On 22 June ECMWF’s Council approved the Italian city of Bologna as the location of ECMWF’s new data centre. This is a vitally important decision: the Centre’s premises in Reading are unable to support the growth in computing power on which continued progress in numerical weather prediction (NWP) depends. To understand this dependence, we need to remember that the skill of global weather forecasts which we take for granted today has been hard won. It is the result of many incremental improvements: in the number and quality of Earth system observations; in data assimilation; in Earth system modelling and ensemble methods; and in the resolution at which we can analyse and predict the state of the atmosphere and related parts of the Earth system.
These improvements have been made possible by scientific advances, of course, but also by sustained growth in computational capacity and substantial efficiency gains. To extend the range of skilful ensemble forecasts further, as our ten-year Strategy demands, progress needs to be made on all these fronts. The upgrade of our Integrated Forecasting System (IFS) described in this Newsletter is a case in point. It brings improvements both in modelling and in the number and use of observations, notably leading to better predictions of tropical cyclones. But it also includes changes in software infrastructure which increase efficiency and prepare the ground for future improvements.
ECMWF’s new interpolation package MIR presented in this Newsletter is another illustration of how NWP research and software as well as hardware development must go hand in hand. MIR has been developed in response to the introduction of different NWP grids and parameters over the years as well as the development of new software and hardware technologies. Its flexible design and links to the IFS’s Atlas library ensure that software and operational services will be able to react fast to new research developments.
Meanwhile, the results of monitoring thin sea ice in the Arctic presented in this Newsletter illustrate the way in which a growing number of satellite observations can feed into better Earth system modelling for NWP. A new sea-ice thickness product provided by the University of Hamburg, based on satellite observations from the European Space Agency’s SMOS mission, is helping to evaluate the performance of ECMWF’s OCEAN5 ocean/sea-ice model implemented last year.
A final remark on Bologna: the data centre decision was the subject of the first ever tweet posted by the Centre on its official Twitter account. While we have had a corporate presence on LinkedIn for some time, using Twitter and Facebook marks a new departure for the Centre. Our plan is to use those platforms to better share our science with users and partners around the world. ECMWF will begin to use these channels regularly from September.