In this section, we examine the warming of the deep water in the 1950s. Fig. 5 shows all temperature observations from the depth range between 2900m and 3100m between 1952 and 1966 and linear fits of the data up to 1962 including and excluding the ATKA 1962 cruise. The Atka data point lies above the trend of the earlier and later data. Extrapolation excluding the ATKA data predicts a temperature of -1.225 C for 1962, instead of -1.183 C as observed. However, a significant warming trend of 0.005 C per year is present even excluding the ATKA data point. Therefore, we conclude a warming of the deep water of the central Greenland Sea in the late 1950s took place, but it may not have been as pronounced as suggested by the ATKA data.
 |
Fig. 5: All direct temperature observations between 2900m and 3100m depth between 1952 and 1966, and the linear fit of the data between 1952 and 1962 including and excluding the ATKA 1962 cruise. The value of the ATKA 1962 cruise is possibly too high. However, even without the ATKA value, a temperature increase of 0.005 C per year is observed. |
|---|
Since the temperature increase in the lower layer in the 1980s and 1990s coincides with a period of lower deep water formation rates as demonstrated by the transient tracer time series, the temperature increase observed during the late 1950s may also reflect an earlier period of decreased deep water formation. Since we do not have transient tracer observations from the 1950s, we have to examine the available data for their significance for this period. Due to its half-life of 269 years, Ar-39 data indicate the average deep water formation rate of the last several decades preceding their sampling date. The very limited Ar-39-data obtained 1981 by Smethie et al. [1986] (88 % modern) for the deep Greenland Sea suggest that average deep water formation rates were high preceding 1981 (the time of observation). However, since the Ar-39-data average over a large period of time, they do not exclude short periods of low deep water formation rates. Bönisch and Schlosser [1995] point out that the tritium concentrations of the deep water observed during the 1970s require a high deep water formation rate since at least 1965. Therefore, we speculate on the basis of admittedly limited data that deep water formation rates of about 0.5 Sv reflect the normal conditions for the Greenland Sea for the past several decades, but for several years deep water formation rates could have been lower preceding 1965. The period of low deep water formation rates in the 1980s and early 1990s was very likely not the only event of this kind. Deep water formation rates were probably low in the late 1950s as well.