Because the attenuation of seismic waves is sensitive to variations in temperature and to partial melting, mapping seismic Q in the upper mantle is an important tool for assessing the processes responsible for lateral heterogeneity. Utilizing a digital data base of over 150 seismograms and a spectral ratio technique, we have measured SS-S differential attenuation in the North Atlantic region. Differential attenuation is positively correlated with SS-S travel time residual, and both differential attenuation and travel time residual decrease with increasing seafloor age. In an earlier paper we showed that the age dependence of SS-S travel time residual can be explained entirely by the cooling of the oceanic lithosphere, i.e., contributions from the asthenosphere or from a mantle melt fraction are not required. The assumption that plate cooling also dominates the variation of differential attenuation with age permits the derivation of an empirical relation between Q-1 and temperature for the oceanic thermal lithosphere, i.e., the depth interval 0--125 km. However, both the absolute values and the depth distribution of Q we obtain under this assumption are at variance with the results of surface wave attenuation studies, and the variation of Q-1 with temperature that we derive is not as strongly dependent on temperature as indicated by some laboratory measurements.

We therefore develop models for Q in which lateral variations include contributions from the asthenospheric low-Q zone as well as from lithospheric cooling. The Q models obtained under this alternative assumption are in good agreement with those obtained from surface wave studies and are therefore preferred over those models with lateral variations confined to the upper 125 km. Systematic long-wavelength (1000--7000 km) variations in differential attenuation, corrected for seafloor age, are evident along the axis of the Mid-Atlantic Ridge. These variations can be qualitatively correlated with long-wavelength variations in SS-S differential travel time residuals and are attributed to along-axis differences in upper mantle temperature. ¿ American Geophysical Union 1992