Multi-analysis Ensamble in Polar Regions

LUCINI, MARÍA MAGDALENA; SHEPHERD, THEODORE; MCFARLANE, NORMAN; POLAVARAPU, SAROJA

Conferencia; American Geophysical Union 2010: The meeting of the Americas; 2010

American Geophysical Union (AGU)

<!-- @page { margin: 2cm } P { margin-bottom: 0.21cm } --> t is widely known that an ensemble of model analyses (multi model analyses) outperforms the analysis of a single model. How to combine these models is a matter of debate, while some authors say that averaging the ensemble members is accurate and reliable enough, some others propose other methods trying to improve results obtained from an average of model analyses.Our goal is to develop a statistical framework to obtain optimal estimates of the weights assigned to each ensemble member, identifying and quantifying at the same time model errors and uncertainties.In this work we present the methodology used for this task and we present a case study using an ensemble of middle atmospheric analyses from different operational and research centers and observations of temperature and ozone, all of them specially produced for the International Polar Year 2007-2008 (IPY) project, to produce a best estimate of the state of the atmosphere in the polar regions.