Innovating in the collection
and use of data

Upgrading and expanding the observation system

Le radar mis en service sur les hauteurs du golfe d’Ajaccio, le 14 décembre 2017. © Météo-France.
Radar in operation on the heights of the Gulf of Ajaccio, 14th December 2017. © Météo-France.


Maintaining an observation network in an operational condition is accompanied by multi-annual system regeneration programmes in order to anticipate programmed obsolescence and reduce the cost of ownership by streamlining systems. Renewing surface networks continued in 2017 with the deployment of 140 automatic Mercury ground measurement stations and gradual replacement of mechanical anemometers with ultrasonic wind sensors. This initiative will continue into 2018 and will be complemented by the onset of automation of the Réseau climatologique d’État (RCE - French Climatological Network). Automation of the RCE consists of replacing daily readings carried out by volunteer operators with automatic stations recording temperature and rainfall.

In expansion for over thirty years, the radar network has reached its target size with 40 devices operating in mainland France and overseas territories. Two new radar sites were created in France in 2017, one in Saint-Rémy (X-band) and the other in Ajaccio (C-band), completing the network. A regeneration programme for the radar network is committed to replacing 1 to 2 radars per year. In 2017, the radar in Nancy was replaced by a C-band dual polarisation radar by the company Selex.

Recent automation of the altitude network is yet to reach its target size. Two new automatic radiosonde systems will be rolled out on the island of Réunion in the first quarter of 2018, and in Tahiti in the second quarter. An aerosol lidar network was rolled out in 2016. A regeneration programme for the altitude network programme should therefore begin by 2020 to replace the first helium robot-probes with hydrogen robot-probes in order to reduce ownership costs.

New radar data processing software

Traitement des données radar lors d’un épisode neigeux : avec l’ancien logiciel SYCOMORE (à gauche) ; avec le nouveau logiciel SERVAL (à droite). © Météo-France
Radar data processing during an episode of snow: using the old SYCOMORE software (left); and the new SERVAL software (right). © Météo-France.

Meteorological radars localise precipitation, measure its intensity in real-time and provide information on its nature (rain, snow, hail, etc.). Météo-France’s new radar data processing software, known as SERVAL (French abbreviation of ‘radar services and centralised visualisation development software’) has been in operation since late 2017. The device facilitates a major step forward in the development of radar products and opens the door to many possibilities. It centralises all raw data from radars in operation across France at the Weather Forecast Centre in Toulouse. Additionally, the software’s architecture has been simplified around a central processing chain and offers the possibility of using raw data available on a simultaneous and coherent basis. The new architecture makes it easy to take into account data from new radars (1 to 2 per year) which will be rolled out within the network regeneration programme framework, and to significantly improve the quality of current mosaic services (composite images). This improvement is of particular significance for the characterisation of low intensity rain, drizzle and snow events. By concentrating all raw radar data within a single process, the SERVAL system significantly improves 3D rainfall descriptions. The system’s 3D descriptions will enable development of high-quality and high-resolution services, such as altitudes of precipitation nuclei, by 2020.

Météo-France’s mobile application becomes collaborative

Configuration de l'application avec accès aux données d'observation saisies par les autres mobinautes. On note, sur cette capture d'écran du 16 octobre 2017, des pictos « poussières » sur la Bretagne, conséquence des incendies de forêt d'Espagne et du Portugal, ainsi que du sable du Sahara, apporté par le flux de Sud. © Météo-France.
Configuration of the application with access to observational data entered by mobile users. Note that there are a number of ‘dusty’ pictograms across Brittany on this screenshot from 16th October 2017, the consequence of forest fires in Spain and the Portugal and sand from the Sahara, brought by South flowing air. © Météo-France.


Automation of the observation network and the appearance of new sensors have reduced the percentage of human observations within weather data collection. However, the latter forms a crucial element, as it facilitates the real-time collection of valuable information at high spatial resolutions for the validation and use of ‘real-time’ models and systems output, as well as for the monitoring of prominent weather events and their impacts, particularly in the case of vigilance. As such, Météo-France has adopted an approach which aims to develop participatory observation via various initiatives. In this context, Météo-France’s mobile application supports crowdsourcing by integrating a module enabling users to enter real-time observations.

Its roll-out was gradual, with 10,000 testers initially recruited for the beta version before production began on a version that satisfied both the following expectations of the general public and the various needs of the agency:

  • feed back a higher number of observations from mainland France and overseas territories to forecasters;
  • assist changing strategy concerning volunteer observers;
  • offer citizens the opportunity to contribute to one of the first mobile consultation applications, with an average of over 350 requests per second.

This operation was crowned with success, with over 30,000 time-sensitive observations in a single day, fed back by mobile users for the most disrupted weather situations.

For more information: Devenez acteur de la météo.

Discover the other chapters of the current part