Forecasting the weather accurately relies on a combination of cutting edge forecasting models developed by our research scientists and the skill of interpreting these models by expert meteorologists.

We’ve recently completed a “Forecasting Experiment” which brought together our research and operational expertise for two weeks.

The aim of this was to intensively evaluate and develop our current and experimental models, forecasting techniques, new scientific products and the interaction between them. The overall aim was to evaluate our forecasting capability so we can identify ways to continue improving the way we do things.

Experts from a range of backgrounds analyse model output in the ‘Forecast Experiment’

We have a number of computer models which can represent weather processes which occur on ever smaller scales.

These models cover different areas of the globe: from the entire world, down to a model which is confined to the UK area and even an experimental model which looks at the weather over an area the size of a large city.

All these models are incredibly complex and have strengths and weaknesses which can be difficult to determine.

By bringing together specialists with considerably different skills and experience in the Forecasting Experiment, it is easier to identify characteristics of the models. This can then inform better interpretation of all the information available to forecasters and can also help with planning improvements in future generations of the models.

The Forecasting Experiment this year was focused on summer weather over the UK. The experiment ran over the end of June and beginning of July.

The first week was characterised by a succession of weather fronts moving eastwards from the North Atlantic over the UK. The second week was dominated by very warm air from Spain moving north across the UK and bringing the hottest July day on record (36.7 °C at Heathrow) with severe thunderstorms to much of the UK, especially the North of England and Scotland.

This mixture of weather types allowed researchers to test models and techniques under a whole range of summer conditions, including high impact weather.

Adrian Semple,who led the experiment, said: “Not only did this experiment promote an exchange of skills, knowledge and experience between the participants, but it also provided a unique environment in which we could critically assess the way in which we produce our forecasts. The experiment will therefore have immediate effects as skills and knowledge are shared and spread throughout the Met Office, but the results can also be used to influence our longer term scientific research and improve future forecasting models.”