Chapter 13
THE ECMWF PERSPECTIVE
David Anderson, Magdalena Balmaseda, and Arthur Vidard
ECMWF, Reading, UK
Abstract:
ECMWF is involved in forecasting on time-scales from the medium range
(days) to seasonal (6 months) ahead. This requires the preparation of initial
conditions from which to start forecasts. For the medium range, a huge effort
has been devoted to developing the most advanced assimilation strategies for
analyzing the atmospheric state. Considerable effort has also been devoted to
retrospectively analyzing the atmospheric state (ERA-40) using a more
advanced assimilation system than was available at the time of the operational
analysis. For forecasts beyond the medium range, coupled atmosphere-ocean
models are used, requiring analyses of the ocean state. These in turn depend
heavily on atmospheric analyses and reanalyses. Aspects of the atmospheric
and oceanic assimilation systems used operationally are discussed, together
with some limitations of current systems.
Key words: ECMWF, forecasting, atmospheric analyses, reanalyses.
1.
Introduction
In considering the merits of a data assimilation system, it is necessary to
have the application in mind. That is after all what sets the measure by
which one decides if the scheme works well or not. It is one thing to
formulate an assimilation strategy theoretically but altogether a different
matter to develop a practical implementation. No ‘operational’ assimilation
system conforms to its theoretical optimum configuration.
At ECMWF (European Centre for Medium-Range Weather Forecasts),
data assimilation is used in a number of applications. The key forecast
applications are: deterministic forecasts for the medium range, ensemble
prediction system (EPS) for the medium range, an ensemble of monthly
forecasts and an ensemble of seasonal forecasts. The time ranges are 10 days
for medium range, 31 days for monthly forecasts, 190 days for seasonal
forecasts. As might be expected, the resolution of the model is not the same
in all the applications. The model resolution decreases as one increases the
361
E. P. Chassignet and J. Verron (eds.), Ocean Weather Forecasting, 361-379.
© 2006 Springer. Printed in the Netherlands.
THE ECMWF PERSPECTIVE
David Anderson, Magdalena Balmaseda, and Arthur Vidard
ECMWF, Reading, UK
Abstract:
ECMWF is involved in forecasting on time-scales from the medium range
(days) to seasonal (6 months) ahead. This requires the preparation of initial
conditions from which to start forecasts. For the medium range, a huge effort
has been devoted to developing the most advanced assimilation strategies for
analyzing the atmospheric state. Considerable effort has also been devoted to
retrospectively analyzing the atmospheric state (ERA-40) using a more
advanced assimilation system than was available at the time of the operational
analysis. For forecasts beyond the medium range, coupled atmosphere-ocean
models are used, requiring analyses of the ocean state. These in turn depend
heavily on atmospheric analyses and reanalyses. Aspects of the atmospheric
and oceanic assimilation systems used operationally are discussed, together
with some limitations of current systems.
Key words: ECMWF, forecasting, atmospheric analyses, reanalyses.
1.
Introduction
In considering the merits of a data assimilation system, it is necessary to
have the application in mind. That is after all what sets the measure by
which one decides if the scheme works well or not. It is one thing to
formulate an assimilation strategy theoretically but altogether a different
matter to develop a practical implementation. No ‘operational’ assimilation
system conforms to its theoretical optimum configuration.
At ECMWF (European Centre for Medium-Range Weather Forecasts),
data assimilation is used in a number of applications. The key forecast
applications are: deterministic forecasts for the medium range, ensemble
prediction system (EPS) for the medium range, an ensemble of monthly
forecasts and an ensemble of seasonal forecasts. The time ranges are 10 days
for medium range, 31 days for monthly forecasts, 190 days for seasonal
forecasts. As might be expected, the resolution of the model is not the same
in all the applications. The model resolution decreases as one increases the
361
E. P. Chassignet and J. Verron (eds.), Ocean Weather Forecasting, 361-379.
© 2006 Springer. Printed in the Netherlands.