The most remarkable feature observed in the intermediate water (200m - 2000m) is the decrease in density between 1981 and 1994 (Fig. 2c). Fig. 6 shows the evolution of the average salinity and temperature values in this layer. Before 1980, the temperatures ranged between roughly -1.05 C and -1.2 C with little variations. Since 1981 the temperature increased continuously, reaching -0.9 C in 1994 (Fig. 6a).
 |
Fig. 6: Depth weighted mean values of (a) theta, (b) salinity from the depth range between 200m and 2000m in the central Greenland Sea versus time. |
|---|
Although vertical salinity variations in this layer can be large, varying from 34.855 at 200m depth to 34.898 at 2000m depth (1991) at the same station, the total salt content indicated by the mean value of the gridded data shows only small variations in time (Fig. 6b). An increase from 34.888 in 1981 to 34.891 in 1989 is followed by a decrease, reaching 34.883 in 1994. Formation of a 25 cm thick layer of sea-ice from water with a salinity of 34.87 (winter surface salinity observed by Schott et al. [1993]) would reset the average salinity from 34.883 to 34.888. Since the spatial salinity variations are high (up to 0.043) compared to the observed temporal changes (0.005 in 13 years), these temporal variations may not be significant.
The total heat content of the intermediate water (calculated relatively to the pressure and salinity dependent freezing point) increased from 6.93E4 MJm in 1981 to 7.17E4 MJm, an increase of 0.24E4 MJm, which represents an average net interannual heating rate of 5.3W/m2 (Fig. 7). Cooling rates during deep convection are about 150W/m2 to 200W/m2 (Visbeck et al. [1995]) and persist usually for about two months each year. Therefore, the total heat loss per year typically would not exceed 0.12E4 MJm. This implies that it would take about 2 years of cooling and deep convection to remove the heat accumulated over 13 years.
 |
Fig. 7: Total heat content calculated relatively to the salinity and depth dependent freezing point of the water column between 200m and 2000m in the central Greenland Sea versus time. |
|---|
Deep water formation rates decreased in 1980 (see above). NSDW, however, is formed as a mixture of water from the deep Greenland Sea and EBDW (Swift and Koltermann [1988]). If the production of new deep water is not high enough to balance the implied loss of deep water required for continued NSDW formation, the most dense (sigma2 > 37.45) faction of the intermediate water may sink below 2000m and enter or mix with the deep layer. Since intermediate waters have higher transient tracer concentrations than the deep water (e.g., about 1.2 pmol/kg CFC11 of 800m depth and about 0.8 pmol/kg in 2700m depth in 1982), the roughly constant transient tracer concentrations in the deep water observed during the 1980s and 1990s (Figs. 3 and 4) could also be explained as a mixing product of the most dense intermediate water and the deep waters from the Norwegian Sea and the Arctic Ocean (see Fig. 4). Such a scenario would be consistent with the observed density decrease in the intermediate water (Fig. 2c).
Fig. 8a shows the deepening of the 37.33, 37.43, and 37.45 isopycnals over time, Figs. 8b, c, and d the development of the CFC11 concentrations, the decay-corrected tritium concentrations and the tritium/He-3-age on these isopycnals, respectively. The uncertainties in these quantities are relatively high (the transient tracer values had to be interpolated onto isopycnals). On the highest density isopycnal (37.45), the decay-corrected tritium and CFC11 concentrations remained more or less constant over time, whereas the tritium/He-3-age increased from 7 years in 1981 to 20 years in 1994, i.e., at a 1:1 ratio with the calendar year. This indicates that on this isopycnal was either not mixing with older or younger water, or the mixing was exactly balanced; a behaviour very similar to that observed in the deepest water. On the 37.43 isopycnal, the transient tracer concentrations decreased and the tritium/He-3-age increased, suggesting lateral advection of older water or vertical mixing with underlying water. Constant decay-corrected tritium concentrations and increasing CFC concentrations indicate a strong influence of younger water on the 37.33 isopycnal. In contrast to the CFCs, the tritium concentration on this isopycnal remains constant as a result of the declining surface concentrations.