Pb and Zn.
During peak discharge, as the banks of the Kuparuk River
erode and the river transports higher concentrations of TSS
(⬎50 mg/L), concentrations of particulate Fe, Pb and Zn correlate well (r ⫽ 0.98, 0.76, 0.93, respectively) with levels of
particulate Al and the metal/Al ratios agree within 15% with
the results found for the Sagavanirktok River (Fig. 6). However, when TSS concentrations decrease to an average of 4
mg/L in the Kuparuk River during off-peak discharge in June,
the average Fe/Al ratio increases by ⬃35% relative to the
Sagavanirktok River (Fig. 6b). These data suggest that elevated
concentrations of particulate Fe (6400 nmol/L or 3.7% dry
weight) relative to Al (14,600 nmol/L or 3.5% dry weight)
during off-peak discharge may be influenced by high concentrations of dissolved Fe (1150 nmol/L) in the Kuparuk River
that potentially enhance the formation of Fe hydrous oxides and
increase concentrations of particulate Fe. Similarly, concentrations of particulate Pb and Zn increase with Fe and are elevated
above levels that would be predicted from the Al concentration
and possibly result from scavenging by hydrous Fe-oxides.
Average concentrations of particulate Cu, Fe and Pb in the
Colville River averaged 25 to 40% higher than concentrations
found in the Sagavanirktok and Kuparuk rivers (Table 7).
However, particles from the Colville River also had 50% higher
concentrations of Al indicating that sources of suspended sediments in the western Brooks Range are richer in fine-grained
aluminosilicates or less diluted by non-aluminosilicate minerals
(Table 7). When concentrations of particulate metals are plotted
vs. Al, they agree with the trend found for metals in the
Sagavanirktok River (e.g., Fig. 6).
486
R. D. Rember and J. H. Trefry
Table 7. Concentrations of particulate trace metals and particulate organic carbon (POC) in the Sagavanirktok, Kuparuk and Colville Rivers during
June 2001. Peak and off-peak concentrations are presented for the Kuparuk River due to distinct differences in the concentrations of some trace metals
during those periods.
River
Sagavanirktok (n ⫽ 17)
Kuparuk Peak
discharge (n ⫽ 3)
Kuparuk Off-peak
discharge (n ⫽ 6)
Colville (n ⫽ 5)
Mean
SD
Range
Mean
SD
Range
Mean
SD
Range
Mean
SD
Range
Al
(%)
Ba
(␮g/g)
Cu
(␮g/g)
Fe
(%)
Pb
(␮g/g)
Zn
(␮g/g)
POC
(%)
5.8
⫾0.8
4.0–7.7
5.2
⫾0.5
4.6–5.8
3.5
⫾0.5
2.7–4.0
8.3
⫾0.3
8.0–8.9
729
⫾105
542–1008
615
⫾88
497–743
469
⫾70
397–559
989
⫾125
890–1190
32.9
⫾2.9
29.7–39.2
32.3
⫾5.8
25.7–37.2
29.9
⫾1.3
28.6–31.2
40.8
⫾2.3
36.9–42.6
3.4
⫾0.3
2.6–4.0
3.6
⫾0.2
3.3–3.8
3.7
⫾0.3
3.1–4.2
5.0
⫾0.3
4.7–5.3
18.3
⫾2.1
16.0–23.4
15.7
⫾0.7
14.7–16.6
14.4
⫾8.3
4.0–22.5
24.0
⫾6.8
19.8–36.0
124
⫾8
115–135
124
⫾13
110–145
92.1
⫾21
69–116
130
⫾11
116–147
1.60
⫾0.49
0.7–2.7
4.6
⫾1.1
3.3–5.9
4.9
⫾2.0
2.8–7.6
2.24
⫾0.56
1.7–3.3
4.4. Transport of Water, Sediment, Trace Metals and
Organic Carbon
Values for transport of metals and organic carbon by rivers
to the coastal Beaufort Sea are difficult to calculate due to
limitations in flow and chemical data as well as possible contributions from summer rainstorms. Nevertheless, by comparing transport during peak discharge in June with transport
during the remainder of the year, a sense of the relative importance of spring floods to annual budgets is obtained. Calculations for material transport are presented here for the Kuparuk
and Sagavanirktok rivers because data for water flow are available (USGS, 1971–2001).
Water discharge in the Kuparuk River was divided into three
periods: (1) 0.3 ⫻ 1012 L (25% of total) during 3 d of flood
discharge in June, (2) 0.4 ⫻ 1012 L (33% of total) during the 21
off-peak days in June and (3) 0.5 ⫻ 1012 L (42% of total)
during the remaining 90 d of the water year (USGS, 1971–
2001). By combining water flow data with average concentrations of DOC for each time period (Tables 4 and 6), 42% of the
annual load of DOC ([0.25][1170 ␮mol/L]/[(0.25)(1170
␮mol/L) ⫹ (0.33)(728␮mol/L) ⫹ (0.42)(375 ␮mol/L)] ⫻
100%) of the Kuparuk River is delivered to the Beaufort Sea in
3 d. Using the same approach shown above for DOC, the
following fractions of the annual loads of dissolved metals are
carried by the Kuparuk River in 3 d of peak flow: Ba (16%), Cu
(36%), Fe (67%), Pb (47%) and Zn (46%).
Using the USGS data for the Sagavanirktok River and assuming that it represents flow for 20% of the system, then the
annual flow can be scaled up and grouped as follows: (1) 1.0 ⫻
1012 L (17% of total) of peak flow during 12 d in June, (2) 1.2
⫻ 1012 L (18% of total) of off-peak flow during 16 d in June
and (3) 4.2 ⫻ 1012 L (65% of total) during the remaining 90 d
of the water year (USGS, 1971–2001). Using DOC data (Tables 4 and 6), 33% of the annual load of DOC ([0.17][480
␮mol/L]/[(0.17)(480 ␮mol/L) ⫹ (0.18)(196 ␮mol/L) ⫹ (0.65
⫻ 201 ␮mol/L)] ⫻ 100%) is transported to the coastal Beaufort
Sea in 12 d. Using the same approach, the fractions of the
annual transport of dissolved metals carried by the Sagavanirktok River in 12 d were as follows: Ba (15%), Cu (38%), Fe
(74%), Pb (50%) and Zn (40%).
Calculations for annual transport of elements with suspended
sediment are based on two time periods (1) peak flow and (2)
off-peak flow in June plus the remainder of the summer, because sediment transport is predominantly during peak flow
and because differences in concentrations of elements in the
particulate form (on a dry weight basis) vary only slightly
among time periods. During the 3 d of peak water flow in the
Kuparuk River in June 2001, TSS averaged 63 mg/L to yield a
sediment discharge of ⬃19,000 t (90% of total). During the
off-peak period in June and the remainder of the summer, TSS
values averaged 4 and ⬃0.5 mg/L, respectively, resulting in an
additional total of 1900 t of sediment discharged (Tables 4 and
6). Thus, ⬎90% of the annual transport of particulate Pb occurs
during peak flow in the Kuparuk River during 3 d ([0.90][15.7
␮g/g]/[(0.10)(14.4 ␮g/g) ⫹ (0.90)(15.7 ␮g/g)] ⫻ 100%). Using
the same approach, ⬎89% of the other particulate metals and
POC were transported to the coastal Beaufort Sea by the
Kuparuk River in 3 d.
For the Sagavanirktok River, TSS averaged 267 mg/L during
peak discharge in June to yield 267,000 t of sediment (88% of
total) during the 12-d peak period. Concentrations of TSS
average 31 and 1.3 mg/L during the off-peak period in June and
the remainder of the summer, respectively, transporting an
additional 36,000 t of sediment to the coastal Beaufort Sea.
Particulate metal concentrations in the Sagavanirktok River do
not vary greatly during peak and off-peak periods (Table 7) and
therefore, ⬎88% of particulate Pb, the other metals and POC
were transported during the 12 d of peak discharge ([0.88][18.3
␮g/g]/[(0.12)(18.3 ␮g/g) ⫹ (0.88)(18.3 ␮g/g] ⫻ 100%).
Results from this study also provide a direct comparison of
the relative abundances of dissolved and particulate forms of
metals and organic carbon in the arctic rivers of Alaska (Tables
4 and 7). During the brief periods of peak flow, the dissolved
fraction accounts for 57% (480 ␮mol/L/[480 ␮mol/L ⫹
(16,000 ␮g/g/12 ␮g/␮mol ⫻ 0.267 g/L) ⫻ 100%]) and 83% of
DOC transport in the Sagavanirktok and Kuparuk rivers, respectively (Table 9). When combined with the off-peak data,
the results indicate that DOC is the dominant form of OC
discharged to the coastal Beaufort Sea from these Alaskan
arctic rivers. Dissolved Ba accounts for ⬎14% of total Ba
transport during peak flow and at least 59% of Ba transport
during off-peak flow in June (Table 9). In the Sagavanirktok
Trace metals and organic carbon in Alaskan Arctic rivers
Table 8. Correlations for particulate trace metals and particulate
organic carbon (POC) in the Sagavanirktok River during June 2001.
Al
Al
Ba
Cu
Fe
Pb
Zn
POC
1
0.88**
0.81**
0.97**
0.82**
0.77**
–0.35
Ba
1
0.83**
0.86**
0.92**
0.63**
–0.18
Cu
1
0.85**
0.87**
0.51*
–0.25
Fe
1
0.83**
0.64**
–0.25
Pb
1
0.59
–0.28
487
Table 9. Percentage of trace metals and organic carbon in the
dissolved phase during peak and off-peak discharge in the Sagavanirktok and Kuparuk rivers.
Zn
1
–0.29
** Correlation is significant at the 0.01 level.
* Correlation is significant at the 0.05 level.
River, ⬍3% of the total Fe, Pb and Zn are in the dissolved
phase during both peak and off-peak periods, whereas dissolved Cu is ⱖ9% of the total Cu, most likely due to complexation by DOC (Cabaniss and Shuman, 1998).
5. CONCLUSIONS
The 3- and 12-d spring flooding events in the Kuparuk and
Sagavanirktok rivers during June 2001 account for ⬃25 and
Fig. 6. Concentrations of particulate Al vs. (a) Ba and (b) Fe for the
Sagavanirktok River (●), Kuparuk River during peak discharge (Œ),
Kuparuk River during off-peak discharge in June (‚) and Colville
River (䊐) during June 2001. The solid lines are from least-squares
regressions of the data. Dashed lines show 95% prediction intervals.
Data points for off-peak discharge in the Kuparuk River are not
included in the linear regression calculations for the Al vs. Fe plot.
Sagavanirktok River
Dissolved (%)
Dissolved (%)
Kuparuk River
Dissolved (%)
Dissolved (%)
Ba
Cu
Fe
Pb
Zn
OC
Peak
Off-peak
14
59
9
29
0.5
0.6
0.6
2
0.7
3
57
83
Peak
Off-peak
33
93
28 9
87 30
4.6
44
83
98
5
41
⬃17%, respectively, of the annual discharge of water (USGS,
1971–2001). High-resolution sampling of the Sagavanirktok
River during peak flow revealed that concentrations of dissolved Cu, Fe, Pb, Zn and organic carbon increased by threefold to twenty-fivefold at maximum discharge. Strong positive
correlations for concentrations of dissolved metals vs. DOC
suggest that soil interstitial water and surface water that are
flushed from the drainage basins are primary sources of metals
and DOC. During off-peak discharge in August, when flow and
levels of TSS decrease, concentrations of dissolved metals in
the Sagavanirktok and Kuparuk rivers are among the lowest
values reported for world rivers.
Trends for concentrations of dissolved trace metals and
organic carbon in the Kuparuk and Colville rivers are similar to
those observed in the Sagavanirktok River. However, concentrations of dissolved metals and DOC are, on average, higher
than those observed in the Sagavanirktok River during June
2001. These higher concentrations in the Kuparuk and Colville
rivers appear to be related to differences in lithology and
vegetation in the drainage basins. Soils westward of the carbonate rich Sagavanirktok River are more acidic. These lower
pH soils are likely to support enhanced chemical weathering
and increased concentrations of dissolved metals and DOC.
In contrast with Cu, Fe, Pb and Zn, concentrations of dissolved Ba remained relatively constant in each river throughout
the sampling period with the Sagavanirktok River at 234 ⫾ 14
nmol/L, the Kuparuk River at 171 ⫾ 16 nmol/L and the
Colville River at 376 ⫾ 24 nmol/L. These results suggest that
concentrations of dissolved Ba are not significantly influenced
by increases in water discharge or concentrations of TSS and
DOC. Concentrations of dissolved Ba vary in proportion to
concentrations of total particulate Ba on a dry weight basis for
the Sagavanirktok, Kuparuk and Colville rivers and suggest
that dissolved Ba concentrations may be controlled by concentrations of particulate Ba. Separate analyses with additional
samples are needed to determine what fraction of the particulate Ba is exchangeable. If this fraction proves to be proportional to concentrations of total particulate Ba, then dissolved
Ba levels may be controlled by ion exchange with suspended
clay minerals.
Concentrations of particulate Cu, Fe, Pb, Zn and OC do not
follow the trends observed for the dissolved fraction and maintain relatively constant concentrations (on a dry weight basis)
throughout the spring floods. In addition, concentrations of
particulate metals correlate well with particulate Al but are
poorly correlated with POC showing that clays rather than
organic matter control particulate metal concentrations.
Increases in water discharge during the spring flood events
488
R. D. Rember and J. H. Trefry
coincide with increased concentrations of TSS, dissolved metals, and OC and account for a large fraction of the annual
transport. In the Kuparuk and Sagavanirktok rivers, ⬎16% of
the water flow and ⬎80% of the annual sediment discharge
occurs during 3- and 12-d periods, respectively. These large
pulses of water carry more than one-third of the annual dissolved load of Cu, Fe, Pb, Zn and OC and ⬎80% of the
particulate metals to the coastal Beaufort Sea.
Acknowledgments—We thank Bob Trocine and Michelle McElvaine
from Florida Institute of Technology, Mark Savoie and Gary Lawley of
Kinnetics Laboratories, Anchorage, and John Brown of Battelle for
their efforts in logistical support, sample collection and preparation.
Our thanks to BP/ARCO for providing lab space, lodging and logistics
throughout the project. We also thank Phillips Petroleum for providing
helicopter support to sample the Colville River. We especially thank
Dick Prentki of MMS, Anchorage, for numerous discussions and input
to the research. This study was supported by Minerals Management
Service (Contract No. 143501-99-CT-30998). Finally, we thank two
anonymous reviewers and the associate editor for their helpful and
thoughtful comments.
Associate editor: K. F. Falkner
REFERENCES
Arnborg L., Walker H. J., and Peippo J. (1967) Suspended load in the
Colville River, Alaska, 1962. Geog. Annaler. 49A, 131–144.
Boyer E. W., Hornberger G. M., Bencala K. E., and McKnight D. M.
(1997) Response characteristics of DOC flushing in an alpine catchment. Hydrol. Process. 11, 1635–1647.
Brown J., Grant C. L., Ugolini F. C., and Tedrow J. C. F. (1962)
Mineral composition of some drainage waters from arctic Alaska. J.
Geophys. Res. 67 (6), 2447–2453.
Cabaniss S. E. and Shuman M. S. (1988) Copper binding by dissolved
organic matter: I. Suwanee River fulvic acid equilibria. Geochim.
Cosmochim. Acta 52, 185–193.
Cauwet G. and Meybeck M. (1987) Seasonal fluctuations of carbon
levels in a temperate river: The Loire (France). In Transport of
Carbon and Minerals in Major World Rivers, Part 4 (eds. E. T.
Degens, S. Kempe, and G. Weibin), pp. 349 –358. Mitt. Geol.Palaont. Inst. Univ. Hamburg, SCOPE/UNEP Soderbd.
Cauwet G. and Sidorov I. (1996) The biogeochemistry of the Lena
River: Organic carbon and nutrients distribution. Mar. Chem. 53,
211–227.
Craig P. C. and McCart P. J. (1975) Classification of stream types in
Beaufort Sea drainages between Prudhoe Bay, Alaska, and the
Mackenzie Delta, N.W.T., Canada. Arct. Alp. Res. 7, 183–198.
Dai M. H. and Martin J. M. (1995) First data on trace metal level and
behavior in two major Arctic river-estuarine systems (Ob and Yenisey) and in the adjacent Kara Sea, Russia. Earth Planet. Sci. Lett.
131, 127–141.
Depetris P. J. and Paolini J. E. (1991) Biogeochemical aspects of
South American rivers: The Parana and the Orinoco. In Biogeochemistry of Major World Rivers, SCOPE Report 42 (eds. E. T.
Degens, S. Kempe, and J. E. Richey), pp. 105–125. John Wiley,
Chichester, UK.
Dupre B., Viers J., Dandurand J.-L., Polve M., Benezeth P., Vervier P.,
and Braun J. J. (1999) Major and trace elements associated with
colloids in organic-rich river waters: Ultrafiltration of natural and
spiked solutions. Chem. Geol. 160, 63– 80.
Gordeev V. V., Martin J. M., Sidorov I. S., and Sidorova M. V. (1996)
A reassessment of the Eurasian input of water, sediment, major
elements, and nutrients to the Arctic Ocean. Am. J. Sci. 296, 664 –
691.
Guay C. K. and Falkner K. K. (1998) A survey of dissolved barium in
the estuaries of major Arctic rivers and adjacent seas. Cont. Shelf
Res. 18, 859 – 882.
Guieu C., Huang W. W., Martin J. M., and Yong Y. Y. (1996) Outflow
of trace metals into the Laptev Sea by the Lena River. Mar. Chem.
53, 255–267.
Hinzman L. D., Kane D. L., Gieck R. E., and Everett K. R. (1991)
Hydrologic and thermal properties of the active layer in the Alaskan
Arctic. Cold Reg. Sci. Technol. 19, 95–110.
Kling G. W., O’Brien J., Miller M. C., and Hershey A. (1992) The
biogeochemistry and zoogeography of lakes and rivers in arctic
Alaska. Hydrobiol. 240, 1–14.
Kriet K., Peterson B. J., and Corliss T. L. (1992) Water and sediment
export of the Kuparuk River drainage of the North Slope of Alaska.
Hydrobiol. 240, 71– 81.
Lock M. A., Ford T. E., Fiebig D. M., Miller M. C., Hullar M.,
Kaufman M., Vestal J. R., Peterson B. J., and Hobbie J. E. (1989)
A biogeochemical survey of rivers and streams in the mountains
and foothills province of arctic Alaska. Arch. Hydrobiol. 115,
499 –521.
Martin J. M. and Windom H. (1991) Present and future roles of ocean
margins in regulating marine biogeochemical cycles of trace elements. In Ocean Margin Processes in Global Change (eds. R. F. C.
Mantoura, J. M. Martin, and R. Wollast), pp. 45– 67. John Wiley,
New York.
Martin J. M., Guan D. M., Elbaz-Poulichet F., Thomas A. J., and
Gordeev V. V. (1993) Preliminary assessment of the distribution of
some trace elements (As, Cd, Cu, Fe, Ni, Pb and Zn) in a pristine
aquatic environment: the Lena River estuary (Russia). Mar. Chem.
43, 185–199.
McNamara J. P., Kane D. L., and Hinzman L. D. (1998) An analysis of
streamflow hydrology in the Kuparuk River Basin, Arctic Alaska: A
nested watershed approach. J. Hydrol. 206, 39 –57.
Meybeck M. (1982) Carbon, nitrogen, and phosphorus transport by
world rivers. Am. J. Sci. 282, 401– 450.
Michaelson G. J., Ping C. L., Kling G. W., and Hobbie J. E. (1998) The
character and bioactivity of dissolved organic matter at thaw in the
spring runoff water of the arctic tundra north slope, Alaska. J.
Geophys. Res. 103 (D22), 28939 –28946.
Moran S. B. and Woods W. L. (1997) Cd, Cr, Cu, Ni and Pb in the
water column and sediments of the Ob-Irtysh Rivers, Russia. Mar.
Poll. Bull. 35, 270 –279.
Nakashima S., Sturgeon R. E., Willie S. N., and Berman S. S. (1988)
Determination of trace elements in seawater by graphite-furnace
atomic absorption spectrometry after preconcentration by tetrahydroborate reductive precipitation. Anal. Chim. Acta 207, 291–
299.
Osterkamp T. E. and Payne M. W. (1981) Estimates of permafrost
thickness from well logs in Northern Alaska. Cold Reg. Sci. Technol.
5, 13–27.
Parkinson R. J. (1977) Genesis and Classification of Arctic Coastal
Plain Soils, Prudhoe Bay, Alaska. M.S. thesis, Ohio State University,
Columbus.
Payne T. G., Dana S. W., Fischer W. A., Yuster S. T., Krynine P. D.,
Morris R. H., Lathram E. and Tappan H. (1951) Geology of the
arctic slope of Alaska. U.S. Geol. Surv. Oil Gas Investig., map
OM-126.
Pokrovsky O. S. and Schott J. (2002) Iron colloids/organic matter
associated with transport of major and trace elements in small boreal
rivers and their estuaries (NW Russia). Chem. Geol. 190, 141–179.
Robinson R. S. and Johnsson M. J. (1997) Chemical and physical
weathering of fluvial sands in an Arctic environment: Sands of the
Sagavanirktok River, North Slope, Alaska. J. Sed. Res. 67, 560 –570.
Telang S. A. (1985) Transport of carbon and minerals in the Mackenzie
River. In Transport of Carbon and Minerals in Major World Rivers,
Part 3 (eds. E. T. Degens, S. Kempe, and R. Herrera), pp. 337–344.
Mitt. Geol.-Palaont. Inst. Univ. Hamburg, SCOPE/UNEP Soderbd.
Telang S. A., Pocklington R., Naidu A. S., Romankevich E. A.,
Gitelson I. I., and Gladeyshev M. TI. (1991) Carbon and mineral
transport in major North American, Russian Arctic, and Siberian
Rivers: the St. Lawrence, the Mackenzie, the Arctic Alaskan Rivers,
the Arctic Basin Rivers in the Soviet Union, and the Yenisei. In
Biogeochemistry of Major World Rivers, SCOPE Report 42 (eds.
E. T. Degens, S. Kempe, and J. E. Richey), pp. 75–104. John Wiley,
Chichester, UK.
Trace metals and organic carbon in Alaskan Arctic rivers
Trefry J. H., Trocine R. P (1991) Collection and analysis of marine
particles for trace elements. In Marine Particles: Analysis and Characterization, Geophysical Monograph (eds. J. Hurd and D. Spencer),
pp. 311–315.
U.S. Geological Survey (1971–2001) Water Resources Data for
Alaska. Anchorage, AK.
489
Walker D. A. and Webber P. J. (1979) Relationships of soil acidity and
air temperature to the wind and vegetation at Prudhoe Bay, Alaska.
Arctic 32, 224 –236.
Zhulidov A. V., Headley J. V., Robarts R. D., Nikanorov A. M., Ishenko
A. A., and Champ M. A. (1997) Concentrations of Cd, Pb, Zn and Cd
in pristine wetlands of the Russian Arctic. Mar. Poll. Bull. 35, 242–251.