For reasons of safety and passenger comfort aircraft often fly around large thunderstorms, and thus air traffic management need advance notice of thunderstorm activity to enable them to adjust air traffic flow before major disruption occurs. An initial project investigating short range thunderstorm forecasts in the London area from the GANDOLF (Generating Advanced Nowcasts for the Deployment of Operational Land-based Flood forecasts) system, showed that GANDOLF forecasts could provide useful information up to about 90 minutes ahead. Ideally, forecasts of disruption due to thunderstorms would be as far in advance as possible, whilst maintaining good spatial accuracy. Hence, this follow-up project aims to quantify the trade-off between the spatial and lead time accuracies of thunderstorm forecasts, for three different types of Met Office forecasts up to 12 hours ahead. These forecasts will be for thunderstorm activity that occurred over the UK during summer 2001, chosen using lightning data. They will be verified in terms of storm intensity and coverage, as studies of pilots' storm penetration/deviation behaviour have shown these to be two key decision factors. Verification will be done against radar observations and forecast performance assessed at each combination of spatial resolution and lead time. This will provide air traffic management with full information as to the utility and accuracy of the Met Office thunderstorm forecasts.