Dynamic Constraints On Thermal Structure Of Subduction Zones

During subduction processes, slabs continuously have heat exchange with the ambient mantle, including both conduction and advection effects. The evolution of slab thermal structure is one of the dominant factors in controlling physical and chemical property changes in subduction zones. It also affects our understanding of many key geological processes, such as mineral dehydration, rock partial melting, arc volcanism, and seismic activities in subduction zones. There are mainly two ways for studying thermal structure of subduction zones with geodynamic models:analytical model and numerical model. Analytical model provides insights into the most dominant controlling physical parameters on the thermal structure, such as slab age. velocity and dip angle, shear stress and thermal conductivity, etc. Numerical model can further deal with more complicated environments, such as viscosity change in the mantle wedge, coupling process between slabs and the ambient mantle, and incorporation of petrology and mineralogy. When applying geodynamic modeling results to specific subduction zones on the Earth, there are many factors which may complicate the process, therefore it is difficult to precisely constrain the thermal structure of subduction zones. With the development of new quantitative methods in geophysics and geochemistry, we may obtain more observational constraints for thermal structure of subduction zones, thus providing more reasonable explanations for geological processes related to subduction zones.Slab temperature and slab heat flux in the deep mantle play a key role in the thermal evolutionary history of the Earth’s mantle. Although the thermal structure of slabs at the shallow depths (< 300 km) has been intensively studied, little work has been done to investigate slab thermal structure in the deep mantle with adiabatic and viscous heating effects considered. Here, we use a 2-D spherically axisymmetric mantle convection model to study slab temperature and heat flux variations at different depths. We explore the contributions of different physical mechanism which causes these variations. The results can be summarized as follows. (1) Slab temperature increases faster with depth than its adiabatic profile. The slab temperature below 300 km depth can be described as an exponential function which is similar to the expression of a mantle adiabat. (2) The variations of slab temperature and slab heat flux are mainly caused by the effects of advection, diffusion, and adiabatic heating. The effects from viscous heating and radiogenic heating are negligibly small. (3) Slab deficit temperature at the 2590 km depths can be constrained as～10% of the core-mantle boundary temperature. Our results are not sensitive to different Rayleigh numbers, internal heating rates, activation energy, and depth-dependent thermal expansivity and diffusivity.Understanding the change pattern of the thermal structure during transition from oceanic subduction to continental subducion is a key in understanding such transition process and plate tectonics. By collecting the previsous P-T(pressure-temperature) data of HP-UHP (high pressure-ultra high pressure) metamorphic rocks, we found thress distinct ocean-continent transition zone:the Tongbai-Hong’an orgoen that changing from hot to cold, the Qilian-Qaidam orogen that changing from cold to hot and the western Alps orogen that stable. We perform convection models under 2-D Cartesian coordinates to estimate different parameters that may affect the thermal structure during the transition processes. We find the porssible reasons that may cool the ocean-continent transition zone:(1) The initial thermal structure of the subducted continental plate is colder than the subducted oceanic plate. (2) The slab dip decreases. (3) The subducting velocity increases. (4) The surface heat flow of the upper plate decrases, and the vice versa. For different ocean-continent transition zones, the reasons that may cause changes of the thermal structure is quite differen, and more observations are needed to provide better constraints.