PhD Project: Forecast calibration and combination: Bayesian assimilation of seasonal climate predictions




 Caio Augusto dos Santos Coelho

Department of Meteorology – University of Reading

Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq - Brazil


Supervisor: Dr. David B. Stephenson

Department of Meteorology – University of Reading


Co-Supervisor: Dr. Francisco J. Doblas-Reyes

 European Centre for Medium-Range Weather Forecasting - ECMWF



The main proposes of this PhD project are:


1)      To produce improved seasonal forecasts of probability of South American rainfall through the development of statistical seasonal forecasting methods;

2)      To check the skill of the Southern Hemisphere ensemble seasonal climatic forecasts that are going to be made at ECMWF using the state-of-the-art coupled model as part of the DEMETER project. This will hopefully allow us to judge how predictable climate is over South America on seasonal time scales during ENSO events.



Methodology: The skill of ENSO and South American rainfall seasonal forecasts will be studied through the use of statistical techniques applied to climate analyses. A probabilistic Bayesian approach will be used to combine historical information with seasonal forecasts. Parametric approaches will be used to estimate probability density functions (p.d.f´s).