The Nowcasting Demonstration Project: case study of line convection on 28 May 2012
Temperatures had risen above 27 °C somewhere in the UK each day from 23 to 28 May 2012. On 28 May, thunderstorms associated with a weak weather front produced lightning strikes and flash floods and brought the heatwave to an end.
All times in this article are in Coordinated Universal Time (UTC).
The thunderstorms started at about 13:30 in London and then more generated in a line from London SE to NW to Leicester with separate cells in London, near Luton and Market Harborough forming by 14:00. The line moved north-eastwards with more cells being generated along the line with time. Up to 300 lightning strikes were reported between Bedfordshire and Kent and roads flooded in Rainham, Essex, with the hour to 16:00 seeing more than an inch of rain.
At this time the NDP was producing just 6 hour forecasts so that the first cycle which could predict the outbreak of the line of convection was the 08:00 cycle and indeed it did forecast cells in Kent, London and near Luton but missed the cell in Leicestershire. All cycles from 08:00 onwards forecast convection in the line with slightly varying degrees of accuracy.
The figure below shows, on the left, the locations of the 129 detected lightning strikes in the hour preceding 15:00 on 28 May 2012 and, on the right, the convergence visible in the Doppler radar winds at 15:00 seen from the Chenies radar located at Flaunden, Hemel Hempstead, Herts.
Click on "Enlarge" to see the following figure in full. It shows comparisons of (top right) the observed radar derived surface precipitation rate at 15:00 with (top left) the NDP 5 hour forecast from 10:00, (bottom left) the current operational UKPP nowcast from 10:00 using extrapolated radar derived rain rates blended with a 4 km resolution UK forecast from 03:00 and (bottom right) the latest available The UKV model - Kilometre-scale forecasting over the UK with the Unified Model 12 hour forecast from 03:00. The NDP has a very good forecast of the location of the line of convection, the current operational nowcast has nothing because the latest UK 4 km forecast from 03:00 had no convection at 15:00 and the UKV has some convection but too far east. This is a very interesting case as there was no precipitation between 09:30 and 10:30 when the 10:00 NDP was initialized so the improved forecast has not come from Doppler radar winds or latent heat nudging.
In the figure below (again, click "Enlarge" to see in full) the UKV 6 hour forecast for 15:00 from 09:00 is shown top left. This would have been available just after the 10:00 NDP. This forecast has more convection but still too far east and too widespread. The UKV forecast includes GPS water vapour column data which arrived too late for the NDP forecasts. Including GPS data in the NDP forecasts caused spurious convection in a similar location to the UKV indicating that improved ways of using the GPS data in the 3D-Var and 4D-Var data assimilation may be required. The 15:00 UKV analysis is shown below top right and also has an error in the location of the line of thunderstorms and looks very similar to the T+6 forecast. The current operational UKPP nowcast analysis at 15:00 is shown bottom left, this is essentially just the radar derived rain rates interpolated to a 2 km grid. The NDP 15:00 analysis is shown bottom right which is in reasonable agreement with the radar derived rain rates. The reasons for the differences between the NDP and UKV analyses and forecasts are now being investigated.
More Information on the NDP system and example forecasts can be found by following links below:
- The Nowcasting Demonstration Project for the London 2012 Olympic and Paralympic Games for the London 2012 Olympic and Paralympic Games
- The Data Assimilation System for the Nowcasting Demonstration Project (NDP)
- The Nowcasting Demonstration Project: case study of 7 May 2012 Oxfordshire tornado Oxfordshire tornado