Overall assessment of the models

The limited interaction mantle models that involve extraction of the atmosphere from the upper mantle do not appear to be viable due to the lack of a simple relationship between atmospheric and MORE Ne and Xe. The models also are opposed to geophysical arguments for whole mantle convection. 

The steady state upper mantle is compatible with much of the presently available noble gas constraints. However, the present formulation, with a mantle boundary at a depth of 670 km, stands contrary to geophysical evidence for substantial mass fluxes across this depth. The mechanism for interactions between mantle reservoirs is promising, however, and might be adapted to other configurations, although not necessarily without time-dependent distributions. 

The idea of a break up of previously layered mantle reconciles the geochemical characteristics that can be explained by long-term evolution in a layered mantle (according to one of the layered mantle models) with geophysical observations regarding recent mantle behavior. However, numerical modeling of the thermal state of the mantle does not support this. 

Models with noble gas reservoirs in other mantle layers. 

The lower boundary layer model avoids some of the geophysical problems of larger convective layers, but must resolve how such a layer is initially stabilized with gas-rich material. Also, it must be combined with other mechanisms to account for the mantle evolution of all the noble gases. However, the model illustrates how a deep boundary may be important. 

---