A combined used of Meteosat water vapor imagery and potential vorticity inversion to improve the numerical prediction of the Algiers 2001 superstorm
In spite of recent improvement in operational forecasting, numerical weather predictions regularly failed to reproduce intensity and distribution of precipitation of heavy rainstorms. Generally, forecast errors arise from imperfect parametrizations of the numerical model used or errors in the representation of the initial state of the atmosphere, especially over data sparse areas such as Mediterranean sea and Atlantic ocean.
Several methods have been developed in order to improve the prediction of intense precipitation. For example, new methods consisting of assimilating mesoscale data (such as temperature, humidity) in initial fields provided to high resolution model have been recently used and have proven their ability to improve precipitation forecast.
This study proposes an alternative method using the link existing between water vapour (WV) Meteosat imagery and potential vorticity (PV) analysis. In fact, a close relationship linking Meteosat WV imagery and PV positive anomalies allows one to detect significant upper-level dynamical structures associated with synoptic developments. This technique is used to assess and correct initial conditions of high resolution numerical simulations performed with the French non-hydrostatic Meso-NH model.
The Algiers 2001 superstorm is investigated using this approach. It corresponds to a deep tropopause dynamic anomaly crossing western Europe and north-west Africa between 9 and 11 November 2001 and followed by a surface cyclogenesis associated with strong convection which developed over Algiers. Thus, during the morning of 10 November, more than 130 mm of precipitation were recorded at Bouzareah and resulted in mudslides which devastated the Bal-el-Oued district. This disaster caused more than 700 casualties and catastrophic damage.
In this case study, an assessment of initial conditions provided to Meso-NH revealed a clear mismatch between PV analysis and Meteosat WV imagery leading to a bad prediction of the cyclogenesis development and of the rainfall distribution. It is demonstrated that a correction of this error improve significantly precipitation forecast principally thanks to a better prediction of the cyclone trajectory.
HyMeX – Hydrological cycle in the Mediterranean Experiment 2010-2020