Advances in recent decades in the science of seasonal-to-interannual climate prediction have created a technology with potentially substantial societal benefit. It is now possible to have more accurate advance warning of the range of seasonal climate patterns to expect with a lead-time of a few months, at least for many parts of the tropics and in a few higher latitude regions as well. In recent years, as attempts to use the information have intensified, it has become clear that a further step was needed: to translate the seasonal climate forecasts into the information that more directly mattered for decisions in such sectors as agriculture and water resource management. Often, this meant interpreting the seasonal climate forecasts at smaller spatial and temporal scales and in terms of new environmental variables and features - activities that have become variously referred to as downscaling. The approach in this course is from this perspective of a user or generator of seasonal forecast information wishing to understand more the regional scales of climate forecasting in order to have a better knowledge base for addressing regional and local problems.
To appreciate the potential to generate information at smaller scales and for new environmental variables, it is first necessary to grasp how seasonal prediction is possible at all. That is, to see how the nature of the coupled ocean-atmosphere interaction on our planet leads to modifications of the large-scale wind patterns to expect in the coming season (Part 1). Understanding how climate models capture that predictability provides further insight to the problem (Part 2). Downscaling approaches for seasonal prediction are placed in the context of the suite of current seasonal forecast methods (Part 3), leading to a detailed treatment of the current issues and potential for generating the more detailed forecast information to support decisions (Part 4). Each of the lectures has an accompanying practical exercise to illustrate some of the key concepts in the lectures. For their own selected geographical region of interest, students will analyze the observed seasonal rainfall variability since 1950, learn how to evaluate the potential seasonal predictability for that region, based on analysis of climate datasets (such as sea-surface temperature) and prediction output from Global Climate Models (GCMs), and finally gain experience in the difficult problem and associated uncertainties, in interpreting downscaled predictability.
The course assumes a general understanding of basic concepts underlying environmental science and numerical analysis (such as correlation and regression). However, the technical content is kept to a minimum, seeking more to provide conceptual understanding. It is anticipated that the course will be useful as part of introductory material for climate science or as part of a course in a related discipline, as well as for professionals involved in creating, using or communicating climate information.