Researchers at the Universities of Oxford and Helsinki describe OpenIFS, a version of ECMWF’s Integrated Forecasting System (IFS), as an “invaluable” tool for research and training.
Heikki Järvinen is a professor of meteorology at the University of Helsinki. He is also a passionate supporter of ECMWF’s OpenIFS project, which makes an easy-to-use version of the operational IFS available to universities and research institutes.
“You can take almost any problem in meteorology and approach it using the OpenIFS,” Heikki says. Speaking on the sidelines of the OpenIFS user meeting taking place at ECMWF from 10 to 12 June, he points out that using an operational model in meteorological training brings a number of benefits.
On the one hand, it helps to spread knowledge of ECMWF’s operational system in the meteorological community. On the other, it equips students with skills that are relevant to the job market. “Skills in coding, computing and methods are all important and can be practised using the OpenIFS,” Heikki stresses.
Professor Heikki Järvinen worked for ECMWF in the nineties and has been a member of the Centre’s Scientific Advisory Committee for eight years.
He mentions the example of an advanced course in numerical meteorology in which he used the OpenIFS earlier this year. “The students were required to download, compile and build the model, run it using a given initial state, plot some forecast fields, and then participate in group work,” he says, adding that all 18 participants obtained credits for their work.
Benefit to taxpayers
Applying the model in research projects at universities and other research centres is also a win-win situation, Heikki believes. ECMWF can apply the results in model development, while the researchers can see the impact their work is having in practice, he points out.
“When you use idealized models, the connection between researchers and model developers is more remote. Here you know that your work is useful as it can be applied operationally, to the direct benefit of European taxpayers,” he says.
All in all, Heikki feels the OpenIFS project is “an invaluable initiative”. As a member of ECMWF’s Scientific Advisory Committee for the past eight years, he has long been a keen observer of the project’s progress, and he helped to organise the first user meeting at the University of Helsinki in June 2013.
Now he has some ideas on how to develop OpenIFS activities further. “You could for example set up a competition for the best research proposal submitted by European students, and the winners could be offered a month-long internship at ECMWF,” he suggests.
“Great support”
Aneesh Subramanian, a post-doctoral scholar in the Predictability of Weather and Climate group at the University of Oxford, came to know about OpenIFS much more recently than Heikki. But he is no less enthusiastic about the project.
Aneesh points out that most members of his group, including about eight post-doctoral researchers, work with ECMWF’s IFS model.
Dr Aneesh Subramanian joined the Predictability of Weather and Climate group at the University of Oxford after obtaining a PhD at the Scripps Institution of Oceanography in San Diego, California.
Aneesh is testing the stochastic parametrization of dynamical cores using the OpenIFS in an idealized setup. He is also using the OpenIFS single-column model in a research project to test and tune convection parametrization.
Filip Váňa, who supports the single-column model in the OpenIFS team at ECMWF, provided “great support” in preparing the model, which was “easy to set up on the system at Oxford”, Aneesh says.
In turn, Aneesh helped ECMWF’s OpenIFS coordinator, Glenn Carver, with some of the practical sessions at this year’s user meeting. And now he is looking forward to applying some of the exercises in his teaching at Oxford.
“The big thing I’d like to take back from the workshop is the great exercise for ensemble forecasting prepared by Glenn and others, including the Metview group. I see myself using this exercise as a teaching tool on a course on ensemble and probabilistic forecasting that we’d like to set up over the next year,” he says.
Four years after the OpenIFS project was started at ECMWF, there is a growing sense that the efforts of its founders are beginning to bear fruit.