| There are a number of simplifying assumptions in modeling Earth's deep interior. These are mostly simplifying assumptions that make the mathematics simpler either for less complicated modeling or for numerical efficiency purposes. The aim of this study is to investigate the effects of some of these simplifying assumptions on 2D and 3D mantle convection models. In particular, the cases with variable coefficients of thermal expansion, alpha, and the inclusion of mineral phase transitions and viscosity stratification have been studied.;Also an existing version of a 3D parallel mantle convection model, MC3D, from Lowman et al. [2001] have been modified to include the temperature- and depth-dependent alpha. In the 3D study it has also been investigated that how the effects of temperature- and depth-dependent alpha vary with or without lithospheric plates.;There are at least two mineral phase transitions in Earth. There is an exothermic phase boundary at 410km below the surface and an endothermic phase boundary at 660km below the surface. For simplicity, most mantle convection models do not consider any of the phase boundaries. Some consider only the endothermic phase boundary. A 2D cylindrical model from Shahnas and Jarvas [2005] has been employed to investigate the effects of considering both phase boundaries compared to models with either no, or one, phase boundary. Different viscosity stratifications have been used in addition to the phase boundaries.;The coefficient of thermal expansion is temperature- and depth-dependent in Earth. But for simplicity, it has been considered as constant in most mantle convection models and only depth-dependent in others. 2D mantle convection models (2D Cartesian and 2D cylindrical) have been created based on an existing model from Jarvis [1992] to investigate the effects of temperature- and depth-dependent alpha on mantle convection compared with the simplified cases. |
