At the end of 2020, 26 national hydrological and meteorological service (NHMS) providers from Europe and North Africa decided to join forces to improve short-range weather prediction. The move saw three regional consortia, ALADIN, LACE and HIRLAM, enter into a larger partnership called the ACCORD consortium, which stands for ‘A Consortium for COnvection-scale modelling Research and Development’. Dr Saji Varghese, head of research at Met Éireann, the Irish NHMS, and chair of ACCORD’s science committee, explains more
Why has Accord been set up?
ACCORD was set up to broaden and deepen the research collaboration on developing advanced high-resolution weather prediction capability for local areas. It will develop world-leading weather forecasting systems to provide the best possible support to society, based on knowledge leading research and pioneering supercomputing technology.
ACCORD is a merger of the three existing consortia, ALADIN, HIRLAM and LACE, resulting in a unique collaboration of 26 countries of Europe and Northern Africa. It was formed to further advance and strengthen research in the complex science of weather prediction, develop new algorithms to take advantage of the opportunities from new computer architectures and tackle the challenge of massive data flows with our combined expertise and effort. We will enhance and evolve the development of our numerical weather prediction model code and cooperate on new components of the modelling system.
Which partners are involved?
Partners involved are:
ALADIN: Algeria, Austria, Belgium, Bulgaria, Croatia, Czech Republic, France, Hungary, Morocco, Poland, Portugal, Romania, Slovakia, Slovenia, Tunisia, and Turkey.
LACE: Austria, Croatia, Czech Republic, Hungary, Poland, Romania, Slovakia, and Slovenia.
HIRLAM: Denmark, Estonia, Finland, Iceland, Ireland, Lithuania, Netherlands, Norway, Spain, and Sweden.
ACCORD is the largest international weather prediction research consortium in the world.
How will it work?
Prior to the formation of the ACCORD consortium, a research and development strategy for five years was jointly developed by the consortium members. Based on the strategy, a rolling work plan was devised and more than 100 scientists will contribute toward its implementation. The management structure consists of a program manager who will have the overall responsibility for implementation of the ACCORD work plan, area leaders for the different research areas and three leaders from the merged groups.
ACCORD consortium will also focus on convergence of different model configurations over the years and will address interoperability and integration issues.
While the consortium will collaborate on science and development, the production and delivery of operational weather forecasts will continue to be the responsibility of the national weather services. The scientific and technical innovations of the consortium will be implemented in the regular upgrades of the operational numerical weather prediction systems.
What stage is Accord at currently?
On November 27 2020, 26 national meteorological services signed an agreement reinforcing their collaboration to improve short-range weather predictions, and deliver a better service to society.
ACCORD is at a very early stage. However, it has the advantage that it is building on the previous partnership between ALADIN, LACE and HIRLAM consortiums. A robust governance and management structure along with a strategy for research and development is already in place. At this stage, all relevant personnel to manage the consortium have been appointed.
What meteorological networks will be used for predictions?
ACCORD builds on Europe’s impressive tradition of cooperation in meteorology. EUMETSAT provides Europe with advanced weather satellite capability; EUMETNET coordinates weather and climate observations across Europe; while ECMWF is the undisputed leader in global weather prediction. ACCORD will collaborate closely with each of these international organizations to provide the very best regional weather prediction capability for its members.
Does Accord plan to fill in any data/coverage gaps?
ACCCORD aims to build the world’s best numerical weather prediction system through research and development for the citizens of the countries involved in the consortium. While it uses observational data from different sources and platforms, it is not directly involved in measurements. However, through its research it can inform the different national weather services, data providers, EUMETNET and EUMETSAT on data coverage issues to be addressed.
ACCORD will also take advantage of the new meteorological observations from both conventional and non-conventional sources such as third party data sets, smartphones, automobiles, etc.
How does it aim to improve weather forecasting?
By bringing together the scientific and technological expertise from its 26 members, ACCORD has deployed the best talent to implement its strategy over the next five years. It will leverage the latest computational technology and develop hectometric scale models with high density observations for nowcasting and short-range forecasts. Coupling important components of the Earth system physical processes will enhance the forecast skills and contribute to impact-based forecasting. Further developments in ensemble forecasting and meteorological quality assurance will definitely strengthen forecasting capabilities.
What key solutions will it focus on first?
With rapidly increasing resolution of global models, moving towards hectometric resolutions of limited area models to maintain their added-value is a primary focus of ACCORD. Further development of ensemble forecast capabilities also would need immediate attention for many members. To achieve these goals, major scientific work is planned in the areas of dynamics, physics, surface, ensembles and verification. Development of nowcasting systems based on the NWP tools is also an important area.
What are the key challenges associated with delivering Accord?
Collaboration between 26 services is a major challenge for ACCORD. To mitigate this, modern collaborative tools will be deployed to facilitate efficient communication and code integration and evolution. The technical validation tools will also be further developed and made available.
From an operational perspective of short-term forecasting, one of the challenges is to improve extreme weather prediction. As mentioned earlier, ACCORD is the largest numerical weather prediction consortium in the world and leveraging its combined resources enables it to deal with complex issues in a collaborative manner.
A key objective of the consortium is to move toward development of hectometric scale models. ACCORD will strive to improve the representation of the Earth system and the various interactions between the atmosphere and the components of our environment. The modelling of the atmospheric composition and the application of AI will be introduced.
The consortium will work on adapting the NWP systems to future computer architectures. Probabilistic forecasting will be an important area of research. Data assimilation and use of a variety of observation sources will also be an essential part of ACCORD’s research program.
ACCORD will build on the previous ALADIN-LACE-HIRLAM partnership and strengthen the coordination with robust governance and management. It is important to maintain the current excellence in research while exploring new opportunities.
How will data/expertise be shared to develop one model?
The collaboration and work plans are arranged in such a way that there will be plenty of opportunities for scientists in the consortium to work together and also make exchange visits. Working weeks and annual staff meetings will be organized. Furthermore, plans are also in place to work toward a common configuration of model and to ensure knowledge-sharing among the consortium members.
How will national met agencies benefit from the new consortium?
With our joint effort, we will develop more integrated and efficient working methods, and minimize duplication of work by achieving a high degree of interoperability of the common codes. This will accelerate the improvement of weather forecasts. In this way we will help our fellow citizens to reduce weather related threats to life, health, economy and property, and support authorities to adapt their services to climate change.
Are there any key trends currently present in the meteorological sector that you believe will have an impact on the progress of Accord?
Global NWP systems such as those from ECMWF are increasing resolution and improving quality of the forecasts. It is important that regional weather models remain ahead both in terms of resolution and accuracy of forecasts.
High performance computing is a key element in weather forecasting. HPC technology continues to evolve rapidly, and it would be difficult to precisely predict the type of processors that would be available in five years’ time. Model codes should be adapted to ensure resilience toward future technology changes and take advantage of new possibilities. This is particularly important in the context of increasing horizontal spatial resolution in NWP systems.
The use of observations and appropriate treatment of physical processes is another area which will have a significant impact on the quality of forecasts. Use of new observation platforms and types have been increasing over the past years and continues to do so. Some examples are Mode-S aircraft data, surface-based remote sensing data, crowd-sourced data and GNSS data.
As in many other applications, use of AI and machine learning are also beginning to emerge in meteorological applications and I expect this to play an important role in the development and use of weather models.
What is the ultimate aim of Accord?
By sharing its limited resources, ACCORD aims for the efficient development of a world-leading operational NWP system. It will deliver a set of common codes that can be assembled under diverse configurations to support NWP systems on regional domains for its members.