Since we all now use apps that rely on GFS model data, I think I might highlight improvements to the forecast system coming on stream in January 2015 here
A significant change is an increase in the number of datapoints
‘Increase horizontal resolution of the first segment of the forecast from Eulerian T574 (~27 km) to Semi-Lagrangian T1534 (~13 km), and extend the length of forecast from 192 hours to 240 hours’
There is also an interesting PPT on how the weather data in our apps is improving: here
Enjoy
Thanks for posting.
Most in the link reads like a foreign language that I do not dominate, but I did see the words ‘moisture’, ‘orographic’ and ‘mountain’.
Could this mean that the forecast of non-convective clouds and weather effects of local geographic features now become better?
Or would that be wishful thinking of this incurable optimist?
It is interesting as they have been working to improve the output to match the better European ECMWF model which has higher resolution amongst other differences. Getting the GFS model better is important for us as it is free vs the very expensive European output.
You can view the test data from the New GFS model and comparen Euro vs GFS here.=
Unfortunately, the increased horizontal resolution, the extended the length of forecast, together with the change from JPEG2000 compression to complex packing with second order spatial differentiation increased the available amount of data by roughly a factor of 10, which seems to now have finally brought the data server to its knees…
Getting the GFS model better is important for us as it is free vs the very expensive European output.
Jason, In the US we pay a lot of taxes which pays for the National Weather Service, which runs the weather model. Therefore, the weather data belong to the people that pay the taxes, hence the data are free. I thought Europeans pay taxes too?
I had an interesting chat about forecasting and models at the Boat Show with a met office forecaster. Currently, models have a fairly rough terrain model in them, and are resolved at 1.2km “pixel” size, and local forecasting knowledge applied to get the best fit for the forecast.
The Met Office are trialling a new model with 330m resolution initially for areas needing especially accurate wind and weather forecasts (it was first used for the sailing at the 2012 Olympics), and also to predict winds at Heathrow to forecast arrival rates and spacing required. With the new Met Office supercomputer, they are hoping ultimately to go to that resolution across the UK.
