Research at the frontiers of knowledge to develop an integrated global model of the Earth system to produce forecasts with increasing fidelity on time ranges up to one year ahead. This will tackle the most difficult problems in numerical weather prediction such as the currently low level of predictive skill of European weather for a month ahead.
Operational ensemble-based analyses and predictions that describe the range of possible scenarios and their likelihood of occurrence and that raise the international bar for quality and operational reliability. Skill in medium-range weather predictions in 2016, on average, extends to about one week ahead. By 2025 the goal is to make skilful ensemble predictions of high-impact weather up to two weeks ahead. By developing a seamless approach, we also aim to predict large-scale patterns and regime transitions up to four weeks ahead, and global-scale anomalies up to a year ahead.
The latest operational change implementing this strategy is the wide-ranging upgrade of ECMWF’s Integrated Forecasting System (IFS), implemented on 5 June 2018, which brings better global weather forecasts, with particularly consistent gains in the extended range. A key plank of the upgrade is enhanced dynamic coupling between the ocean, sea ice and the atmosphere.
The upgrade improves forecast quality in the tropics, enabling better predictions in the extended range, improves 2-metre temperature forecasts, brings better predictions of rain in coastal areas as a result of improved cloud physics, and improves high-resolution forecasts in situations marked by rapid interactions between sea ice, the ocean and the atmosphere, such as the passage of tropical cyclones or rapid changes in sea ice cover.
Up to now, dynamic ocean, sea ice and atmosphere coupling has only been applied in medium-range ensemble forecasts (18 km horizontal resolution), extended-range forecasts and seasonal forecasts. The upgrade extends this coupling to ECMWF’s medium-range high-resolution forecasts (9 km horizontal resolution). As a result, the IFS is now more consistent and seamless across different timescales and spatial resolutions.
Other changes from IFS Cycle 43r3 to IFS Cycle 45r1 include a better use of observations, notably through a scheme to account for the horizontal drift of radiosondes during their ascent; the use of more satellite observations, such as non-surface-sensitive infrared channels over land and all-sky microwave sounding channels over coasts; and the introduction of new products useful in the prediction of severe weather, including forecasts of lightning flash density, and of maximum convective available potential energy (CAPE) as well as CAPE-shear (CAPES) over the last six hours of the forecast.
Current developments to further improve the Earth system Assimilation and Modelling system are wide-ranging.
Particular emphasis is given to the underlying atmospheric physical processes, numerical methods, and model uncertainty representation in high-resolution ensembles of assimilations and forecasts, as well as coupled processes including land-surface, ocean, wave and sea-ice interactions. Atmospheric composition aspects of model development such as ozone radiation interactions or improved 3D aerosol climatology as well as land surface developments targeting both the hydrological and the carbon cycle are included. Substantial efforts are devoted to the technical infrastructure and code development to ensure adaptivity to future high performance computing architectures, with a flexible, modular and scalable forecast and data assimilation system at the forefront of scientific knowledge and computing ability. Algorithmic efficiency gains through the use of single precision, options for alternative discretization in the dynamical core, and the use of domain-specific language concepts to adapt flexibly to a range of not necessarily compatible future hardware architectures for the atmosphere as well as the other Earth system components are key development targets. Coupled Earth system modelling investigates optimal coupling strategies of the Earth’s surface with the atmosphere and improves the Earth surface sub-systems (land, ocean, sea-ice, waves, glaciers, and surface-layer). A new 9-layer soil and a 5-layer snow model with improved hydrological cycle are being tested.
The concept of Continuous Data Assimilation has recently developed as a major new theme. Continuous DA will be used to enable full use of all observations that arrive before the forecast model needs to be run, whilst retaining current operational schedules. Work to improve the efficiency and accuracy of the Ensemble of Data Assimilations (EDA) will make it possible to deliver a 50 member EDA at approximately the same computational cost of the current 25 member EDA. As the Ensemble forecast (ENS) is also composed of 50 members, the 50 member EDA is an important step towards a seamless assimilation and forecast system and will already provide significant benefits in terms of analysis and forecast accuracy and reliability. The data assimilation strategy is now also extending to other components of the Earth system, and towards coupled data assimilation. Current IFS cycles now include the modular possibility to use coupled assimilation following different approaches, ranging from weakly coupled assimilation (WCDA) to quasi strongly coupled assimilation (QSCDA).
In addition to these developments, major areas of work are the optimal use of observations, including satellite data, and predictability studies investigating the potential improvements which can be brought to the extended and seasonal ranges.
The presentation will highlight recent results along the lines of research described above, as well as some diagnostics on the performance of the ECMWF’s IFS forecasts.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner