OBSERVATIONS OF TEMPORAL CHANGES OF TRITIUM-HE-3 AGE IN THE EASTERN NORTH-ATLANTIC THERMOCLINE - EVIDENCE FOR CHANGES IN VENTILATION

A compilation of fifteen years of tritium and He-3 measurements is use d to examine the ventilation of the eastern North Atlantic subtropical gyre with specific emphasis on the temporal character of the tracer a ge field. A multivariate regression analysis in the form of a spatiote mporal Taylor expansion is applied to observations interpolated onto i sopycnal surfaces. The time-dependent component of the tracer age fiel d is found to be statistically significant, explaining approximately 1 0% of the Variance of the tracer age observations in the upper thermoc line (sigma(theta) = 26.5) and increasing to roughly 50% of the varian ce in the lower thermocline (sigma(theta) = 27.0). The observed transi ent tracer age increases over the 15 years of observations with the fr actional rate of change of the age field varying between 2% and 5% per year. The largest observed changes occur on the deepest, most slowly ventilated isopycnal surfaces. The second derivative of the tritium-He -3 age with time suggests that the tracer age field may be approaching a steady state. If tritium-He-3 age is interpreted as a true measure of the advective ventilation age, the temporal changes in age would im ply a slackening of the ventilation of the lower main thermocline of g reater than 50% from the late 1970's to the early 1990's. However, con sideration of the full advective-diffusive balance of tritium-He-3 age reveals that the changes in tracer age field represent a time-depende nt adjustment of the transient tracer concentrations in conjunction wi th a steady local circulation field. integral approximations of the up stream evolution of the tracer field also fail to demonstrate evidence for decadal changes in ventilation. The integral balance along the pa th of subduction yields an improved estimate of the true ventilation a ge based on the temporal tendency of the age field along the path of v entilation. An approximation of this integral suggests that actual ven tilation ages can be up to 40% larger than the measured tracer age in the deeper portions of the North Atlantic thermocline. Proper accounti ng of the time-dependent biases of the tracer age dating technique are a prerequisite for examining transient tracer measurements for eviden ce of changes in the physical ventilation of the upper ocean.