common navigation bar
FASTEX: a quick tour To first To previous To next To last

JavaScript fast nav Menu:

About adaptive observation to improve cyclone prediction

  • the new approach to observation suggested by recent work on the predictability of cyclones has been tested for the first time in real time and with real facilities on real cases: see below, Fig. F, G and H)
  • unique datasets have been obtained by the ships on turbulent fluxes under strong winds and high seas, but also on oceanography-related topics.
Example of cyclogenesis forecast Figure F: An early forecast for 19 Feb. 97 12UT, made from the analysis of 16 Feb. 00UT (the range is 84h) showing a possible Low 41. The low resolution system employed in real time is then asked ``where should we observe on 17 Feb. 18UT in order to improve specifically the following (42h) forecast of Low 41 ?'' in the verification area. The field shown is the pressure at the mean sea level, interval 3 mbar. The model employed is ARPEGE on a regular grid at resolution T63.
Zone where to observe and flight plan Figure G: The answer to the question asked with Fig. T1 is shown here by considering the isolines: they represent the 700 mbar temperature perturbation of the most unstable singular vector that can develop between 17 Feb. 18UT and 19 Feb. 12UT (the amplitude is arbitrary; negative contours are dashed). A small error on the amplitude of this 3D perturbation will amplify as rapidly as this perturbation and is likely to wreck the forecast. A flight plan designed to collect extra data where this structure has a maximum amplitude at that time is superimposed, the dots showing where dropsondes were to be launched. This flight plan was proposed before the Gulfstream flew to StJohn's, but had to be confirmed several times afterwards, using more recent forecasts.
Impact of adapted observations

Figure H: The dropsonde data obtained during the 17 Feb. flight has been used operationnaly as well as after the field phase, in a better controlled environment. The figure shows the impact of this dropsonde data (shaded area, interval 2 mbar) on a high resolution forecast derived from a suite run without any of the special FASTEX data, except this particular flight. The impact in this case is to change the amplitude towards a better, less deep, value.

The groups that provided methods, computer programs for finding where observations should be taken and that contributed to fund flight time of the aircraft were NOAA/NCEP associated with Penn State university, NOAA/ETL and the Naval Research Laboratory, all in the USA and Météo-France and CNRS in France.

The figures on this page are by T. Bergot, from Météo-France.

To previous To next

|quick tour home|scientific objectives|field experiment|adaptive observation|data and the Data Base|

|home|atlas|project info|data info|data selection|

fastex-dba@cnrm.meteo.fr
for questions or comments

The FASTEX Home site and data base are hosted byTo Meteo-France site

Project info pages: all rights reserved on figures, images and texts.© 1997-1999 by Météo-France (A. Joly),