Richer data, better
performing models

Collecting and using all-terrain data

Using microwave radiometer data in the AROME model: case study in an Alpine valley

Série temporelle de profils de température observés par radiosondage (a), issus des prévisions 1 heure du modèle AROME (b), obtenus après assimilation 1D des observations de radiomètre micro-onde (c), lors de la campagne Passy-2015 du 6 au 20/02/2017. © Météo-France.
Temporal series of temperature profiles observed by radiosonde (a), from 1-hour forecasts from the AROME model (b), obtained after 1D assimilation of microwave radiometer observations (c), during the Passy-2015 campaign from 6th to 20th February 2017. © Météo-France.

The town of Passy in the Arve Valley, Haute-Savoie, France, is regularly affected by peaks of intense pollution. These occur as a result of anticyclonic winter conditions, during which weather forecasts are often uncertain.

In order to improve understanding and forecasting of these events, Météo-France launched the Passy-2015 measurements campaign in 2015 and installed a microwave radiometer on the ground. It takes continuous measurements of temperature and humidity profiles as well as content integrated into water from the atmosphere. Its unrivalled temporal resolution (of around several minutes) enables evolution of the boundary layer’s life cycle to be monitored; in other words, the lower part of the atmosphere located just above the Earth's surface, in order to incorporate this data into the numerical prediction model, AROME. The initial state of the atmosphere taken into account by the model is therefore closer to reality, which ultimately improves the forecasts it provides.

Work conducted during the Passy-2015 campaign highlighted the limitations of the AROME model which largely underestimated cooling that occurs near the surface, with forecasting errors reaching 12°C during peaks of pollution. By combining microwave radiometer observations with the model’s 1-hour forecasts, the error observed in the model’s initial state could be greatly reduced, both on the surface (going from 8°C to 0.5°C) and at the base of stratus clouds. Beyond the initial state, the temporal evolution of this boundary layer is also far better represented in the model thanks to its consideration of these high-frequency observations.

This work led to studies on operational assimilation of the data in the AROME model. It should bring about improvement in the forecasting of a number of events that pose significant economic and social issues: pollution, but also fog and intense rainfall events.

PIRATA network: 20 years of data from the tropical Atlantic

L'ancienne bouée PIRATA est saisie sur le pont du Thalassa. © IRD, Bernard Bourlès.
The old PIRATA buoy hoisted onto the Thalassa bridge. © IRD, Bernard Bourlès.

The tropical Atlantic Ocean plays a key role in weather forecasting, seasonal forecasting and climate studies in the Americas, Africa and Europe. French, Brazilian and American oceanographers and meteorologists have been developing and operating a network of 18 anchored buoys for 20 years, in order to study and monitor the basin. Known as PIRATA (Prediction and Research Moored Array in the Tropical Atlantic), this device, along with campaigns at sea to keep it maintained, provides high quality data on air and depth temperatures, atmospheric pressure, salinity, and even the partial pressure of carbon dioxide. The latter feed atmosphere and ocean prediction models in real-time and are used by researchers to study the ocean’s whims and climate change.

As the result of international cooperation, PIRATA responds to both scientific and operational objectives, federating research organizations; Institut de recherche pour le développement (IRD - French Institute of research for development) and the Brazilian Instituto Nacional de Pesquisas (INPE - National Institute for Space Research), with meteorological services; the American National Oceanographic and Atmospheric Administration (NOAA), and Météo-France. Observing long-term variability of the tropical Atlantic is essential. This area is central to many research requirements, but also to very practical weather and extreme event forecasting applications such as African monsoons, droughts in the Northeast of Brazil, hurricanes. As such, PIRATA buoys have recorded and transmitted valuable information on Hurricane Irma, which enabled American and French services to refine their forecasts and issue tailored warnings to save Caribbean populations.

Better forecasting and alerting in overseas territories

New coastal wave model for Réunion and Mayotte

Hauteur significative des vagues de la mer totale (m) de WW3 le 31 janvier 2013 à 18 h UTC, lors du passage de l’ouragan Felleng (catégorie 3) à La Réunion. Les flèches violettes foncées représentent la direction de la houle primaire, les claires la houle secondaire et les noires la mer de vent.
Significant wave height of the total sea area (m) in WW3 on 31st January 2013 at 18:00 UTC, when Hurricane Felleng (category 3) struck Réunion. The dark purple arrows represent direction of the primary sea swell, the light arrows represent secondary sea swell and the black arrows represent sea wind. © Météo-France

As of late 2015, the second phase of the HOMONIM project (History, observation, modelling of sea levels), led by Météo-France and the Service hydrographique et océanographique de la marine (SHOM - French Naval Hydrographic and Oceanographic Service), was launched with the support of the Ministry for the Ecological and Inclusive Transition (MTES) to improve surge and wave forecasts on the coast, notably in overseas territories. After being implemented on the Guyanese and Caribbean coasts in 2016, a high-resolution configuration of the wave model, WaveWatch 3 (WW3) using an irregular grid, was deployed in Réunion and Mayotte in 2017.

This configuration benefitted from coastal bathymetry modelling at 100 m, recently produced by the SHOM. Several parametrisations describing sea conditions and interaction with the sea bed have been tested on prominent events to have occurred in recent years. These episodes largely correspond to hurricanes, such as Hurricane Felleng (category 3) in January 2013 (see illustration).

One area covers the islands of Réunion and Mauritius and the other covers the Comoro Islands. These configurations are forced by winds from the AROME Outre-Mer atmospheric model, at a resolution of 2.5 km, and are interlinked in Météo-France’s regional wave model MFWAM, at a resolution of 10 km. WW3’s resolution reached 100 m in the lagoon of Mayotte, enabling clear representation of channels and bathymetric details, and 200 m from the coasts of Réunion.

The lagoon of Mayotte’s specific context requires forcing to be tested using a water level and currents model (tide and rises in sea level). This forcing is due to be implemented by mid-2018 at a resolution of 200 m. In addition, these configurations will be replayed on new episodes documented in order to perfect their validation and/or calibration.

Extension of APICs in overseas territories

The Intense rainfall warning system at local authority level (APIC) provided by Météo-France, informs municipalities when observed rainfall reaches exceptional levels in their municipality or surrounding municipalities. Implemented in 2011 in mainland France, the crisis management support system was rolled out in Réunion in 2016 and in Martinique and Guadeloupe in October 2017. It is expected to be rolled out in New Caledonia in 2018.

Data and modelling for the Copernicus programme

Wave modelling for Copernicus (global and regional)

Waves play an important role in movement, heat and gas exchanges between the ocean and the atmosphere. In the framework of Copernicus Navy services (Copernicus Marine Environment Monitoring Service or CMEMS) Mercator Ocean is responsible for project management and Météo-France (DirOP/MAR) for the provision of wave services across the globe using the MFWAM model at a resolution of 20 km. The global model is forced by CEP winds and assimilates height data from a number of satellites. Services are adapted according to 17 parameters describing sea conditions (notably wave height, average period and Stokes drift).

The MFWAM wave model is also used for the regional area of Iberian Biscay Ireland (IBI), at a resolution of 10 km within the framework of a collaboration with Puertos del Estados, AEMET, CESGA and Marine Institute. The CMEMS wave services were launched on 1st April 2017. In the upcoming version of the CMEMS-IBI operating system, V4, the MFWAM wave model will be coupled with the NEMO ocean circulation model (Nucleus for European Modelling of the Ocean).

Three coupling processes have been implemented: the first is the Coriolis-Stokes forcing term, which uses Stokes drift calculated by the wave model; the second process consists of taking into account stress of the flow of the amount of movement provided to the ocean; lastly, the third process takes into account turbulence related to wave surge pumped into the ocean mixed layer. Initial results from coupling between MFWAM and NEMO showed significant impact on key parameters, these being surface temperature, current components, salinity, and the ocean mixed layer.

Copernicus Atmosphere: an European air quality service for citizens and decision-makers



Copernicus Atmosphere (CAMS) aims to develop and make available useful environmental monitoring elements to all European citizens. The regional CAMS service, led by Météo-France and the INERIS (French National Institute for Industrial Environment and Risks), provides up to 4-day forecasting, and analysis and reanalysis of air concentration for approximately 10 pollutants across Europe on an operational basis. It includes concentration forecasts for various types of pollen. Today, these forecasts are developed by combining 7 Chemistry-Transport models, including Météo-France’s MOCAGE model and INERIS’ CHIMERE model, as well as models developed by the best air quality specialists in Europe.

A set-theoretic service has also been developed using these models. All services are open-access and are intended to be reused in expert or downstream systems. The Copernicus Atmosphere regional service currently offers an operational service based on the latest research developments, offering a high level of availability. In France, the service complements the country’s national platform Prev'air as well as AASQA local platforms (Associations Agréées pour la Surveillance de la Qualité de l'Air - Officially Approved Associations for Air Quality Monitoring). The CAMS project is due to be further enriched in the future, using satellite data from Sentinel programmes.

CAMS notably provides forecasts for the atmospheric concentration of pollen. Birch pollen forecasts were implemented in Europe in 2013, before being extended to include olive tree and grass pollen in 2016. These species are recognised by the WHO as some of the most allergenic, with monitoring and forecasting their concentration classed as a strategic public health issue.

Pollen emission flows are calculated by cross referencing weather data (wind, humidity, and precipitation charts) with information relating to the biological cycles of pollen-emitting plants. ‘Heat’ accumulation from the start of the year is also taken into account for olive trees. Parameterisation of grasses is based on emission probability surrounding pollination dates derived from past observations.

The inclusion of ragweed pollen, which has increasing impacts on health, is under development within the service and forecasting is expected to go live in the Copernicus service in 2018.

For more information, visit: COPERNICUS.

Discover the other chapters of the current part