You are viewing an old version of this page. View the current version.

Compare with Current View Page History

Version 1 Current »

Line convection on very active cold fronts can be difficult for the models to resolve.  Typically the line convection is quite narrow while the resolution of IFS is 9km.

In Cy49 (and earlier) the convection scheme any precipitation that is developed is allowed to fall to the surface immediately within the grid square.  Thus there can be some cells showing high precipitation rates and totals.  However, there is some under-indication of maximum precipitation rates.

In Cy50 the convection scheme transfers about 40% of forecast convective precipitation to the broad scale scheme.  This advects the precipitation while allowing it to fall driven by the underlying winds.  In this way the precipitation is effectively smeared downwind and the forecast precipitations appear reduced.

Clearly, the peak rates which exceeded 50mm in reality are not shown, as they were too small scale for a 9km resolution model to pick up.  There are hints in both model runs of more intense precipitation band associated with two cold fronts but not definitive in intensity nor location.

This effect should be taken into account where users anticipate line convection, and modifications made to output when advising customers.


Fig9.6.8-1: 

Example of a very active cold front on 13 March 2026 compared with T+24 forecasts from Cy49 and its replacement Cy50 DT 00UTC 12 Mar 2026, VT 00UTC 13 March 2026.  Radar imagery shows line convection crossing England.  The diagrams show precipitation rates.  Naturally the peak rates which exceeded 50mm in reality are not identified, as they were too small scale for a 9km resolution model to pick up.  Convective precipitation in Cy49 falls immediately within the grid square.   Some of the convective precipitation in Cy50 is transferred to the broad scale scheme and advects with the ambient winds while falling to the surface.  Thus the convective precipitation is spread over several grid squares and therefore is shown as less intense.  There are hints in both model runs of more intense precipitation band associated with two cold fronts but not definitive in intensity nor location. 

  • No labels