21 Jul (CLIM) Mary Korendyke, PhD Defense

Mary Korendyke, Doctor of Philosophy in Climate Dynamics

Holistic Forecasting of z500: A Combined Oscillation-Regime Framework

Mon, 21 Jul, 1:00pm, Exploratory Hall rm 3301

Adviser: David Straus

The difficulty in predicting the atmospheric circulation and weather beyond two weeks in advance arises because the predictability arising from knowledge of the initial conditions becomes small, while the information that comes from knowledge of the boundary conditions (such as tropical ocean temperatures) is useful on seasonal time scales. Improving the skill of weather prediction in the 2-4 week range would provide useful and actionable information for decision makers in many sectors of society, from water managers to forestry officials to emergency response planners. However, achieving that level of skill requires that traditional techniques used for forecasting the weather out to 10 days must be augmented by statistical techniques. Different techniques have been independently investigated in the literature for potential weather forecasting skill in the 2-4 week range. Each has strengths and weaknesses. In this dissertation, I use fundamental oscillations in the atmosphere to enhance a statistical model trained on the past evolution of large scale atmospheric patterns. These oscillations can be interpreted in terms of preferred patterns and results from this model can be interpreted in terms of circulation regimes. I analyze the strengths, weaknesses, and potential increase in weather forecasting skill of this enhancement to determine the viability of such holistic forecasting methods for the 2-4 week range.