Understanding the thermal evolution of earth

Studying the thermal history of Earth's mantle can provide a better understanding of Earth's evolution on a planetary scale. In this work, several mechanisms affecting the thermal evolution of Earth's mantle are investigated. The Nusselt-Rayleigh power law relationship (Nu(Ra)) was calculated from the results of a series of models with three dimensional spherical geometry and free slip boundary conditions. Basally and internally heated convection was examined. For Nu(Ra) = aRaP, (5 was found to be 0.294 0.004 for basally heated systems and 0.337 0.009 for internally heated systems. Model cases were extended to Rayleigh numbers higher than any previous study (109). 0 was not observed to reduce at high Rayleigh number, therefore, as this mechanism cannot be invoked to moderate thermal flux in the past, the influence of time dependent layering on thermal evolution was considered. A parameter space exploration of Rayleigh number and 660 km phase change Clapeyron slope demonstrates that present day Earth could have a partially layered mantle and that full two layer convection is possible in the past at higher Rayleigh numbers. Evolution of mantle temperature was modelled, with the models cooling from an initially layered state. As layering breaks down at high Rayleigh numbers, the mantle passes through a wide domain of partial layering before achieving whole mantle convection. The partially layered regime is characterised by a series of avalanches from the upper into the lower mantle. When an avalanche reaches the core mantle boundary it triggers a pulse of plume-like instabilities in the opposing hemisphere, producing a pulse in global surface heat flux. As the mantle cools, the avalanche-pulse events evolve towards higher frequency and lower magnitude. If this mechanism occurs within Earth, the gradualist view of Earth's thermal evolution may need to yield to a more event-driven model. The mechanics of avalanche-pulse events could also provide an explanation for geochemical observations of periodic maxima in melt extraction from the mantle. The modelling of Earth's mantle produces large data volumes. A distributed computing solution to the data storage problem was investigated. The system, MantleStor, is based on Peer-to-Peer technology and intended to operate over hundreds of standard workstations. A trial implementation demonstrates that MantleStor is able to safely store data in a challenging network environment. Data integrity was maintained with over 30% loss of storage machines. MantleStor is an example of an e-Science project, a discussion of e-Science and its implications is presented