Numerical simulations of future climate performed by use of general circulation models all agree that climate warming will occur first and most intensively in arctic and subarctic regions (e.g., Manabe and Stouffer, 1993; Cattle and Crossley, 1996; Shindell et al., 1999; Flato and Boer, 2001). Surface air temperature observations reveal that the largest increase in recent decades has occurred over Northern Hemisphere land areas from about 40 to 70°N (Serreze et al., 2000). Due to warming of the world oceans (Levitus et al., 2000) the sea-ice cover in the Arctic has decreased by ∼14% since the 1970s (Johannessen et al., 1999). This decrease has led to a prolongation of the open-water period off the north coast of Russia and in the Greenland Sea, the Barents Sea, and the Sea of Okhotsk (Parkinson, 1992). Higher atmospheric temperatures will increase the transport of water vapor toward the North Pole, resulting in an increase in precipitation and freshwater supply to high latitudes. This prediction agrees well with the observed general increase in precipitation in the 55–85°N latitude band during the last century (Serreze et al., 2000).
The research and monitoring station ZERO (Zackenberg Ecological Research Operations) was established in 1995 in Young Sound, northeast Greenland, with a view to monitoring effects of climate changes in a high arctic environment. The study area is considered to be a sensitive indicator of climate change because of its contact with water masses from the Greenland Sea and direct meltwater flux from the Greenland Ice Sheet. It is believed that the predicted climate changes will dramatically alter the conditions in the Greenland Sea as well as coastal hydrological conditions. A long-term marine monitoring program will supplement the ongoing monitoring program of the terrestrial environment in 2003. Detailed quantitative studies of the marine ecosystem processes are being conducted in the area, including seasonal measurements of primary production by phytoplankton (Rysgaard et al., 1996, 1999), sea-ice microalgae (Kühl et al., 2001; Rysgaard et al., 2001; Glud et al., 2002a), brown algae (Borum et al., 2002), benthic microalgae (Glud et al., 2002b), and coralline red algae (Roberts et al., 2002). Furthermore, seasonal measurements of mineralization rates in the water column (Rysgaard et al., 1999; Levinsen et al., 2000) and within the sediment (Rysgaard et al., 1998; Glud et al., 2000; Berg et al., 2001) have been made, along with estimates of abundance and production of dominant infauna (Sejr et al., 2000; Sejr et al., 2002) and walrus abundance and predation (Born et al., 1997; Born et al., 2003). In an attempt to establish carbon and nutrient budgets and models for carbon and nutrient cycling in this high arctic fjord (Young Sound), detailed data on bathymetry and hydrographic conditions are needed.
In this study, we surveyed the bottom topography of Young Sound and combined the resulting data with existing topographical data from land to provide hypsometry and catchment data on the fjord. Young Sound is the first fjord in northeast Greenland to be fully surveyed, and our investigation covers climatic data, water discharge from the catchment area, and sea-ice conditions during 2000–2001, along with hydrographic conditions during the short summer thaw. We used a hydrodynamic model to evaluate the barotrophic signals entering Young Sound from the adjacent sea. The model is based on the topography, freshwater input, and wind conditions as input variables and is used to estimate water transport in the fjord. Furthermore, measurements of nutrients and total organic carbon in the water column are used in a volume-mass model to estimate a carbon budget for the outer fjord area. Finally, simulations of future changes in climatic conditions using a regional climate model are presented and discussed in relation to expected changes in the ice-free period, water discharge, and hydrographic conditions in Young Sound.
This investigation was carried out in the fjord Young Sound (74°18′N, 20°18′W) situated in the Northeast Greenland National Park (Fig. 1). Long-term climatic records (1958–1999) are available from Daneborg, situated in the outer part of the fjord (Cappelen et al. 2001), and from Zackenberg Ecological Research Operations (1995–present) located farther inside the fjord (Meltofte and Thing, 1996, 1997; Meltofte and Rasch, 1998; Rasch, 1999). The duration of the midnight sun and polar night periods in the area is 101 days and 81 days, respectively, with a maximum incoming radiation of ∼800 W m−2 during summer (Fig. 2a). Winters are generally very cold, with mean monthly temperatures as low as −22°C (Fig. 2b). Mean temperatures (1961–1999) are below freezing 9 months of the year, and only the months June to August have positive mean air temperatures of up to 4.0°C. During summer the outer area of Young Sound is often affected by fog from the ice-filled sea, the temperature of the fog being only a little above 0°C. The mean annual wind velocity (1958–1999) in the area is 4.9 m s−1, most frequently coming from the north (Fig. 2c). However, during June–August (1958–1999) the dominant wind direction is easterly, with a mean wind velocity of 4.1 m s−1. Annual precipitation varies from 214 to 320 mm, approximately 75% of which is snow (Ohmura and Reeh, 1991; Soegaard et al., 2001). Snow cover is usually formed in September and disappears through June–July. Fast ice in the fjord breaks in July, and new ice begins to form in September, growing to a thickness of 10–15 cm in early October (Fig. 2d).
The outer part of the fjord (Glud et al., 2000) and an area near Zackenberg (Nielsen and Rasch, 1997) have been mapped earlier using a relatively coarse net of transects. In this study, we improved data cover in these areas as well as surveying the bottom topography of the remaining part of Young Sound. In particular, the outer part of the sound around the sill was mapped in great detail. The survey was performed from a rubber dinghy using a dual-frequency echo sounder and GPS receiver. Data were recorded continuously along predefined transects. The surveyed transects totaled 850 km, and a total of 200000 data points was collected. These data were integrated with the topographical data from the Geological Survey of Denmark and Greenland (GEUS) and a Landsat TM satellite image to construct a map of the entire area. This combined data set was used to estimate the catchment area of Young Sound.
The freshwater discharge from the Zackenberg River has been available since 1995 from automatic water-level recordings and frequent discharge measurements (Rasch et al., 2000).
Continuous measurements of water level, salinity, temperature, current velocity, and direction were made at water depths of 5 and 25 m at 2 positions using Aanderaa RCM 9 current meters (Aanderaa instruments, Bergen, Norway). One station was placed at the mouth of the fjord near the small island of Sandøen (74°14.995′N, 20°11.111′W), and the other about 30 km upstream near Zackenberg (74°26.450′N, 20°27.157′W). These stations will be referred to as the Sandøen station and the Zackenberg station, respectively (Fig. 1b). We made additional measurements using an Aanderaa DCM 12 current profiler near Sandøen (Aanderaa instruments, Bergen, Norway). This instrument, an Acoustic Duppler Current Profiler (ADCP), continuously measures the current velocity and direction at 6 water depths as well as the water levels.
Daily measurements of water temperature and salinity profiles (0–100 m) were made using a Conductivity Temperature Detector (CTD) profiler (SIS CTD + 100) together with daily samplings of water (1, 5, 10, 15, 20, 30, and 50 m) from 2 stations: the Sandøen station and the Basaltø station (74°19.336′N, 20°19.828′W) situated near the island of Basaltø (Fig. 1). During a tidal period at the end of the field program, hourly CTD profiles were recorded and water samples collected in the same depth intervals at which daily measurements were made. We also conducted length and cross-section measurements of salinity and temperature profiles in the fjord. Salinity was measured using the Practical Salinity Scale, and all sensors were calibrated immediately after the field campaign.
Water samples for determination of NO3− and PO43− were filtered through glass fiber filters (Whatman GF/F) and frozen (−18°C) within 1–3 h after collection in the field. Samples for total organic carbon (TOC) and total nitrogen (TN) determinations were frozen in acid-washed 30-ml vials for later analysis. Concentrations of NO3− + NO2− were determined according to the scheme of Braman and Hendrix (1989). Phosphate and TN concentrations were determined by standard methods as described by Grasshoff et al. (1983). TOC was analyzed with a Shimadzu TOC-5000 Analyser using high-temperature catalytic techniques described in Buesseler et al. (1996). TOC calibrations were made with 45-μl injections of C8H5KO4 in UV-oxidized Q-water. All carbon data were corrected for instrument blank of 12 μM, and each measurement represents the mean of 3 to 7 injections.
Young Sound covers an area of 390 km2 and is ∼90 km long and 2–7 km wide (Fig. 3). The volume of the fjord is 40 km3, and the average depth in the fjord is ∼100 m, the maximum depth of 360 m being found in Tyrolerfjord (Region 4). A 45-m sill is found in the outer part of the fjord and another further inside the fjord (Region 5). Sea-floor area and volume of the different regions (0–6) are given in Tables 1 and 2, respectively.
The topographic map was used to measure the catchment area of Young Sound (Fig. 4). The area was found to be 3109 km2, although the borders of the watershed on the Inland Ice sheet are uncertain. Water discharge from the Zackenberg River peaked at 60 m3 s−1 around 30 June (Fig. 5). The total water discharge in 2001 was 175 × 106 m3, and 65% occurred from 8 June to 20 July. The values for 1997, 1998, and 1999 were 174 × 106, 256 × 106, and 186 × 106 m3, respectively. Assuming that the water discharge in the Zackenberg River, representing a catchment area of 512 km2 (Rasch et al., 2000), is representative of the entire Young Sound catchment area (3109 km2), the total water discharge to the fjord is 1063 × 106 m3 yr−1.
Sea ice in the fjord represents a maximum freshwater input of 540 × 106 m3, in the event that all ice within the fjord should melt. However, the sea ice broke on 27 July 2001 (Fig. 2d), when it was around 20–30 cm thick, and was exported from the fjord to the Greenland Sea within 1 d. This export represents ∼100 × 106 m3 fresh water. Drifting sea ice was absent from Young Sound until 11 August, when dense pack ice suddenly entered Young Sound and drifted within the fjord for 1–2 d before being exported to the Greenland Sea once again. Regular sea ice was established again by early October.
Young Sound is a deep-sill fjord with significant buoyancy input in the form of fresh water from the melting of snow and ice in the catchment area and the melting of sea ice. During the ice-free period, the freshwater input and mixing by wind and tides results in an estuarine circulation, where lighter water of salinity <30 is moved seaward above denser water from the East Greenland Current with a salinity of 31.5–33 (Fig. 6). The primary pycnocline is found at a depth of 5 m, and the surface-layer salinity is in the range of 14–30 (Fig. 6). The freshwater content above the sill depth decreases gradually from 0.07 in the Tyrolerfjord to 0.026 at Basaltøen and 0.014 at Sandøen. The temporal variation at Basaltøen and Sandøen causes the errors of the mean freshwater content to be 0.002 and 0.004, respectively. However, the surface-layer temperature is 2–7°C and reaches 6–7°C in the inner parts of the fjord (Fig. 7). The density difference across the pycnocline has a magnitude of 5–10 kg m−3. At larger depths salinity increases while temperatures are below zero. The subzero temperature zone gradually deepens farther inside the fjord (Fig. 7b). The cool and more saline water from the Greenland Sea submerges below the surface water of Young Sound, and weak tidal mixing decreases its salinity and increases its temperature by entrainment of warmer low-salinity surface water. The horizontal density differences caused by differences in salinity exceed those caused by temperature. Consequently, the baroclinic circulation is dominated by differences in salinity.
A tidal amplitude of 0.8 to 1.5 m was observed at both the Sandøen and Zackenberg sites (Fig. 8a). A hydrodynamic model simulation of tidal amplitude agreed well with the measured data set on both monthly (Fig. 8a) and diurnal (Fig. 8b) scales. Furthermore, the model predicted high current velocities at the entrance to the fjord between Sandøen and Clavering Ø (Fig. 8c) in an area where polynyas are observed regularly (Fig. 8d). Direct measurement shows surface current velocities of up to 120 cm s−1 near Sandøen, with a mean velocity of 45 cm s−1. Calculations of the currents revealed that the mean velocity in the entire cross-section at Sandøen is 17 cm s−1. Inside Young Sound the mean velocity is lower by a factor 10 because the barotrophic flow is distributed over a larger cross-sectional area compared to the area at the entrance of the fjord.
During the study period, the TOC concentration in the water column ranged from approximately 60 to 110 μM at the Sandøen and Basaltø sites (Fig. 9a). Elevated concentrations were found in the upper 10–25 m in association with the pycnocline in the period 2–12 August. TN concentrations ranged from 5–15 μM at the Sandøen station and 8–20 μM at the Basaltø station, with an average TOC:TN ratio of 9 at both stations (data not shown). Nitrate concentrations ranged from 0–4 μM in the upper 50 m of the water column, with values below 0.5 μM in the upper 10–25 m (Fig. 9b). A vertical concentration gradient of PO43− was also observed at the two sites during the study period, with concentrations below 0.3 μM in surface waters and up to 1.5 μM around a depth of 50 m (Fig. 9c).
Young Sound is an East Greenland fjord similar to numerous other large and deep fjords in the region that often penetrate hundreds of km inland (Fig. 3). The fjords are typically U-shaped valleys incised into high mountain plateaus and continuing westward underneath the Greenland Ice Sheet (Sugden, 1974). With lengths of up to 300 km and depths to 1600 m, the East Greenland fjords are among the longest and deepest in the Northern Hemisphere (Funder et al., 1998; Cofaigh et al., 2001). The Young Sound–Tyrolerfjord system is narrow and steep-walled in its interior (Tyrolerfjord) and wider and more shallow toward the mouth (Young Sound) (Fig. 3). This topographical contrast is typical of glacially eroded fjords (Bennet and Glasser, 1996) but has become more prominent in the Young Sound–Tyrolerfjord system due to differences in bedrock geology between the inner part of the fjord, consisting of hard Caledonian gneisses, and the outer part of the fjord, of softer Cretaceous and Tertiary basalts and sandstones (Escher and Watt, 1976). Although data on bathymetry in fjords along the East Greenland coast are sparse (e.g., Cofaigh et al., 2001; this study) it is likely that most of the fjords have one or more sills (Funder, 1989; Funder and Hansen, 1996; Funder et al., 1998). As the presence of sills restricts horizontal water exchange along the longitudinal axis of the fjord, variations in bathymetry and water depth at the sill between fjords will lead to different water-exchange conditions from one fjord to another, as has been shown in Norwegian fjords (Aure and Stigebrandt, 1989).
Typically, meltwater from the Inland Ice and/or rivers drain into the inner parts of the East Greenland fjords. The total freshwater discharge in Young Sound takes place over a 3-mo period during June–September, when air temperatures exceed 0°C, and 65% of the discharge occurs within 6 wk in 2001 (Figs. 2b, 5). The outflow of fresh water and mixing result in a seaward surface slope and also causes the primary pycnocline in the depth range 5–10 m (Fig. 6) to have a landward inclination.
Evaluation of the tidal current and circulation using the 2D hydrodynamic model (MIKE21, DHI, Hørsholm, Denmark) showed that transport of tides was delayed less than 15 min from the mouth of Young Sound to the Zackenberg station due to low friction in the fjord system (Fig. 8). Thus, the energy loss of the barotrophic signal is minimal. However, the tidal forces generate high current velocities (up to 120 cm s−1) at the mouth of Young Sound, causing a delay in establishment of ice cover in the outer sill area (Fig. 8). Openings (polynyas) in the sea ice have regularly been observed to form in this area during winter and may in some cases initiate a larger opening from the Sandøen location and several km inside the fjord (Military Division Sirius, personal communication).
During summer, current velocities have a magnitude of 0.05 m s−1, while the mean wind velocity is 4.9 m s−1. This represents a production of turbulent kinetic energy of 0.1 × 10−3 W m−2 and 5.6 × 10−3 W m−2, respectively. The energy produced by wind is calculated according to the procedure of Rasmussen (1997), while the tidal production is calculated as a bottom shear stress times the average tidal velocity applying a friction coefficient of 10−3. The wind mixing may support an entrainment from below the pycnocline of 6.4 × 10−7 to 1.3 × 10−6 m s−1 at the present surface-layer thickness (5 m) and the density difference across the pycnocline (5–10 kg m−3). However, the quantities suggest that during the ice-free period, wind forcing is the dominant mixing agent of the surface waters, while tidal mixing and internal waves give rise to mixing in the deeper areas of Young Sound. Little change in density was seen in the water mass below a depth of 25–30 m during this investigation. At the entrance to Young Sound, tidal currents are stronger and have a strong influence on the mixing of the surface layer.
In the period of ice cover, wind mixing is absent, while both internal waves and tides mix the water column of Young Sound. Earlier measurements of salinity and temperature in bottom and surface waters during winter (February) have revealed a well-mixed water column (Rysgaard et al., 1999). Thus, total or partial renewal of the water in Young Sound most likely occurs during autumn, when stratification weakens due to cooling of surface waters and reduced freshwater inflow from the surrounding land, as observed in northern Norwegian sill basins (Eilertsen et al., 1981). In addition, the development of sea ice leads to formation of high-density brine that sinks toward the bottom, resulting in buoyancy-driven circulation. It has recently been demonstrated that brine leaking out of sea ice under freezing conditions is supersaturated with respect to O2 (Glud et al., 2002a). Brine drainage may therefore also be responsible for a transfer of oxygenated water to deeper water layers. Thus, Young Sound belong to the “Arctic circulating category” presented in the model by Gilbert (1983) where mixing and permanently oxic conditions to the bottom is typical of fjords in the Canadian Arctic and west Greenland and give rise to rich and diverse benthos communities (Dale et al. 1989).
NUTRIENTS, TOTAL CARBON, AND NITROGEN
Irradiance causes the upper water layers to reach temperatures of 6–7°C (Fig. 7). Irradiance is sufficient to increase the temperature down to depths of 20 m (light attenuation is 0.19 m−1). The irradiance further stimulates primary production to deplete NO3− within this depth range, while the concentration of PO43− is only marginally reduced above the second pycnocline (Fig. 9). Nitrogen fixation has been shown to be of minor importance in the fjord (Rysgaard et al., 1996), and according to the Liebig's classical way of evaluating nutrient limitation through comparison of observed P:N concentrations with the Redfield ratio, 16N:1P, the very low NO3−:PO43− ratio in the upper 20 m observed during the present investigation (Fig. 9b and c) supports earlier conclusions that NO3− limits primary production in the fjord (Rysgaard et al., 1999). The finding that phosphorus was present in excess throughout our investigation is also in accordance with observations from arctic waters in general (e.g., Harrison and Cota, 1991).
The elevated TOC concentrations found in the upper 10–25 m in association with the second pycnocline from 2–12 August (Fig. 9a) reflect increased chlorophyll values due to the onset of primary production following sea-ice breakup, as shown earlier at the same location (Rysgaard et al., 1999). This finding is further supported by the low TOC:TN ratio of 9 observed in the water column. The increase in nutrient concentrations below 10–25 m (Fig. 9b and c) is due to bacterial nutrient regeneration in the water column (Ottosen et al., unpublished data) and sediment (Rysgaard et al., 1998; Glud et al., 2000). In these previous studies from Young Sound it have been shown that nutrients (NO3−, NH4+, urea, PO43−, and Si) were released from the sediment to the overlying water throughout the year due to active mineralization in the sediment.
Earlier investigations from the outer region of Young Sound suggest that mineralization exceeds primary production, and as such the system is net heterotrophic, requiring additional carbon input to balance mineralization (Glud et al., 2000, 2002b). The previous mass-balance study was based on direct measurements of primary production and mineralization in the pelagic and benthic compartments and vertical export and burial within the sediment. In order to make an independent estimate of the net heterotrophic versus net autotrophic status of Young Sound, a net carbon budget for the outer fjord area was established using techniques independent of the previous studies (Fig. 10). The carbon export toward the sea (Nc, mol s−1) was estimated through two separate transects (Sandøen and Basaltø—Fig. 3, region 1) as:Knudsen, 1900; Officer, 1980; Rasmussen and Josefson, 2002). The saltwater transport is calculated as: Figs. 2, 5). Equations 1 and 2 are used to calculate the transports through the Sandøen and Basaltøen cross-sections (Fig. 3, Region 1). Thus, the transport in the Sandøen cross-section becomes: Table 3. The freshwater runoff from the catchment area of Young Sound is estimated at 75 m3 s−1 during the investigation period, using discharge data from the Zackenberg River (Fig. 5) and adjusting for catchment area (Fig. 4). At the Sandøen cross-section, the freshwater-driven TOC export from the fjord (Cs,S R) is 5.4 mol C s−1 is lower than the TOC import from the Greenland Sea arising from the concentration difference and the saltwater exchange (((Cs,S−Cb,S)/fS)R = −39.3 mol C s−1). Note that the negative value represents an import into the fjord. Consequently, the TOC flow is directed into the fjord at the Sandøen cross-section (Fig. 10). At the Basaltø cross-section, we find that the freshwater-driven TOC transport (Cs,B R = 6.0 mol C s−1) into Region 1 is lower than the flux out of Region 1 caused by the saltwater exchange (((Cs,B−Cb,B)/fB)R = −13.1 mol C s−1) and, hence, the TOC transport is directed away from Region 1 and further into the fjord (Fig. 10).
Bringing the calculation one step further toward the net budget, Equation 5 enhances the dependency of the concentration differences and estimated parameters (fB and fS). The net flux of TOC to Region 1 related to the freshwater volume flow ((Cs,S−Cs,B)R) is −0.6 mol C s−1 which corresponds to a net retention of 0.6 mol C s−1 in Region 1 (Fig. 10, Table 1). Furthermore, the net flux of TOC into Region 1 from saltwater exchange ((Cs,S−Cb,S)/fS − (Cs,B−Cb,B)/fB)R is −26.2 mol C s−1 corresponds to a net retention of 26.2 mol C s−1 in Region 1. In total, the net retention becomes 26.8 mol C s−1, corresponding to 27.8 t C d−1 for Region 1 (Fig. 10).
This mass and volume budget supports previous findings of a net consumption of carbon in the outer part of Young Sound (Glud et al., 2000). However, it is difficult to assess the accuracy of this net carbon consumption estimate (–26.8 mol C s−1) because of the uncertainty associated with the freshwater supply to the fjord. In the calculations of the total freshwater input to the fjord from the surrounding land, we assumed that the Zackenberg catchment area of 512 km2 is representative of the entire catchment area of 3109 km2 (Fig. 4). This estimate, though crude, is the best estimate at present because of the limited data available from other areas than the Zackenberg catchment area. If the freshwater input is subject to an uncertainty of 50%, the net carbon consumption (–26.8 mol C s−1) will be in the range of 15–50 mol C s−1, corresponding to 15–50 t C d−1 to Region 1. More work is needed in future to obtain a better estimate of the entire freshwater supply to Young Sound.
The HIRHAM regional model (Christensen and Kuhry, 2000; Christensen et al., 1998; Dethloff et al., 1996) was used to predict changes in wind, temperature, and precipitation minus evaporation (P−E) conditions in East Greenland. The model has a 50-km horizontal resolution and is nested within 2 transient climate-change simulations (Roeckner et al., 1996; Stendel et al., 2000). The ability of the model to simulate realistic present-day arctic conditions has been thoroughly documented (Christensen and Kuhry, 2000; Dethloff et al., 1996). The model predicts small (<5%) changes in the average 30-yr wind conditions in the northeast Greenland region at the end of this century (2071–2100), as compared to present-day (1961–1990) values. In contrast, it predicts a dramatic increase in the average 30-yr atmospheric temperatures of up to 8°C in the region by the end of this century (Fig. 11). Furthermore, the average 30-yr P−E conditions in the region are predicted to increase 20–30% in the same period (Fig. 11). The effect of increased temperatures and P−E can be expected to cause dramatic changes in future sea-ice conditions in Young Sound. Reliable predictions of sea ice changes in Young Sound cannot be made with the HIRHAM model due to insufficient horizontal resolution, but may be made by another approach. Today, sea ice covers the fjord for 9–10 mo of the year and grows to a thickness of ∼1.5 m (Fig. 2). Because the water mass underlying the sea ice in Young Sound consists of polar water, the oceanic heat flux is negligible, and the only factors affecting sea-ice thickness inside the fjord are air temperature and snow cover. Thus, the sea-ice thickness and seasonal length can be estimated from the number of degree-days of freezing since the beginning of winter (Lebedev, 1938; Bilello, 1961), with a further treatment for melt rates in summer (Bilello, 1980). Through this technique and applying the air-temperature rise of 9.3°C ± 1.5 during December–February; 4.7°C ± 1.3 during March–May; 0.4°C ± 0.3 during June–August, and 8.6°C ± 2.1 during September–November predicted by HIRHAM (scenario B2) in 2071–2100, the winter fast-ice thickness in Young Sound can be predicted to decline approximately from 1.4–1.9 (depending on snow thickness) to 0.8–1.3 m, with the expected increase in the ice-free season from 2.5 mo today to 4.7 mo at the end of the century. The scenario A2 data in the estimation further decreases sea-ice thickness and increases the ice-free season to 5.3 mo.
An increased ice-free period will enhance the heat flux from the atmosphere to the surface water and increase the freshwater runoff from land to the fjord. Evaluation of the effect of increased freshwater flux into the fjord according to Officer (1976) reveals that in fjords dominated by entrainment mixing, the surface-layer thickness (e.g., upper low-salinity layer) will change only marginally, whereas the transport of salt water from the Greenland Sea to Young Sound below the halocline will increase significantly due to increased estuarine circulation. (Fig. 12). An increase in the ice-free period will enhance biological productivity in the area due to increased light availability for primary producers (Rysgaard et al., 1999; Kühl et al., 2001; Roberts et al., 2002; Borum et al., 2002; Glud et al., 2002b). Because the surface layer thickness will change only marginally by the end of the century, the phytoplankton bloom will continue to occur in a subsurface layer, but as net transport increases, production will benefit from an increased import of nutrients and organic matter from the Greenland Sea. Improved food availability will stimulate bivalve growth and production in the area (Sejr et al., submitted). Furthermore, an increase in the ice-free season will prolong the period in which birds and marine mammals, e.g., walruses, have access to the food-rich coastal area and thus improve their foraging conditions (Born et al., 2003).
The predicted changes in temperature, ice-free conditions, and precipitation in the area at the end of this century suggest that physical conditions in Young Sound will become more similar to present-day conditions farther south, e.g., at Kap Tobin (Fig. 13). Thus, the area extending from Young Sound and a few hundred km south represents a climate gradient reflecting this century's climate change. We suggest that north-south transects in this region may be highly valuable in evaluating adaptations in biological processes and species to different physical settings in relation to future climate change in the Arctic.
The study was supported financially by DANCEA (Danish Cooperation for Environment in the Arctic, Danish Ministry of Environment). We acknowledge the Danish Polar Center for logistical support. The Danish Military Division, Sirius, is thanked for its hospitality. In addition, we wish to thank Egon R Frandsen, Jan Damgaard, Kitte Gerlich, Marlene Skjærbæk, and Tanja Quottrup for technical assistance in the field and laboratory. We thank Ronnie N Glud, Michael Sejr, and Morten Rasch for helpful discussions. Robert Gilbert is thanked for his comments on the manuscript. Finally, we are grateful to Anna Haxen for linguistic corrections.