Proposal: To consider the behaviour of two 'toy' models of the atmosphere (3 and 28 dimensional) and the High Resolution Limited Area Model (HIRLANI). I will investigate the effects on predictability, of both selection and uncertainty of the initial state. I will examine measurement of reliability based on the evolution of an ensemble of similar initial conditions.
I wish to assess improvements in prediction, when using larger ensembles of initial conditions, and improving both the model, and the quality of observations.
Also, I hope to consider some observational time series and distinguish between the relative strengths of deterministic and probabilistic modelling in the context of long range predictions.
I hope to disseminate, promote and exploit the findings by collaborating with the European Centre for Medium Range Weather Forecasting (ECNIWF Reading) and the Irish Meteorological Service.