ECMWF plans to begin the production of a second ECMWF Re-Analysis (ERA-40) during the first half of 1999. The plans for ERA-40 are outlined in the sister presentation by Gibson et al. at this AMS conference.
The first ECMWF reanalysis, ERA-15, has been available to scientists for some time, and much experience has been gained on its strengths and weaknesses. Some were discussed at the first WCRP reanalysis conference in Silver Springs in October 1997 and further work has revealed new details. Before commencing the ERA-40 production in earnest, it is essential to validate the intended system against the ERA-15 experience. In particular we need to demonstrate that known weaknesses and errors in ERA-15 will be eliminated or at least reduced in the revised system for ERA-40. Some of these deficiencies, such as the use of representative surface wind observations, are relatively easy to correct and validate. Others, such as imbalances in the air-sea exchanges and suspect "trends" caused by changes in the observing systems, as shown in the presentation by Fiorino et al. at this conference, will require substantial work.
Thus, a comprehensive set of experiments will be carried out before the actual production. Among these experiments are: 1) observing system experiments (OSEs), particularly on the presence or absence of satellite data, 2) spin-up and other properties of the air-sea exchanges, 3) the hydrological cycle, 4) the performance of the intended coupled wave model and 5), the behaviour of the high resolution stratosphere including ozone. Some results and comparisons with ERA-15 will be presented