Ensemble simulations with perturbed physical parametrizations: Pre‐HyMeX case studies. (28th November 2013)

• Record Type: Journal Article
• Authors: Hally, A.; Richard, E.; Fresnay, S.; Lambert, D.
• Is Part Of: Quarterly journal of the Royal Meteorological Society. Volume 140:Number 683(2014:Jul.)
• Page Start: 1900
• Page End: 1916
• Publication Date: 2013-11-28
• Subjects: Meteorology -- Periodicals; 551.5
• DOI: 10.1002/qj.2257
• Languages: English
• ISSNs: 0035-9009

Abstract

<abstract abstract-type="main" xml:lang="en"> <title>Abstract</title> <p>Heavy precipitation events (HPEs) affect the southeastern area of France frequently during the months of September–November. Very high amounts of rain can fall during these events, with the ensuing flash floods causing widespread damage. The cases of 6 September 2010 and 1–4 November 2011 represent the different large‐scale conditions under which these episodes can occur. These HPEs are forecast with differing levels of skill by the Méso‐NH model at 2.5 km resolution. The case of 6 September 2010 is used to test different methods of addressing cloud physics parametrization uncertainties. Three ensembles are constructed, where the warm‐process microphysical time tendencies are perturbed by different methods. Results are compared by examining the spatio‐temporal distribution of the precipitation field as well as looking at ensemble statistics. The ensemble methodology that induces the most dispersion in the rainfall field is deemed the most suitable. This method is then used to examine the sensitivity of four cases from November 2011 to errors in the microphysical and turbulent parametrizations. It appears that the sensitivity to microphysical perturbations varies according to the model skill for the HPE. Events where the model skill is high (low) show low (moderate) sensitivity. These cases show a stronger sensitivity to perturbations performed upon the turbulent tendencies, while perturbing the microphysical and turbulent tendencies together produces even greater dispersion. The results show the importance and usefulness of ensembles with perturbed physical parametrizations in the forecasting of HPEs.</p> </abstract>