Abstract
A strong low S velocity anomaly at 300–500 km depth coincides with the western boundary of the Russian Platform. The anomaly is too large to be explained by a simple temperature anomaly or by compositional variations, nor is it an artifact induced by seismic anisotropy. We present a model that explains the anomaly through the injection of water at the time of closure of the Tomquist Ocean that separated the continents of Avalonia and Baltica in the early Paleozoic. Dense hydrous magnesium phases are the most likely agents for transporting water to the transition zone, but an important role may also be played by nominally anhydrous clinopyroxene. When these minerals are brought out of their stability field, water is released. It may accumulate in β‐spinel or K‐amphibole and may be released when the mantle warms up after subduction halts. The effect of this water is to induce weakening of the shear modulus or even creation of a heavy melt. By a conservative estimate, a subduction episode lasting 85 m.y. would inject enough water to saturate a large volume of mantle rock with 0.3% H2O. Low velocities or low Q anomalies have also been observed in the transition zones near currently active slabs.