The Kongsfjorden Transect: Seasonal and Inter-annual Variability in Hydrography

Abstract

The Kongsfjorden conductivity, temperature and depth (CTD) Transect has been monitored annually since 1994. It covers the full length of the fjord and the shelf, and the upper part of the shelf slope outside Kongsfjorden. In addition to CTD profiles, data from vessel-mounted Acoustic Doppler Current Profiler (ADCP) and moorings have been collected. Previous studies noted that Atlantic Water (AW) from the West Spitsbergen Current was observed in the fjord every summer, but to a varying extent. The prolonged monitoring provided by the Kongsfjorden Transect data set examined here reveals continuous variations in AW content and vertical distribution in the fjord, both on seasonal and inter-annual timescales. Our focus in this paper is on this variable content of AW in Kongsfjorden, the forcing mechanisms that may govern the inflow of this water mass, and its distribution in the fjord. We classify three winter types linked to three characteristic scenarios for winter formation of water masses. During the historically typical winters of type “Winter Deep”, deep convection, often combined with sea ice formation, produces dense winter water that prevents AW from entering Kongsfjorden. Summer inflow of AW starts when density differences between fjord and shelf water allows for it, and occurs at some intermediate depth. During winters of type “Winter Intermediate”, AW advects into the fjord along the bottom via Kongsfjordrenna. Winter convection in Kongsfjorden will then be limited to intermediate depth, usually producing very cold intermediate water. Deep AW inflow continues during the following summer. A winter of type “Winter Open” seems to develop when open water convection produces very dense shelf water, and AW winter advection into Kongsfjorden occurs at the surface. Summer AW inflow is rather shallow after such winters. We find that variations between Winter Deep and Winter Intermediate winters are due to inherent natural variability. However, the Winter Open winters seem to be a consequence of the general trend of atmospheric and oceanic warming, and, more specifically, of the decreasing sea ice cover in the Arctic region. The Winter Open winters have all occurred after an unusual flooding of AW onto the West Spitsbergen shelf in February 2006.

Appendix A: The Kongsfjorden Transect Hydrography from Individual Years

Here we show temperature, salinity and density distribution for each year with enough available data to construct the Kongsfjorden Transect, separated in Winter Deep, Winter Intermediate and Winter Open winter data (January–May) and their respective following summers (July–September).

The five Winter Deep winters with sufficient CTD data available to grid the transects are shown in Fig. 3.23, although in two of them, 1997 and 2009, the data coverage is poor in the central basin of Kongsfjorden. In 1997, the data coverage is also poor across the shelf-edge front. Focusing on those transects with good data coverage, all these years the WSC had a narrow warm core confined to the shelf-edge region, and reaching the surface. The surface shelf water had similar or lower density than the surface-layer part of the WSC core. The deepest part of the shelf water column was, however, generally denser than the water at the same depth in the WSC core, and the density differences across the front were weak at intermediate depth level.

Fig. 3.23

Mean winter temperature, salinity and potential density along the Kongsfjorden transect, based on all available Kongsfjorden Transect data from winters 1997, 2001, 2002, 2006 and 2009, defined as “Winter Deep”

Such a density distribution favors eddy overturning with AW exchange across the front dominating at the depth level where the density differences vanish, that is intermediate depth (Tverberg and Nøst 2009). This is indeed what can be observed over the shelf. The overturning cell indicated in the principle sketch (Fig. 3.17) of shelf-edge processes, would in such cases apply to the upper part of the water column only. In 2009, AW was found both at intermediate and deep water level, and cross-frontal density differences were weak at both levels. In the transects from the 3 years with good data coverage, it can be seen that Kongsfjorden interior was less influenced by AW than the shelf. We also note that in the mouth region of Kongsfjorden, a depression of the deepest isopycnals is visible each of these years. It may be possible that these depressions are associated with the coastal current at the mouth region of Kongsfjorden forcing the path of AW advection to be modified there (geostrophic control). The exact location of the depressions varies among the years, which may confirm that the location of the geostrophic control is located at the common mouth of Kongsfjorden-Krossfjorden (Kb0) in some years and at the mouth of the Kongsfjorden central basin (Kb1) in others. This is in line with the typical situation during a Winter Deep winter; denser water in the fjord in the deepest part of the water column enhances the bottom speed of the coastal current on the shelf, past the fjord.

Mooring data add valuable information from two of these years; 2006 and 2009. Most of the winter in 2006 was actually a Winter Intermediate winter with exceptionally strong deep AW advection, culminating in a rather short period with strong convection and mixing (see Fig. 3.5), and it must have been during this final part of the winter that the largest volumes of winter water were produced. The winter 2009 was opposite; during most of the winter the mooring data reveal deep convection and only weak indications of AW advection, except from a short period in late April with deep AW inflow (see Fig. 6.2 in Hegseth et al. this volume). That inflow was evidently not strong enough to replace all the winter water produced earlier that winter.

The five Winter Intermediate winters with sufficient CTD data available to grid the transects are shown in Fig. 3.24, although two of them have limited data coverage (1998 and 1999). The data coverage in the WSC, however, is good for all 5 years, revealing a WSC that tends to be isolated from the surface. Shelf water spreads westwards on top of the WSC, and AW from the WSC tends to enter the shelf in the deep water and more pronounced than during the Winter Deep winters. The WSC was clearly denser than shelf water at all depths. This leads to a weakening of the WSC with depth (thermal wind effect), perhaps contributing to enhanced baroclinic instabilities at the shelf-edge front. The resulting eddy overturning will bring AW onto the shelf in the deep and shelf water off-shelf in the surface. The water column in Kongsfjorden interior had lower density than the shelf, meaning no geostrophic control at the entrance. Mooring data confirm deep AW advection inside Kongsfjorden during three of these winters (2004, 2005 and 2010), combined with homogeneously cold water above the AW inflow. In 2010, the AW advection reached rather shallow depths in January–March, while it almost disappeared in April–May and was replaced by a thick layer of homogeneously cold water, reaching almost 200 m depth (see Fig. 6.2 in Hegseth et al., Chap. 6).

Fig. 3.24

Mean winter temperature, salinity and potential density along the Kongsfjorden transect, based on all available Kongsfjorden Transect data from winters 1998, 1999, 2004, 2005, and 2010, defined as “Winter Intermediate”

The three Winter Open winters with sufficient CTD data available to make transects (2007, 2008 and 2014) are shown in Fig. 3.25. The WSC reaches the surface, is less dense than shelf water in the whole water column, and tends to spread onto the shelf, being most pronounced in the surface. The thermal wind effect on the WSC in such a situation will enhance the current speed with increasing depth. This might guide the WSC northwards past Kongsfjordrenna at depth, while the eddy overturning will spread AW onto the shelf in the surface layer. However, except for 2007, the density differences were weak, so topographic steering of geostrophic AW advection in Kongsfjordrenna can be significant, with AW advection in the whole water column. In 2007 it looks like the AW might not be passing Kb1 in the deep part of the water column (due to geostrophic control?). In the surface, however, AW entered the fjord freely. The mooring data inside Kongsfjorden indicated a rather homogeneously warm water column, but with a tendency of warmest temperatures in the surface (see Fig. 6.2 in Hegseth et al., Chap. 6). The homogeneously warm water column is particularly evident in the 2012 and 2014 time series, which may indicate no horizontal density differences across the shelf-edge front that year, meaning weak eddy overturning, but substantial geostrophic AW advection, with horizontal eddy diffusion spreading water masses laterally.

Fig. 3.25

Mean winter temperature, salinity and potential density along the Kongsfjorden transect, based on all available Kongsfjorden Transect data from winters 2007, 2008 and 2014, defined as “Winter Open”

The eight summer transects after Winter Deep winters are shown in Fig. 3.26. Inside Kongsfjorden, they are characterized by remnants of cold winter water in the deep, and a core of warm and saline AW or TAW at some intermediate depth. The shelf has in principle the same distribution, however with more pronounced presence of AW or TAW. Some years (e.g., 2001), the deep fjord water was clearly denser than the deep water in Kongsfjordrenna, which indicates that the geostrophic control in the fjord entrance can be in effect throughout long parts of the summer. The AW exchange was extensive across the shelf-edge front, with no pronounced density front. The mooring data confirm rather cold water in the deep in summer 2003 and 2011, while in 2006 and 2009 the cold water resided below the mooring depth (see Fig. 6.2 in Hegseth et al., Chap. 6).

Fig. 3.26

Mean summer temperature, salinity and potential density along the Kongsfjorden transect, based on all available Kongsfjorden Transect data from summers 1997, 2000, 2001, 2002, 2003, 2006, 2009 and 2011, after “Winter Deep” winters

The six summer transects after Winter Intermediate winters are shown in Fig. 3.27 (in 1995 and 1996, the data coverage was too poor to form transects). They are characterized by remnants of cold winter water at intermediate depth level, while AW or TAW were found in the deep. The deep water can comprise remnants of deep AW inflow during the winter, or summer advection of AW. We would expect the summer transects in 2004 and 2010 to be examples of the first situation, because deep fjord water is denser than shelf water, possibly implying geostrophic control at work in the mouth. Mooring data, however, indicate that there was a distinct increase in AW in 2010, similar to 2005 (see Fig. 6.2 in Hegseth et al., Chap. 6), indicating summer advection of AW. Some years the AW exchange across the shelf-edge front appeared to be restricted (e.g. 1998), but in other years it was pronounced (e.g. 2004, 2005, 2010). The shelf-edge front is not a pronounced density front; rather the isopycnals tend to often be terrain following, which can be a long-term effect of eddy exchange across the shelf-edge front (Adcock and Marshall 2000).

Fig. 3.27

Mean summer temperature, salinity and potential density along the Kongsfjorden transect, based on all available Kongsfjorden Transect data from summers 1994, 1998, 1999, 2004, 2005 and 2010, after “Winter Intermediate” winters

The five summer transects after Winter Open winters are shown in Fig. 3.28. They show that large volumes of old winter water were present in the fjord and on the shelf during these summers (winter water from Winter Open winters is relatively warm). The density at intermediate and deep depth levels were generally higher than in the WSC, with two summers (2007 and 2013) being more pronounced in the fjord than on the shelf. Temperature and salinity were generally higher on the shelf than in the fjord. Mooring data confirm a rather warm water column during these summers and only in 2014 there was a pronounced increase in AW content throughout the summer (see Fig. 6.2 in Hegseth et al., Chap. 6). This indicates that there may be rather little water renewal in the fjord during a summer after a Winter Open winter. Our explanation is stronger geostrophic control at the mouth due to the high-density water inside the fjord.

Fig. 3.28

Mean summer temperature, salinity and potential density along the Kongsfjorden transect, based on all available Kongsfjorden Transect data from summers 2007, 2008, 2012, 2013 and 2014, after “Winter Open” winters

The overall winter mean transects of temperature and salinity (Fig. 3.2) seem dominated by the Winter Deep type of winter, with a water mass close to LW definitions filling most of the shelf and fjord. However, they are strongly influenced by Winter Intermediate and Winter Open winters as well; the shelf-edge front has an overturning leading AW onto the shelf in the deep and shelf water over the WSC (Winter Intermediate), rather high bottom salinity on the shelf and in the fjord (Winter Intermediate), and TAW-type water mass close to Kb3 (Winter Open). Kb3 is the only CTD station in the transect that is placed within the path of the topographic steering of the geostrophic AW advection in the fjord. The influence from the Winter Open (AW) winters will thus likely be strongest there. The overall summer mean transects of temperature and salinity (Fig. 3.3), display a clear discrepancy between the shelf and fjord regions, which in reality is only seen after some Winter Open winters (Fig. 3.28). The water mass distribution inside the fjord seems like a mixture of summers after a Winter Deep and a Winter Open winter. TAW fills most of the water column below thin layers of SW and IW, and LW is only found at the head of the fjord, reflecting the glacier influence in the basin inside the Lovénøyane (see Table 3.5 for water mass definitions). A summer after a Winter Deep winter would have had LW in the deepest part of the water column, while a summer after a Winter Open winter would have pure AW instead of TAW filling part of the water column. Comparing these mean summer sections with the time series of water mass distribution in Kongsfjorden (Fig. 3.20), we see that such a distribution has been the most common after 2006, more specifically the two periods 2006–2008 and 2012–2014.

• • Data Availability: doi:10.21334/npolar.2019.074a215c.