Thesis (S.M.)--Joint Program in Physical Oceanography (Massachusetts Institute of Technology, Dept. of Earth, Atmospheric, and Planetary Sciences and the Woods Hole Oceanographic Institution), 2000. / Includes bibliographical references (leaves 78-82). / A simple quasi-two dimensional dynamical model of Thermohaline circulation (THC) is developed, assuming that the mixing only occurs near western and eastern boundary layers. When the surface density is prescribed, the climatically important quantities, such as the strength of overturning and meridional heat transport, are related to the zonal integral over the vigorously mixing regions and scaled as (KvΔx)2/3. The numerical results suggest that the density difference between eastern and western boundaries play an important role in the meridional overturning. The eastern boundary is characterized by the upwelling on top of downwelling. The western boundary layer is featured by the universal upwelling. The inefficiency of diffusion heat transport accounts for the narrowness of sinking region and shallowness of overturning cell in one-hemisphere. The experiments with other surface boundary conditions are also explored. The circulation patterns obtained are similar under various surface temperature distributions, suggesting these are very robust features of THC. The role of boundary mixing is further explored in global ocean. The 2 1/2 dynamical model is extended to two-hemisphere ocean. Additional dynamics such as Rayleigh friction and abyssal water properties are taken into account. A set of complicated governing equations are derived and numerically solved to obtain steady state solution. The basic circulation features are revealed in our dynamical model. An equtorially asymmetric meridional circulation is observed due to small perturbation at the surface temperature in the high latitude. The density differences between eastern and western boundaries are distinct in both hemispheres. This is achieved during the spin-up process. Although the dynamical model results agree well with OGCM results in one-hemisphere, several important dynamics are lacking and exposed in two-hemisphere experiments. We need to consider horizontal advection terms which will effectively advect positive density anomalies across the equator and form the deep water for the entire system. / by Hua Ru. / S.M.