In July-September 1997 two hydrographic lines were done in the western N. Atlantic along longitudes of 52 and 66^o W as part of the WOCE onetime hydrographic survey of the oceans (Fig 1). Each of these two lines approximately repeated earlier ones done during the International Geophysical Year(s) (IGY) and the mid-1980s. Because of this repeated sampling, long-term hydrographic changes in the water masses can be examined. In this report, we focus on temperature and salinity changes within the subtropical gyre mainly between latitudes of 20 & 35^o N and compare our results to those presented by Bryden et al (1996) who examined changes along a zonal line at 24^oN, most recently occupied in 1992. Since this most recent 24^o N section in 1992, substantial changes have occurred in the western part of the subtropical gyre at the depths of the Labrador Sea Water (LSW). In particular, we see clear evidence for colder, fresher Labrador Sea Water throughout the gyre on our two recent sections that was not yet present in 1992 at similar longitudes along 24^o N. At shallower depths inhabited by waters which are an admixture of Mediterranean (MW) and Antarctic Intermediate Waters (AAIW), our recent survey shows an increase in salinity, which can only be attributed to changes in water masses on potential temperature or neutral density surfaces. Furthermore, waters above the MW/AAIW layer and into the deeper part of the main pycnocline have continued to become saltier and warmer throughout the 40 year period spanned by our sections. These latter changes have been dominantly due to a vertical sinking of density surfaces as T/S changes in density surfaces are small, but depths of individual T/S horizons have increased with time. The net change since the IGY shows a mean temperature increase between 800 & 2500m depth at a rate of 0.57^o C /century with a corresponding steric sea level rise of 1 mm/yr, and a net downward heave with small values near the top and bottom, and a maximum rate of –2.7 m/yr at 1800m depth. Changes in the deep Caribbean indicate a warming since the IGY due to temperature increases of the inflowing source waters in the subtropical gyre at 1800m depth, but no significant change in the deep salinity.

Changes at 52^o W (A20)

The sections of potential temperature, salinity and neutral density at 52^o W (Fig2) are contoured with dashed contours in the upper panel for q=1.5, 2.5 & 3.5^oC, middle panel for S=34.85, 34.95 & lower panel for gn=27.9, 27.95 kg/m³. Changes with time have been estimated by vertically interpolating the bottle data from IGY and as well as the CTD data at standard levels for the two modern cruises before horizontal gridding. This procedure assures comparable errors due to curvature in the interpolation of all three data set. Horizontal gridding is onto a 0.5 degree latitude grid where longitudinal differences in the sections are ignored. Resulting differences (Fig3a, Fig3b) have been smoothed using a 100 km gaussian filter. We show potential temperature differences for 80s-IGY (upper), WOCE-80s (middle) and WOCE-IGY (lower). Positive (negative) differences are in red (blue) with the contour interval of 0.1 up to a maximum (minimum) of 0.5 (-0.5) (a). As above but for salinity differences with the contour interval of 0.02 up to a maximum (minimum) of 0.1 (-0.1) (b). The warming of mid-depth (1000-2500m) waters from the IGY to the 80s has disappeared at depths of 2000m (near the core of the LSW) comparing the 80s to WOCE, although recent freshening of the salinity at this depth has occurred. Net changes from IGY to WOCE still show a basin-wide increase in temperature at this depth range, however. We have averaged the properties between the latitudes of 20 & 35^o N in order to reduce eddy variability and to focus on mid-latitude changes away from boundary influences. On the A20 section (52W), this eliminates the strong latitudinal gradients associated with AAIW to the south and masks out the coldest AABW and the core of the DWBC. Mean temperature (Fig4a) and salinity (Fig4b) differences for the section include an error in the mean based on the observed variability and an eddy length scale of 300km. One can see large, offsetting changes in the upper 1000m in the time interval IGY-80s and 80s-WOCE, largely due to the vertical heaving of the main pycnocline. Upper ocean changes are also masked by the eddy variability. Between 1000 and 2000m temperature (but not salinity) changes have been of the same sign and appear re-enforced in the net change The spatially-averaged q/S changes (Fig5) are shown in the upper left panels of the figure and selectively focus on the thermocline (upper right), MW/salinity maximum (lower right) and LSW (lower left). One change not obvious from the previous figures is the salinity increase in the upper thermocline between the first two occupations and the present. Symbols denote changes for depths of 200, 500, 1000, 1500, 2000, 3000 and 4500m. At a depth of 200 (first symbol on figure) this is evident as a shift of the q/S diagram to higher salinities while at 500m depth (second symbol), it is more clearly a change along the mean q/S diagram. The latter type of variability characterizes much of the change throughout the thermocline between depths of 500 & 1000m.

Changes at 66^o W (A22)

A similar analysis has been done for the 66W section, where we have interpolated two bounding IGY sections (Fig1) in order to estimate an iitial 'section' at 66W for the IGY. As above, we show the overall sections (Fig6), changes with time (Fig7a, Fig7b), latitudinal averages from 20-35N for the changes against depth (Fig8a, Fig8b) and in q/S space (Fig9). We basically see comparable changes at the two longitudes but the recent LSW changes are greater at 66W, which can be seen by comparing the salinity differences in the lower left panels of figs 5 and 9. For both sections, the q/S curves are coincident for the IGY and 80s near q = 3.1^o C, but at the core level of the LSW (note the symbols for 2000m depth in both), one notices a difference in character: from IGY-80s, the change indicates a q/S shift along the mean curve for A20 but more of a combined contribution due to vertical sinking of properties as well as a salinity increase for A22. Between the 80s and the present, temperature has changed little, but salinity has decreased on both sections.

Summary

Space does not permit a more complette presentation of the observed changes including those in the Caribbean and the effects of vertical heave and water mass changes on neutral surfaces. However, a manuscript has been prepared and submitted to Deep-Sea Research by the authors, and a more thorough account will become available (eventually). We summarize our results by combining the net changes (IGY-WOCE) for both sections into a grand average for the subtropical gyre (Fig10). Spanning a time interval of about 43 years (WOCE-IGY), we see a maximum temperature increase of 0.6^o C, which is nearly 1.4^o C/ century. Over the depth range where a significant temperature change has occurred, the net change from IGY to present is 0.25^o C (A20) and 0.24^o C (A22), which works out to an increasing temperature of 0.57^o C per century over a depth interval of 1700m. The net steric sea level rise can be computed from the combined contributions due to temperature and salinity. Changes in the latter will act to reduce the net sea level increase, but the overall steric increase, accounting for changes between 800 & 2500m depth, is 4.7 & 4.3 cm for A20 & A22, respectively, which is equivalent to about 1 mm/yr. These figures for sea level and mean temperature change are only slightly greater than those estimated from Bermuda by Joyce and Robbins ( 1996, 0.5^o C per century & 0.7-0.9 mm/yr) but apply to a thicker water column and point to a long-term increase in the stratification between mid-depths and the underlying deep waters. The mid-depth increase in temperature and salinity is dominantly due to heaving. The depth variation of the 'heave' signal (not shown) indicates a maximum negative shift of 112m at a depth of 1800m giving a downward vertical velocity of the density surface at 1800m depth of –2.7 m/yr. We wish to acknowledge the support of an NSF grant (OCE95-29607), the assistance of J. Dunworth-Baker and R. Goldsmith in helping with some of the calculations and finally, the WHOI CTD and hydro groups and the captain & crew of the R/V Knorr for their assistance in obtaining this new WOCE dataset.

References