The value of high resolution models for simulating intense rainfall

In research just published in the Journal of Climate we have compared two Met Office high resolution regional climate model simulations of sub-daily extreme rainfall – a 1.5 km model and a 12 km model, to identify where the benefits of such models may be found.

By comparing the models with observations we found that the 12 km model has lower biases in mean rainfall intensity compared with the 1.5 km model but these biases are, in part, a compensation of too much light rainfall and excessively high extreme intensities.  However, the more detailed model is more successful than the coarser model in representing “very extreme” (multi-) hourly summer events (i.e. those with long return periods).  As accumulation periods lengthen toward daily timescales however, the erroneous 12 km rainfall extremes become more comparable with the observations and with the 1.5 km model.  In winter, both model simulations have comparable simulations of extremes.

We conclude that caution is needed in interpreting the realism of summer extremes in the 12 km model (and coarser resolution models in general) and that the higher degree of realism in the 1.5 km model is a source of confidence in its projections of the future climate.  There is, therefore, added value in the convection-permitting 1.5 km model in the simulation of summer extremes, particularly for the most extreme events.  As the 12 km model appears sufficient to simulate winter extremes however the advantage of the 1.5 km model for this season is relatively low.

The research is available on early view in the Journal of Climate  and is titled “The value of high-resolution Met Office regional climate models in the simulation of multi-hourly precipitation extremes” by Steven Chan et al. 

Last modified: Mon, 16 Jun 2014 10:41:29 BST