Accumulation rates of pollutants and constituents of marine aerosol during the Holocene at Lake ArresjØen, Svalbard, Norway

Stephen A. Norton* (University of Maine), Eirik Fjeld and Sigurd Rognerud (Norwegian Institute for Water Research), and George L. Jacobson Jr. (University of Maine)

Abstract

We collected and examined a ²¹⁰Pb- and ¹⁴C (AMS)-dated lake-sediment core from Danskøye, Svalbard, Norway in 1996. The 72-cm core represents approximately 10,500 calendar years. Radiocarbon dates indicate that the sediment accumulation rate (cm/yr) decreased slightly down-core, generally consistent with compaction (% H₂0); higher frequency variations in % H₂O are associated with % organic matter. Mercury (Hg) concentration increased slowly over the last 5,000 years in parallel to a long-term increase in % organic (Hg (ng/g) = 4.15*LOI(%) + 14.1). Hg was 50 ng/g at about 5,000 yr B.P. and increased to 230 ng/g at the surface. The accumulation rate increased slowly starting in the late 1800s (A.D.) from about 0.25 ng Hg/cm²/yr to a sharp peak about 1975 (A.D.) (0.65 ng/cm²/yr) and then declined to 0.4 ng/cm²/yr by 1996. Over the same time span, the concentration of Pb ranged from ca. 30 to 58 ug/g, with a very weak relationship to % organic matter. The accumulation rate increased from pre-industrial values of 0.03 mg Pb/cm²/yr to ca. 0.11 mg Pb/cm²/yr in the 1970s and then decreased. The peak in accumulation rates of Hg and Pb results partly from a slight increase in sediment accumulation rate and partly from an independent increase in the metal fluxes. Although about an order of magnitude lower in absolute quantity, recent increases in Hg accumulation rates are temporally co-incident with those in continental Europe and southern Scandinavia, implying long-distance transport.

Longer-term long-wavelength variations occur in the concentration and accumulation rates of Br, I, and Se, which are of marine aerosol origin. Their concentrations and accumulation rates vary with sediment accumulation rate, % organic matter, and possibly changes in past climate.

Introduction

Anthropogenic pollutants (Cd, Hg, Pb, PAHs, PCBs, etc.) in the Arctic are of concern because of the bio-accumulation and magnification of these pollutants in Arctic flora and fauna (e.g., Science Tot. Environ. 245). Accordingly, these pollutants are the focus of many Arctic research programs (e.g., ARCUS [U.S.], ARTIS [Canada]). Knowledge of the history of deposition at high latitudes, and sources and redistribution of the pollutants, is critical to understanding biological impacts. Studies of lake sediments and of ice cores have contributed to understanding these concerns. Arctic lake-sediment records (chemistry, pollen, diatoms, grain size, etc.) have also been used to reconstruct climate. We report here on the deposition history of Hg and Pb in a ²¹⁰Pb- and ¹⁴C-dated lake-sediment core from the Svalbard Archipelago, emphasizing the industrial period. Secondarily, we demonstrate chemical linkages between the marine environment and lake-sediment record over millenia.

Our research was supported by the Norwegian Institute for Water Research, Oslo and the U. S. National Science Foundation (Grant ATM#9634345 to the University of Maine).

Methods

Site Description: We collected two sediment cores from Lake Årresjøen, on Danskøye Island, northwesternmost Svalbard Archipelago in August of 1996. Arresjøen is approximately 30 m deep, is ice-locked 9-10 months per year, and is fed by one major inlet which drains a small pond. The lake drains directly to the North Atlantic, which lies approximately 500 m to the west. The catchment ranges from 30 to 250 masl. The watershed, 10 times the area of the lake, is underlain by Precambrian felsic gneisses, with thin discontinuous till and more than 50% exposure of bedrock. No macrophyte vegetation was observed. Water quality was dominated by marine aerosol input. Alkalinity was <50 meq/L.

Coring and pre-analysis processing: Cores were obtained in 1996 with a 10-cm-diameter stationary piston corer suspended from a floating platform. Sediment was sectioned in the field and stored in WhirlPakÔ plastic bags. We sectioned at 0.5 cm intervals from 0 to 20 cm, in 1.0 cm intervals from 20 to 40 cm, and in 2 cm intervals from 40 to 70 cm. Sediment was stored dark until it was returned to Maine and processed. Samples were dried at 35^oC to constant weight for determination of % H₂O. This same fraction was used for determination of the Hg and Pb concentrations. Another dried aliquot was heated to 110^oC and ashed at 550^oC for four hours to determine loss-on- ignition (LOI), an approximation for organic concentration.

Dating of sediment: ²¹⁰Pb age-dating was performed by direct gamma analysis on one 72 cm core by P. Appleby at Liverpool University, U.K. (Appleby and Oldfield, 1978; Appleby et al., 1986). We dated six older intervals by AMS ¹⁴C dating (Beta Analytical Inc.).

Determination of metals: We digested 0.25 g aliquots of each interval using microwave-assisted acid digestion for Hg, and HNO₃ digestion for Pb. Hg was determined with cold-vapor atomic absorption. Pb was determined with flameless atomic absorption spectrophotometry. Precision and accuracy were checked with standard reference material (SRM) analyses, blanks, duplicates, replicate analyses, and standard checks during a run. Two SRMs were digested per core (minimum), a blank was prepared every 14 samples, and 1 of every 9 samples is duplicated. Check standards were run once for every 10 analyses. Concentrations are +/-10% for Hg and +/-5% for Pb. We determined Se, Br, and I using INAA.

Accumulation rates: The net accumulation rate for Hg or Pb (g/cm²/yr) equals:

[(mass of sediment/interval/cm²)(concentration of Hg or Pb in interval)]/(years/interval) (1)

This total flux includes three components. (1) The natural background flux (Hg_b and Pb_b). Commonly, variations in % organic matter in sediment match variations in concentrations of Hg or Pb. We used pre-1850 A.D. sediment for this normalization. (2) Variations in the gross sedimentation rate, which cause variations in the flux of metals (Hg_v and Pb_v). This linkage is estimated by normalizing the sedimentation rate. However, the sedimentation rate for the core was so low that resolution of fewer than 25 years was not possible, except in the upper few cm. (3) Variations in atmospheric deposition of anthropogenic Hg or Pb directly to the lake, and leaching of anthropogenic Hg and Pb from the watershed to the lake (Hg_a and Pb_a). Thus, for example, for Hg:

Total Hg (ng Hg/cm²/yr) = Hg_B + Hg_V + Hg_A (2)

Results and Discussion

Dating: Equilibrium of total ²¹⁰Pb activity with the supporting ²²⁶Ra (corresponding to about 150 years accumulation) was reached at a depth of 3.75 cm. Unsupported ²¹⁰Pb_u declined nearly exponentially with depth, indicating uniform sediment accumulation during the past 100 years or more, except for a small disturbance at 1.25-1.75 cm. The ²¹⁰Pb chronology we used was based on the CRS dating model of Appleby and Oldfield (1978) although their CIC model gave nearly equivalent ages. During the last 100 years, the mass accumulation rate had a mean value of 0.0013+/-0.0002 g/cm²/yr. The years represented by intervals of sediment range from 2 to 25 years between the uppermost interval and 3.5 cm, which is dated at approximately 1850 A.D. The lowest ¹⁴C-dated interval of the core (68-70 cm) yielded 10,490 cal. B.P. The age-depth curve based on six AMS dates, corrected for compaction, is nearly linear and indicates an average mass accumulation rate of 0.0018 g/cm²/yr, slightly higher than that for the last 150 years. The loss-on-ignition (LOI) decreases exponentially from 25% at the surface to a nearly constant 7-8% prior to 6,000 cal. B.P. Water decreases from 100 to nearly 80% in parallel with LOI.

Metals: Hg was measured on all samples from 0 to 20 cm. The concentration declined from about 230 ng/g at the surface to 50 ng/g at 20 cm (Figure 1). Regressing variations of organic matter and Hg concentrations between 5 and 20 cm yields the relationship:

Hg (ng/g) = 4.16(LOI) + 14 (R² = 0.4) (3)

Although the relationship is statistically weak, it indicates that the increase in the flux of Hg to the sediment (Figure 2) is controlled partly by changes in organic content. The accumulation rate of Hg prior to industrial times was about 0.04 ng/cm²/yr, increasing slowly during the last few centuries to about 0.25 ng/cm²/yr by the late 1800s A.D. Much of this increase is linked with organic content. Deposition increased sharply in the middle of the 20^th century to at least 0.6 ng/cm²/yr, apparently as a result of increased atmospheric deposition. The decline in the accumulation rate of Hg starting about 1975 resembles a similar decline in eastern North America, documented in studies of lake-sediment cores, ombrotrophic peat cores, and forest floor soil chemistry (Engstrom et al., 1997; Norton et al., 1997, Evans et al., this volume). The abruptness of the increase and subsequent decrease of Hg is consistent with relatively abrupt changes in atmospheric deposition, and with a watershed that has little capability of retention of Hg because of the lack of organic soil. That is, changes in atmospheric deposition of Hg result almost immediately in proportional changes in sedimentation of Hg in the lake. The absolute and proportional increases in concentration and deposition rate of Hg are remarkable and demonstrate the global atmospheric transport of Hg.

Pb was determined for alternate samples from 0 to 30 cm. The concentration of Pb varied unsystematically down core (Figure 1) and appears to be only weakly controlled by LOI:

Pb (ug/g) = 0.99(LOI) + 26.4 (4)

The intercept for LOI = 0% is a typical concentration for Pb in felsic metamorphosed granites and volcanic rocks, which are common in the catchment. A very small increase (<0.1 ug/g) in the accumulation rate of Pb occurred during the last 150 years but the temporal resolution is poor because of alternate sample analysis and slow sedimentation. Rognerud et al. (1998) and Steinnes et al. (1997) demonstrated a sharp decline in the deposition of Pb from southern to northern Norway. Extrapolation to Svalbard suggests little deposition of excess Pb at the coring site. Apparently any atmospheric deposition of Pb from pollution is so limited at the latitude of Danskøye that the signal is nearly lost in natural variation of Pb originating from the local bedrock.

Marine Aerosols: Br, I, and Se were determined on all samples from 0 to 20 cm. Concentration of Se ranged from 2 to 5 mg/g with no enrichment in recent sediment, consistent with the findings of Rognerud et al. (1997) for high latitude lakes in Norway, Russia, and Svalbard. However, Se and Se normalized to LOI (Se/LOI) had a cyclical pattern with a wave length of 500-1000 years that was closely paralleled by Br, and to a smaller extent by I. Consequently, the accumulation rate of Se in sediment has a cyclical value (Figure 3). These three elements at Lake Arresjøen are derived primarily from marine aerosols and are subsequently incorporated into organic matter. Consequently, we hypothesize that lake sediment records may archive a record of the strength of the flux of marine aerosols to the catchment. This flux is controlled by the vigor of atmospheric circulation and proximity to open ocean, both of which are important climatic variables.

Conclusions

Lake-sediment records from Lake Arresjøen, Svalbard indicate increased deposition of Hg followed by a decline, both related to the period of known pollution in western Europe and eastern North America. The anthropogenic component of the deposition is about 0.1 times the anthropogenic flux farther south, although the relative increase at the site is more than 100%. For Pb the increase in accumulation rate during recent time is relatively small and any pollution signal is indistinct. Both Hg and Pb vary with organic matter concentration in pre-industrial sediment, a relationship that must be considered when using sediments to evaluate pollution history. Deposition of Hg is proportionately higher than for Pb at this high latitude, compared to southern Scandinavia or eastern North America. This probably relates to two factors. First, Pb emissions are typically in the 1-micron-size class, whereas much Hg is emitted as a vapor. Accordingly, the residence time of Pb would be shorter in the atmosphere. Second, condensation of Hg at high latitude may enhance its atmospheric deposition.

The sediment record from Lake Arresjøen contains strong evidence of cyclical and parallel variations in the concentrations and accumulation rates of Br, I, and Se. These elements are derived primarily from marine aerosol input to the catchment. Variations in their sedimentation, partly associated with organic matter, may be controlled by climate factors including strength of atmospheric circulation and proximity to the ocean. Under some geographic conditions, the extent of sea ice may control this flux. Thus, sediment chemistry may be useful for climate reconstruction, and for the reconstruction of sea ice. Such information would be extremely useful for climate modeling.

References

Appleby, P. G. et al. (1986), Hydrobiologia. 141:21-27.

Appleby, P. G., Oldfield, F. (1978), Catena. 5:1-8.

Engstrom, D. R., Swain, E. B. (1997), Envir. Sci. Tech. 31:960-967.

Evans, G. C. et al. (2000), This volume.

Norton, S. A., Evans, G. C., Kahl, J. S. (1997), Water, Air, and Soil Poll. 100: 271-286.

Rognerud, S. et al. (1997), Can. J. Fish. Aquat. Sci. 55: 1512-1523.

Science of the Total Environment. (2000), 245.

Steinnes, E. et al. (1997), Sci. Total Environ. 205: 255-266.

Figure 1: Concentration of Hg and Pb versus depth from a core from Lake Arresjøen, Svalbard.

Figure 2: Accumulation rate of Hg in sediment at Lake Arresjøen, Svalbard.

Figure 3: Accumulation rate of Se in sediment from Lake Arresjøen, Svalbard.