Towards an operational 1km model

The Met Office Operational forecast model has a non-hydrostatic formulation, which offers the scope to run forecasts at very high resolution. A Limited Area Model with a horizontal resolution of 4km is to be introduced operationally in May 2005 and tests have been run with a 1km version of the model. In research mode the model has been run with a horizontal resolution of 200 metres.

The expected benefit of a 1km model comes from both the ability to explicitly resolve convective storms (rather than depend on convection parametrization) and from more accurate representation of the surface orography. It is reasonable to anticipate that such a 'storm-resolving' model should be capable of much greater realism, but does that mean better forecasts?

One of the problems we now face is the difficulty in trying to forecast meteorological events, such as convective storms, that are inherently less predictable on the temporal and spatial scales we are now attempting to resolve. For that reason, considerable care is needed in the way the output is interpreted and verified. It is important to be able to determine the scales over which a model is sufficiently skilful for a particular purpose and present the output on those scales.

Examples will be presented to show that the 1km model does indeed have the potential to improve our ability to forecast severe convective rainfall events, both in terms of improved realism and spatial accuracy. The findings from individual cases are backed up by results from a verification technique that can be used to determine the variation in forecast skill over different spatial scales. This information can be used to provide suitable products for customers.