The day-to-day business of many scientists in the Met Office is focussed on continually improving our Unified Model (UM) and advancing our treatment of observations that are used to initialise the UM to maintain our world-leading forecast accuracy.
Tom Blackmore, a Met Office senior scientist, said: “The Unified Model provides weather and climate forecasts across timescales ranging from hours to decades for the public and government, and a range of commercial customers. We constantly strive for better forecast performance and our work across a wide range of scientific and technical disciplines continually delivers many incremental improvements in skill which collectively add up to significant gains to our forecasting accuracy.
“We have often been asked how we bring new science into our operational forecasting model, which runs every day of the year to provide forecasts to our customers, without interruption. We do this by making changes – based on our scientific and technical knowledge, novel research and collaboration – to a copy of the model and rigorously test these changes.”
Several times a year the successful individual changes are bought together into a ‘package’ and all tested together in a copy of the model that is run in parallel to the operational model. Met Office operational meteorologists and scientists work together to ensure that all the changes in this Parallel Suite work together and provide improvements over the current operational version of the model, from global to very small scales. Once this has been carefully assessed, the Parallel Suite is switched to replace the operational model and thereafter becomes the version of the model that produces our world-leading forecasts.
The latest upgrade to our model was made this spring when our 37th Parallel Suite was switched to operational use. This was the first science upgrade to go operational on our new supercomputer and is the culmination of several years work from scientists from different disciplines across the Met Office.
The package of changes involved better use of the observations we receive from existing weather satellites, the introduction of observations from six new satellites and measurements of humidity from aircraft. One of the largest performance gains came from a better way of automatically and adaptively correcting errors in satellite observations. This allows us to use more, higher-quality, satellite observations in the model – providing a more accurate starting point for each forecast.
Careful monitoring of the Parallel Suite containing these changes, over months of testing and comparisons with our operational model during 2015 and early 2016, showed significant improvements in forecast skill. It became clear this amounted to one of the largest performance gains in our global model in the past decade.
Kirsty McBeath, Private Secretary to the Met Office Chief Scientist, said: “Now that Parallel Suite 37 is operational, our scientists are working on improvements to be considered for the next Parallel Suite, due to become operational later in 2016.
“This process of constant improvement and innovation as usual helps to ensure that the Met Office is consistently one of the top two operational weather forecasting services in the world, and we continually strive to ensure that our forecasts for the public, the government and our customers all over the world are as accurate and useful as possible.”
What utter nonsense. How can you have a “climate forecast” of “hours” duration. This is what sensible unbiased people used to call a weather forecast.