Variations of Winter Water properties and sea ice along the Greenwich meridian on decadal time scales

Abstract

Sea–ice ocean interaction processes are of significant influence on the water mass formation in the Weddell gyre. On the basis of data obtained between 1984 and 2008 from eight repeat hydrographic sections, moored instruments and profiling floats in the Weddell gyre on the Greenwich meridian – almost all of them collected with R.V. Polarstern – we identified variations in the properties of the Winter Water and the sea ice draft. In the Winter Water the salinity was relatively low throughout the 1990s (with a minimum in 1992) and a maximum was observed in 2003. Observations of sea ice draft by moored upward looking sonars are available from 1996 onwards. In the southern part of the transect they display variations on a decadal time scale with a minimum in sea-ice thickness in 1998 and an increase since then. Salinity variations in the Winter Water layer cannot be explained only by variations in sea-ice formation and variable entrainment of underlying Warm Deep Water, but lateral advection of water and sea ice needs to be taken into account as well. Potential sources are melt water from the ice shelves in the western Weddell Sea or transport of water of low salinity entering the Weddell gyre from the east. Accompanying variations of the properties of Warm Deep Water are discussed in detail in a companion paper (Fahrbach et al., 2011).

Introduction

The Weddell Sea is one of the main regions of ventilated water production in the Southern Ocean (Orsi et al., 1999). These waters are transferred into the Antarctic circumpolar water belt and from there they spread equatorwards as part of the lower limb of the global thermohaline circulation (Heywood and Stevens, 2007). This renders the processes occurring in the Weddell Sea some relevance to the global climate.

Dense bottom and deep waters are generated over the shelves of the Weddell Sea (Foster et al., 1987, Foldvik et al., 2004, Nicholls et al., 2009), where salt rejection following sea ice formation leads to the dense surface water component of the nascent deep and bottom waters. The other main mixing ingredient is the Circumpolar Deep Water (locally known as Warm Deep Water), which has been transferred into the Weddell Sea from the Antarctic Circumpolar Current to the north. Apart from this saline mode of ventilated deepwater production, a thermal mode exists which is associated with open ocean deep convection (Gordon, 1991). During the last three decades the saline mode was active; only in the mid-1970s the thermal mode is known to have been dominant, the large Weddell polynya being its expression (Gordon, 1978). Smaller open ocean polynyas occur in the Weddell gyre as transients (Comiso and Gordon, 1996), and they testify of smaller convection events.

The very name “thermal mode” strongly emphasizes the role of temperature as a main factor during deepwater formation. However, the salinity level of the surface layer is an essential co-factor for the overturning potential. The thermal mode can only become active if the surface layer contains sufficient salt. In the Weddell gyre the seasonally varying formation and melting of sea ice is a main factor of variation of surface layer salinity. However, upwelling and entrainment of salt-rich deep water plays a significant role as well (Gordon and Huber, 1990). Variability of one or the other may decisively condition the salinity of the surface layer, which in turn may destabilize the stratification. It should be realized that the local stratification is relatively weak and thus surface layer processes could easily induce convective overturning.

Evidence is accumulating that certain modes of climate variability in the Southern Ocean, of which the Southern Annular Mode (SAM) (Hall and Visbeck, 2002, Thompson and Solomon, 2002) is the most important, are affecting the Weddell gyre. Kerr et al. (2009) suggested a correlation between the SAM and Weddell Sea Bottom Water. Gordon et al. (2007) even suggested a relationship between the changing Southern Annular Mode and the occurrence of the Weddell polynya. Thus, the salinity trends should also be considered in the light of ocean-wide mechanisms. Long-term trends in surface water salinity, resulting in a decreasing salinity of the bottom water in the Indian sector of the Southern Ocean, have been observed (Rintoul, 2007), which was also found by Aoki et al. (2005). Boyer et al. (2005) calculated, based on quality controlled data, significant trends of salinity in all ocean basins, with freshening in the Weddell and Ross Seas. Hellmer et al. (2009) show that surface water salinity anomalies move through the Weddell gyre from the east to the west, emphasizing the role of advection in the salinity issue. Similarly, Park et al. (1998) described the advection of freshwater from the Weddell basin to the Enderby Basin to the east. Jacobs et al. (2002) reported the decrease of salinity in the shelf waters of the Ross Sea, which causes the bottom waters originating from that region to become fresher. Freshening of the northwestern Weddell Sea continental shelf was detected by Hellmer et al. (2010).

On average over the Southern Ocean, the salinity of the surface layer exhibits a strong seasonal cycle with an increase from March to October and a decrease from November to February (Dong et al., 2009). To determine the salinity budget one has to include the major terms advection, diffusion, upwelling/entrainment, and freshwater fluxes. According to Dong et al. (2009), advection and entrainment are the major factors, while freshwater fluxes contribute much less. Locally this may be different, though.

Here we present data of winter surface layer salinity which span a time period of more than 20 years based on traditional hydrographic (Fahrbach et al., 2007) and recent float data. For the purpose of comparing all those data collected in different seasons, the actually observed salinities had to be adjusted to obtain apparent winter values. In addition, ice draft data from moored upward looking sonar instruments are presented for investigating the interactions between the water and the sea ice. These data may give clues as to the probability of a switch from the current saline mode to the thermal mode and thus to the possible formation of a large, non-transient polynya. Since the time lag from section to section varies from one to eight years a quantitative analysis is not feasible. Still, consistent variations on multiannual to decadal time scales suggest correlations between relevant processes which help to understand the functioning of the Weddell system.