Empirical techniques are developed to adjust dynamic model forecasts on the seasonal time scale for southern African summer rainfall. The techniques, called perfect prognosis and model output statistics (MOS), are utilized to statistically "recalibrate" general circulation model (GCM) large-scale fields to three equi-probable rainfall categories for December to February. The recalibration is applied to a GCM experiment where simultaneously observed sea-surface temperature (SST) fields serve as the lower boundary forcing, referred to as the simulation mode experiment. Cross-validation sensitivity tests are first performed over a 28-year climate period to design an optimal canonical correlation analysis (CCA) model for each of the two recalibration methods. After considering several potential predictor fields, the 700 hPa geopotential height field is selected as the single predictor field in the two sets of statistical equations that are subsequently used to produce recalibrated rainfall simulations over a 1 a-year independent test period. Patterns analysis of the predictor and predictand fields suggests that anomalously low (high) 700 hPa geopotential heights over the subcontinent are associated with wet (dry) conditions over land, an association that is supported by observational evidence of rain (drought) producing systems. Additionally, the dominant mode of the recalibration equations is associated with the EI Nino/Southern Oscillation (ENSO) phenomenon. Somewhat higher retro-active skill levels are found using the MOS technique, but the computationally less intensive perfect prognosis technique should also be able to produce usable seasonal rainfall forecasts over southern Africa in an operational forecast environment hampered by the lack of computing resources. / Dissertation (MSc)--University of Pretoria, 2006. / Geography, Geoinformatics and Meteorology / Unrestricted

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Meteorology and Physical Oceanography, 1982. / Microfiche copy available in Archives and Science / Bibliography: leaves 228-233. / Lynne D. Talley. / Ph.D.

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Meteorology, 1981. / Microfiche copy available in Archives and Science. / Bibliography: leaves 210-215. / by William Roy Young. / Ph.D.

Thesis (Sc.D.)--Massachusetts Institute of Technology, Dept. of Meteorology, 1979. / Microfiche copy available in Archives and Science. / Vita. / Bibliography: leaves 242-247. / by Antony Brett Mullan. / Sc.D.

Thesis. 1977. M.S.--Massachusetts Institute of Technology. Dept. of Meteorology. / Microfiche copy available in Archives and Science. / Bibliography : leaves 62-65. / M.S.

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Meteorology, 1979. / Microfiche copy available in Archives and Science. / Bibliography: leaves 231-234. / by James William Anthony Fullmer. / Ph.D.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Meteorology, 1979. / Microfiche copy available in Archives and Science. / Bibliography : leaf 80. / by John Lee Lillibridge III. / M.S.

This thesis addresses the problem of defining the control law for human respiration. Seven different drivers have been identified as possibly having an input to the respiratory controller. These seven represent a combination of feedforward and feedback inputs arising from neural and humoral mechanisms. Using the assumption that carbon dioxide concentrations in the arterial blood have the strongest effect, a control equation with proportional and derivative components based on this driver was evaluated. The methodology for the evaluation was to create a model of the respiratory system incorporating the P/D controller, obtain experimental data of one test subject's respiratory response to exercise, then compare model generated output with experimental data, and adjust the parameters in the control equation to yield optimal model performance. The usual practice of testing controller performance has been to apply single step loads to a model and evaluate its response. A multi-step protocol was used here to provide a better, more generalized test of controller performance. This thesis may represent the first documented use of an approach of this type for evaluating respiratory controller performance. Application of a multi-step protocol revealed a non-linear controller was needed to keep pace with system changes. Respiratory system operation was effectively managed using a controller of the form: VENTILATION = F(dCO₂/dT,Q) + F(CO₂,Q) + CONSTANT. / Master of Science

The onset of summer stratification in temperate lakes and reservoirs forces a decoupling of the hypolimnion from the epilimnion that is sustained by strong density gradients in the metalimnion. These strong gradients act as a barrier to the vertical transport of mass and scalars leading to bottom anoxia and subsequent nutrient release from the sediments. The stratification is intermittently overcome by turbulent mixing events that redistribute mass, heat, dissolved parameters and particulates in the vertical. The redistribution of ecological parameters then exerts some control over the ecological response of the lake. This dissertation is focused on the physics of deep vertical mixing that occurs beneath the well-mixed surface layer in stratified lakes and reservoirs. The overall aim is to improve the ability of numerical models to reproduce deep vertical mixing, thus providing better tools for water quality prediction and management. In the first part of this research the framework of a one-dimensional mixed-layer hydrodynamic model was used to construct a pseudo two-dimensional model that computes vertical fluxes generated by deep mixing processes. The parameterisations developed for the model were based on the relationship found between lake-wide vertical buoyancy flux and the first-order internal wave response of the lake to surface wind forcing. The ability of the model to reproduce the observed thermal structure in a range of lakes and reservoirs was greatly improved by incorporating an explicit turbulent benthic boundary layer routine. Although laterally-integrated models reproduce the net effect of turbulent mixing in a vertical sense, they fail to resolve the transient distribution of turbulent mixing events triggered by local flow properties defined at far smaller scales. Importantly, the distribution of events may promote tertiary motions and ecological niches. In the second part of the study a large body of microstructure data collected in Lake Kinneret, Israel, was used to show that the nature of turbulent mixing events varied considerably between the epilimnion, metalimnion, hypolimnion and benthic boundary layer, yet the turbulent scales of the events and the buoyancy flux they produced collapsed into functions of the local gradient Richardson number. It was found that the most intense events in the metalimnion were triggered by high-frequency waves generated near the surface that grew and imparted a strain on the metalimnion density field, which led to secondary instabilities with low gradient Richardson numbers. The microstructure observations suggest that the local gradient Richardson number could be used to parameterise vertical mixing in coarse-grid numerical models of lakes and reservoirs. However, any effort to incorporate such parameterisations becomes meaningless without measures to reduce numerical diffusion, which often dominates over parameterised physical mixing. As a third part of the research, an explicit filtering tool was developed to negate numerical diffusion in a threedimensional hydrodynamic model. The adaptive filter ensured that temperature gradients in the metalimnion remained within bounds of the measured values and so the computation preserved the spectrum of internal wave motions that trigger diapycnal mixing events in the deeper reaches of a lake. The results showed that the ratio of physical to numerical diffusion is dictated by the character of the dominant internal wave motions.

Thesis (Ph. D.)--Joint Program in Physical Oceanography (Massachusetts Institute of Technology, Dept. of Earth, Atmospheric, and Planetary Sciences; and the Woods Hole Oceanographic Institution), 2008. / Includes bibliographical references (p. [221]-232). / Marine ecosystems reflect the physical structure of their environment and the biological processes they carry out. This leads to spatial heterogeneity and temporal variability, some of which is imposed externally and some of which emerges from the ecological mechanisms themselves. The main focus of this thesis is on the formation of spatial patterns in the distribution of zooplankton arising from social interactions between individuals. In the Southern Ocean, krill often assemble in swarms and schools, the dynamics of which have important ecological consequences. Mathematical and numerical models are employed to study the interplay of biological and physical processes that contribute to the observed patchiness. The evolution of social behavior is simulated in a theoretical framework that includes zooplankton population dynamics, swimming behavior, and some aspects of the variability inherent to fluid environments. First, I formulate a model of resource utilization by a stage-structured predator population with density-dependent reproduction. Second, I incorporate the predator-prey dynamics into a spatially-explicit model, in which aggregations develop spontaneously as a result of linear instability of the uniform distribution. In this idealized ecosystem, benefits related to the local abundance of mates are offset by the cost of having to share resources with other group members. Third, I derive a weakly nonlinear approximation for the steady-state distributions of predator and prey biomass that captures the spatial patterns driven by social tendencies. Fourth, I simulate the schooling behavior of zooplankton in a variable environment; when turbulent flows generate patchiness in the resource field, schools can forage more efficiently than individuals. / (cont.) Taken together, these chapters demonstrate that aggregation/ schooling can indeed be the favored behavior when (i) reproduction (or other survival measures) increases with density in part of the range and (ii) mixing of prey into patches is rapid enough to offset the depletion. In the final two chapters, I consider sources of temporal variability in marine ecosystems. External perturbations amplified by nonlinear ecological interactions induce transient ex-cursions away from equilibrium; in predator-prey dynamics the amplitude and duration of these transients are controlled by biological processes such as growth and mortality. In the Southern Ocean, large-scale winds associated with ENSO and the Southern Annular Mode cause convective mixing, which in turn drives air-sea fluxes of carbon dioxide and oxygen. Whether driven by stochastic fluctuations or by climatic phenomena, variability of the biogeochemical/physical environment has implications for ecosystem dynamics. / by Ariane Verdy. / Ph.D.