• ensemble prediction system;
  • ensemble sensitivity;
  • Southwest Vortex;
  • ensemble analysis;
  • targeted observation;
  • real-time research


In this study, an ensemble analysis and sensitivity experiment of a mesovortex over southwest China (Southwest Vortex, SWV) are performed with a regional ensemble prediction system in order to demonstrate: (i) the usefulness of an ensemble prediction system for diagnosing the dynamical and thermodynamical characteristics leading to the SWV's eastward movement and growth; and (ii) the potential importance of adopting targeted observations in improving short-range weather forecasts by applying a simple technique. Results show significant variations in forecast performance, ranging from good to poor, among 11 ensemble members. By examining two distinct clusters (‘moving’ and ‘staggering’) of the ensemble members it is found that a strong upper-level trough with positive vorticity advection near the vortex centre, its westward vertical tilt and lower-level downstream wet and upstream dry moisture configuration favour the eastward movement and growth of the SWV. Ensemble sensitivity analysis shows that: (i) a simple correlation-based ensemble sensitivity method works well in identifying sensitive areas of mesoscale forecasts of the SWV; (ii) the forecast of the SWV movement and intensity is especially sensitive to initial conditions of horizontal wind and moisture fields; and (iii) unlike the medium-range forecast in which a sensitive region is often traced back to an area far upstream in association with a Rossby-wave packet, sensitive regions in short-range forecasts appear to be directly related to the meteorological system of interest located not far upstream. Based on the results, we suggest that field experiments of targeted observations potentially can be designed using this simple ensemble sensitivity approach to improve short-range weather forecasts associated with SWVs.