(2)Laboratoire d'Océanographie Dynamique et de Climatologie

Subducted temperature anomalies have been invoked as a possible way for mid-latitudes to alter the climate variability of equatorial regions through the so-called `thermocline bridge', both in the Pacic and Atlantic oceans. To have a signicant impact on the equatorial heat balance, however, temperature anomalies must reach the equatorial regions sufciently undamped. In the oceans, the amplitude of propagating temperature (and salinity) anomalies can be altered both by diabatic (nonconservative) and adiabatic (conservative) effects. The importance of adiabatic alter-ations depends on whether the anomalies are controlled by wave dynamics or by passive advection associated with density compensation. Waves being relatively well understood, this paper seeks to understand the amplitude variations of density compensated temperature and salinity anomalies caused by adiabatic effects, for which no general methodology is available, assuming that these can be computed independently of amplitude variations caused by diabatic effects. The rst main result is to show that the ratio T 0=S 0 of density compensated T and S anomalies may potentially undergo large amplitude variations at the gyre scale, because density compensation requires the equality T 0=S 0 = S= to hold along the mean trajectories, and because in the actual oceans th