As important as the ability to run the next generation modelling systems is the ability for scientists to know what the models are doing. This means that they have to be able to get out of the models the quantities that they need to do that analysis. These quantities are called the model diagnostics. It has been estimated that there are about 5000 diagnostics in the current Unified Model (UM) system. Not all of these are used all of the time. It is therefore important that scientists have a way of telling the model which diagnostics they want when. The LFRic Diagnostic Infrastructure project has been busy designing just such a scheme for the new GungHo/LFRic-based next generation atmosphere model.
The completion of the LFRic Diagnostic Infrastructure project marks a significant step in delivering the next-generation modelling capability within the Met Office. The system was built using up to date, object-orientated software engineering methods. This object orientated design gives us a high degree of extensibility. Any future model developments or new metadata standards can easily be incorporated into the system.
We’ve also utilised modern information systems design, such as Single Source of Truth. All metadata are now defined in one place within the model. This means that any change or update to metadata automatically cascades to all parts of the system.
The system was also designed with intuitiveness in mind. In contrast to the existing system, all parts of the system are human readable. This will allow scientists to configure and run next-generation models in a more flexible and intuitive manner, enabling scientists to spend more time on the science and less time on configuring tools.