Possibility of hot anomaly in the sub-slab mantle as an origin of low seismic velocity anomaly under the subducting Pacific plate

Recent seismological studies identified the existence of low seismic velocity anomaly in the sub-slab mantle of the northeast Japan subduction zone around the 410 km depth and argued that it is mainly due to the high temperature anomaly whose magnitude is around 200 K. In this study, to examine this possibility based on the mantle convection theory, two types of models, that is, external and internal origin of hot anomalies, are considered. First, the previous model in which a past hot anomaly was entrained there is reconsidered. We include the previously neglected effects, that is, the effects of the thermal structure inside the subducting/overlying plate and the effects of slab deformation such as a stagnation around the 660 km discontinuity. It is found that the previous model underestimated the effects of the thermal structure inside the slab and the plate, and, as a result, the hot anomaly cools faster than the previous models show. Thus, the previous suggestion that the past Pacific superplume activity may be the cause of the low seismic velocity anomaly observed presently may not be likely. Second, as a model of internal origin of hot anomaly, the model that takes into account the effects of radiogenic heat production, mantle cooling, adiabatic heating, latent heat and viscous heating are considered. It is found that hot anomaly arises in two positions; one is around the 410 km depth and the other is in the lower mantle. The magnitude of both high temperature anomalies is O (200 K) for the earth-like cases, although the former temperature anomaly may be caused by the poor modeling of slab strength. We also found a phenomenon that the high temperature anomaly in the lower mantle occasionally goes up to the upper mantle. In the case that negative Clausius-Clapeyron slope is steeper (-2.8 MPa/K), the amount of hot anomaly going up is smaller and the magnitude of hot anomaly in the lower mantle is higher, compared with the case with gentler slope (-1.3, 0 MPa/K). This phenomenon is similar to the one observed for the interaction between the endothermic phase transition and the cold slab. Considering these results, we propose a combination of above two models as a likely model. In this model, the hot anomaly in the lower mantle produced mainly by the internal, viscous and adiabatic heating goes up occasionally and is then entrained toward the 410 km phase boundary. This hypothesis may be tested by the future studies of sub-slab mantle of various subduction zones. (C) 2010 Elsevier B.V. All rights reserved.