From observations on yearly cruises to the central Greenland Sea betwe
en 1993 and 1996, conclusions are drawn with respect to winter convect
ion and bottom water renewal. The data indicate that winter convection
was extremely weak after 1993, not even ventilating the intermediate
waters. This is remarkable, since the salinities in the upper layers i
ncreased considerably between 1993 and 1995, thus providing presumably
favorable conditions for winter convection. With the absence of deep
reaching winter convective events, the temperatures in the deeper wate
rs of the Greenland Gyre increased steadily by about 0.03 K between 19
93 and 1996. We conclude from the development of mainly the thermal st
ructure on a zonal transect that an explanation for the temperature in
crease can be given by a large-scale downward water movement of about
150 m/yr in the central Greenland Sea. The data indicate that this pro
cess is independent of changes in the dynamically induced density dist
ribution. It is therefore possible that a downward movement, perhaps m
asked by other processes, may continue for many years. If this is the
case, resulting flushing times would be of the order of 20-30 years on
ly. The presence of a large-scale circulation cell with downward movem
ent in the central Greenland Gyre would explain the observed warming o
f the bottom waters without the demand for an actually active heat sou
rce. It is also in accordance with the observed increase of chemical t
racer concentrations in the deep waters.