On Friday, October 13th, we will have an OCP Seminar given by Dr. Tim Palmer, a Royal Society Research Professor in Climate Physics and a Senior Fellow at the Oxford Martin Institute.
This will be a hybrid in-person/zoom seminar taking place in the Monell Auditorium. Please email the event contact for the zoom information. The title and abstract are provided down below.
Title: Circulation Regimes, Local Air-Sea Interaction and the Signal-to-noise Paradox
Abstract: In the early days of ensemble forecast development, the overwhelming challenge was to solve the problem that ensembles were typically under-dispersive, making ensemble forecasts overconfident. The development of singular vectors, stochastic parametrisations and multi-model ensembles helped alleviate this problem (to the extent that in the medium-range at least, contemporary ensemble forecast systems are remarkably reliable). In recent years, however, an opposite problem has emerged, particularly on the seasonal timescale. Here ensemble forecasts of large-scale patterns like the North Atlantic Oscillation appear to be over dispersive - this is known as the Signal to Noise Paradox. We present an argument that this latter problem is associated with a systematic bias in contemporary climate models: that of being unable to represent semi-persistent circulation regimes adequately. Paradoxically perhaps, stochastic parametrisation may help alleviate the problem. However, fundamentally the issue appears to be one of inadequate model resolution. Plans to develop a new generation of km-scale global model ensembles through the international programme EVE will be outlined.