Long-term sediment traps were deployed from September 2005 to August 2007 in the northern Laptev Sea to assess the annual variability in vertical export of particulate organic carbon (POC). The second year of deployment coincided with the record low in Arctic summer ice extent reached in 2007 that resulted in an increase in marine primary production over the Siberian shelves. POC export fluxes increased during ice melt in 2007, leading to a ∼2-fold increase in annual POC export relative to 2005–2006 over the continental slope of the Laptev Sea. These results suggest that the continuous decrease of sea ice extent could sustain increased POC export in the northern Laptev Sea and adjacent seas, potentially altering marine ecosystem structure in the Siberian Arctic.
 A substantial reduction in sea ice extent has recently been observed in the Arctic Ocean, with the 2007 minimum ice extent 23% below the previous low of 2005 and corresponding to the largest recorded decrease in minimum sea ice extent [Arrigo et al., 2008; National Snow and Ice Data Center (NSIDC), Arctic sea ice shatters all previous record lows, http://nsidc.org/news/press/2007_seaiceminimum/20071001_pressrelease.html, Boulder, Colo., 2007]. Using satellite-derived sea ice, sea surface temperature, and chlorophyll a coupled to a primary production algorithm parameterized for Arctic waters, Arrigo et al.  showed that this loss of sea ice led to a marked increase in annual primary production due to a larger open water area and a longer phytoplankton growing season, particularly over the Siberian shelves. This continuous and rapid reduction of Arctic sea ice and the associated increase in primary production may cause an increase in particulate organic carbon (POC) export over the Arctic shelves and have significant repercussions on marine ecosystems and pelagic-benthic coupling [Sakshaug, 2004; Wassmann et al., 2004; Carmack and Wassmann, 2006; Grebmeier et al., 2006].
 Long-term sediment traps were deployed through a collaborative effort between ArcticNet, a Network of Centres of Excellence of Canada, and the Nansen and Amundsen Basins Observational System (NABOS) project to assess the annual variability in vertical POC export in the northern Laptev Sea, Siberian Arctic. In this paper, we report on enhanced POC export during the ice melt period of 2007 in the northern Laptev Sea. The considerable changes currently observed in this region reflect what may occur in upcoming years in regions undergoing similar reduction in ice cover, stressing the importance of monitoring the Siberian Arctic to assess the eventual impact of climate change on carbon export in the Arctic Ocean.
2. Experimental Method
2.1. Ice Concentration
 Weekly averaged sea ice concentrations above the mooring (79°55′N; 142°21′E; Figure 1) were computed from daily sea ice concentrations provided by the NSIDC for the complete deployment period. Data were derived from DMSP-F13 Special Sensor Microwave/Imager (SSM/I) daily brightness temperatures at a grid cell size of 25 × 25 km. Sea ice concentration retrievals were based on the NASA Team algorithm described by Cavalieri et al. .
2.2. Chlorophyll a Concentration
 Satellite-derived cholorophyll a concentration (CHL) data were obtained from the Globcolour project (http://www.globcolour.info/data_access.html). CHL was retrieved using the semi-analytical GSM algorithm [Garver and Siegel, 1997; Maritorena et al., 2002] applied to merged water-leaving reflectances spectra from SeaWiFS, MODIS and MERIS datasets [Maritorena and Siegel, 2005]. GSM was selected over usual empirical algorithms to minimize the impact of optical constituents (colored detrital material (CDM) and non algal particles) that do not covary with CHL on the CHL patterns. Although GSM has not been validated for the Arctic Ocean, the semi-empirically-derived CHL patterns are more realistic than the one obtained using empirical algorithms in waters dominated by CDM absorption [Bélanger et al., 2008]. Merged level-3 products, which are distributed on a 4.63 km resolution Integerized Sinusoidal (ISIN) grid, were mapped on a Lambert conic conformal projection using the Generic Mapping Tools software (http://gmt.soest.hawaii.edu/).
2.3. Sediment Traps
 Two Technicap PPS 3/3 (0.125 m2 aperture) sediment traps were deployed from September 2005 to August 2006 and redeployed from September 2006 to August 2007 at 175 and 850 m respectively on the slope of northern Laptev Sea at station M3 (1350 m deep; Figure 1). The traps were deployed and recovered during the NABOS annual expeditions on board the icebreaker Kapitan Dranitsyn in 2005 and 2006 and on board Viktor Buynitskiy in 2007. A current meter (Aanderaa, RCM-11) was deployed under the shallow sediment trap to record current speed and direction throughout deployment.
 The sediment trap sample cups were filled with filtered (Whatman GF/F, nominal pore size of 0.7 μm) seawater poisoned with formalin (5% v/v) buffered with sodium borate and adjusted to a salinity of 35 with NaCl to preserve samples during deployment and after recovery. In the laboratory, swimmers were removed from the samples using a stereoscopic microscope. The samples were split in replicate subsamples and filtered on pre-weighed GF/F filters pre-combusted for 4 h at 450°C. Triplicate filters were rinsed with distilled water to remove salt, dried for 12 h at 60°C, measured for dry weight, and exposed for 12 h to concentrated HCl fumes to remove inorganic carbon. The filters were dried at 60°C and particulate organic carbon (POC) and particulate nitrogen (PN) were measured on a Perkin Elmer CHNS 2400 Series II analyser. Daily mass (mg m−2 d−1) and POC (mg C m−2 d−1) fluxes were averaged for each collection period (coefficient of variation = 13.6% for mass and 12.8% for POC). POC fluxes were not corrected for solubilization of organic material and should be considered as minimal estimates.
3. Results and Discussion
 Thanks to serendipity, the deployment of the sediment traps encompassed a year of relatively normal ice melt in 2006 and the spectacular regression of the ice cover in 2007 (Figure 2a). The annual vertical export of POC in the northern Laptev Sea was higher in 2006–2007 than in 2005–2006, mainly due to an increase in POC export during and following ice melt in 2007 (Figures 2c and 2d and Table 1). POC fluxes at 175 m were ∼10-fold higher during the last 2 weeks of July 2007 than over the same period in 2006. The increase in POC export was even larger at 850 m for the same period, with POC fluxes ∼24-fold higher in July 2007 than in July 2006. An increase in POC export was also observed following ice melt at 175 m when POC fluxes were ∼8-fold higher in August 2007 than in August 2006. These increases in POC export are attributable to enhanced phytoplankton production due to an earlier start of the phytoplankton growing season above the mooring in 2007, as indicated by satellite-derived CHL (Figure 3). The complete ice melt observed above the mooring in 2007, in contrast with an incomplete melt in 2006, resulted in an earlier phytoplankton bloom and enhanced phytoplankton production, and contributed to the increase in POC export in the northern Laptev Sea (Figures 2 and 3).
Table 1. Annual POC and Mass Fluxes Measured in 2005–2006 and 2006–2007 in the Northern Laptev Sea
Trap Depth (m)
POC Flux (g C m−2 a−1)
Mass Flux (g m−1 a−1)
 Interestingly, the relative increase in total mass fluxes was greater than the increase in POC fluxes, resulting in low %POC in the exported material (Figures 2c and 2d). In the Laptev Sea, ice contributes to the transport of 180 × Gg C a−1 of terrigenous POC into the Arctic Basin [Eicken, 2004]. The increase in total mass and POC fluxes during the melt period of 2007 may therefore reflect a larger release of particulate material from melting ice since completely ice-free conditions were reached whereas ice concentrations remained >20% during the melt period of 2006. The complete ice melt in 2007 perhaps also resulted in a larger release of ice algae that may have contributed to the enhanced total and POC fluxes. Increased C:N ratios during and following ice melt suggest that particles and ice algae released from sea ice contributed to the total mass and POC fluxes, considering that transparent exopolymeric substances produced by sinking ice algae are characterized by high C:N ratio [Engel and Passow, 2001]. Moreover, elevated POC fluxes at both 175 and 850 m in July 2007 probably reflected the fast sinking of large phytoplankton cells, ice algae and particles released from the melting sea ice.
 In contrast to the melting period, POC fluxes were lower under ice cover from November 2006 to March 2007 than during the same period the previous year (Figures 2c and 2d). Elevated POC and total mass fluxes at both depths from January to March 2006 possibly reflect resuspension events caused by strong northeastward currents allowing POC transport to the continental slope, while currents were weaker and POC export remained low during the same period in 2007 (Figures 2b–2d).
 In 1995–1996, Fahl and Nöthig  measured POC export under almost permanently ice-covered conditions over the Lomonosov Ridge at a distance of 140 km NNW from our mooring location. Notwithstanding the effect of differences in trap design on the estimation of fluxes (see Hargrave et al.  for arctic fluxes), the annual POC flux they measured under heavy ice cover (1.5 g m−2 a−1 at 150 m) was approximately 3× and 6× lower than those we measured a decade later in 2005–2006 and 2006–2007 respectively (Table 1). Fahl and Nöthig  identified an interval of increased fluxes in mid-July/August caused by an increase in primary production, whereas increased POC fluxes at 175 m were observed as early as mid-June in 2006 and 2007, indicating that primary production started earlier and that the growing season was longer in the northern Laptev Sea in response to ice cover reduction. A second interval of increased fluxes in September/October was caused by an increased influence of the Lena River and contributed most of the annual export in 1996 [Fahl and Nöthig, 2007]. The Lena River, the second largest Arctic river in terms of annual discharge, sustains an average annual POC discharge of 1.2 Mt a−1 with the maximum discharge occurring during the spring break-up in June [Holmes et al., 2002; Rachold et al., 2004]. In our study, an increase in POC flux in August 2006 possibly reflected the maximum POC discharge from the Lena River. An earlier input of POC in 2006 in comparison to 1995 may reflect an earlier break-up of the Lena River, or may be explained by our mooring being located closer to the Lena River than the mooring deployed on the Lomonosov Ridge by Fahl and Nöthig . In 2007, POC input from the Lena River discharge was not apparent, probably masked by the interval of high POC fluxes associated with ice melt and enhanced primary production. River discharge and POC input from river discharge are expected to increase as a result of climate warming, particularly in the Siberian Arctic where the largest rivers in terms of water discharge enter the Arctic Ocean [Holmes et al., 2002; Peterson et al., 2002]. However, the substantial increase in POC export from mid-June to mid-August 2007 suggests that decreased ice cover and enhanced primary production will have a larger impact on POC export in the northern Laptev Sea than a potential increase of the Lena River discharge.
 Observations indicate that the Arctic sea ice decline is occurring faster than forecasted by model simulations, with current summer minima approximately 30 years ahead of the mean model forecast [Stroeve et al., 2007; Wang and Overland, 2009]. Since the ongoing decline in sea ice is largely occurring in the Siberian Arctic, the increase in POC export observed in the northern Laptev Sea may reflect the impending impact of decreasing ice cover on POC export for other regions of the Arctic Ocean. It is likely that areas currently experiencing similar decreases in sea ice extent and increases in primary production, like the outer shelves of the Kara and East Siberian Seas [Arrigo et al., 2008], are also sustaining increases in POC export.
 A larger POC export over the Arctic shelves is likely to increase carbon input to the benthic ecosystem, amplify pelagic-benthic coupling, and potentially affect benthic carbon cycling [Grebmeier et al., 2006; Morata and Renaud, 2008]. However, retention of POC in the upper water column or export to the deep ocean depends largely on the match or mismatch between the seasonal dynamics of the phytoplankton community and the grazing impact of zooplankton [Wassmann et al., 1996, 2004; Sakshaug, 2004]. Improved conditions for zooplankton due to reduced ice cover and increased primary production may result in a better match between the phytoplankton bloom and the zooplankton stock. For example, greater food availability and higher temperature due to reduced ice cover contributed to earlier recruitment, advanced population development and improved reproductive success for copepods in the North Water polynya relative to Barrow Strait, a non-polynya region in the Canadian Archipelago [Ringuette et al., 2002]. Under these conditions, 79% of particulate primary production was retained by zooplankton in the upper 50 m of the polynya [Tremblay et al., 2006]. Thus, elevated annual POC export as observed in 2007 in the northern Laptev Sea will depend on the adaptation of zooplankton to an early phytoplankton bloom.
 In summary, Fahl and Nöthig  determined that the lithogenic character of the exported material was pronounced on the Lomonosov Ridge a decade ago due to the high input of terrigenous matter by the Eurasian rivers and the relatively low primary production. The large increase in POC export in 2007 suggests that the northern Laptev Sea is in transition to a more productive state due to the large reduction in sea ice extent. However, future increases in primary production may be capped by the exhaustion of present surface nutrient inventories, as suggested by Arrigo et al. . Our results indicate that considerable changes in the organic carbon cycle are occurring in the northern Laptev and reinforce the importance of monitoring the Siberian Arctic to assess the actual impact of climate change on carbon export in the Arctic Ocean.
 We thank W. Chan and D. Barber of the Centre for Earth Observation Science at the University of Manitoba for ice data. This study was funded by ArcticNet, a Network of Centres of Excellence of Canada. Logistics in the Arctic Ocean were provided by ArcticNet and the Nansen and Amundsen Basins Observational System (NABOS) project at the International Arctic Research Center (IARC-University of Alaska, Fairbanks). CL was supported by a postdoctoral grant from the Fonds Québécois de Recherche sur la Nature et la Technologie (FQRNT). This is a joint contribution to Québec-Océan at Université Laval, ArcticNet, NABOS, and the Canada Research Chair on the response of marine arctic ecosystems to climate warming.