The 2020 Global Operational NWP Data Assimilation System at Météo-France

Bouyssel, F.; Berre, Loïk; Bénichou, H.; Chambon, Philippe; Girardot, N; Guidard, Vincent; Loo, Cécile; Mahfouf, J.‐F.; Moll, Patrick; Payan, Christophe; Raspaud, Dominique

doi:10.1007/978-3-030-77722-7_25

Cited by 18 publications

(16 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For more details, Bouyssel et al . (2022) (their table 1) summarize the various observing systems assimilated in ARPEGE (excluding Aeolus) in 2020.…”

Section: Preliminary Resultsmentioning

confidence: 99%

“…• 4D-Var data assimilation (DA) system (6 hr window subdivided into 30 min observation time-slots) with an incremental formulation (Bouyssel et al, 2022):…”

Section: Experimental Designmentioning

confidence: 99%

“…The chosen configuration of the high‐resolution NWP model ARPEGE (Action des Recherche Petit Echelle et Grande Echelle) is based on a version that became operational in July 2019. Its main features are described below:Spectral model with a variable horizontal resolution T L 1798 (spectral truncation) C2.2 (stretch coefficient) – 5 km over France and 25 km over New Zealand – (Figure 1) 105 vertical levels (from 10 m above model topography to 0.1 hPa) 4D‐Var data assimilation (DA) system (6 hr window subdivided into 30 min observation time‐slots) with an incremental formulation (Bouyssel et al ., 2022):Two inner‐loops of minimization: the first one at T L 224C1, 105 levels (40 iterations) and the second one at T L 499C1, 105 levels (40 iterations) Background error variances and correlation lengths taken from an Ensemble of Data Assimilation (EDA) system (4D‐Var at lower resolution: T L 499/T L 224) with 50 members (AEARP: EDA for ARPEGE) …”

Section: Description Of the Observing System Experimentsmentioning

confidence: 99%

“…4D‐Var data assimilation (DA) system (6 hr window subdivided into 30 min observation time‐slots) with an incremental formulation (Bouyssel et al ., 2022):Two inner‐loops of minimization: the first one at T L 224C1, 105 levels (40 iterations) and the second one at T L 499C1, 105 levels (40 iterations) Background error variances and correlation lengths taken from an Ensemble of Data Assimilation (EDA) system (4D‐Var at lower resolution: T L 499/T L 224) with 50 members (AEARP: EDA for ARPEGE) …”

Section: Description Of the Observing System Experimentsmentioning

confidence: 99%

See 3 more Smart Citations

Operational assimilation of Aeolus winds in the Météo‐France global NWP model ARPEGE

Pourret

Šavli

Mahfouf

et al. 2022

Quart J Royal Meteoro Soc

View full text Add to dashboard Cite

The European Space Agency's Aeolus satellite was launched in August 2018. Measurements of wind profiles are provided for the first time from space using an onboard Doppler wind lidar. The quality of Aeolus Level‐2B (L2B) wind products has been found suitable for data assimilation in the Météo‐France global model ARPEGE since April 2020, in particular, when applying a suitable bias correction method. This article describes a series of Observing System Experiments (OSEs) conducted in April–May 2020 to assess the impact of Aeolus horizontal line‐of‐sight winds (HLOSW) on Météo‐France's global numerical weather prediction analyses and forecasts. Innovation statistics and a posteriori diagnostics from a period of July–August 2019 were used to scale the random observation errors provided by the L2B processor (mostly for Rayleigh‐clear winds). Although the HLOSW data represent only 0.42% of the total amount of all observations assimilated in ARPEGE, their contribution to the reduction of the global analysis variance is up to 2.3% (measured by the Degree of Freedom for Signal). The assimilation of HLOSW showed improvement in 6 hr short‐range forecasts which is demonstrated by an overall reduction of innovations statistics for various operational observing systems. From a Forecast Sensitivity to Observations impact (FSOi) study Aeolus is found to be the third most effective observing system (per individual observation) at reducing global 24‐hour forecast errors. For longer forecast ranges, the largest positive impacts are noticed over the tropics, particularly in the lower stratosphere up to 102 hr ahead (with up to 2% root‐mean‐square error reduction for wind and temperature), but also in the troposphere up to 72 hr ahead. To a lesser extent, a similar improvement is observed over the Southern Hemisphere. This positive impact of Aeolus HLOSW in OSEs has led to their operational assimilation at Météo‐France starting in June 2020.

show abstract

“…For more details, Bouyssel et al . (2022) (their table 1) summarize the various observing systems assimilated in ARPEGE (excluding Aeolus) in 2020.…”

Section: Preliminary Resultsmentioning

confidence: 99%

“…• 4D-Var data assimilation (DA) system (6 hr window subdivided into 30 min observation time-slots) with an incremental formulation (Bouyssel et al, 2022):…”

Section: Experimental Designmentioning

confidence: 99%

Section: Description Of the Observing System Experimentsmentioning

confidence: 99%

Section: Description Of the Observing System Experimentsmentioning

confidence: 99%

See 2 more Smart Citations

Operational assimilation of Aeolus winds in the Météo‐France global NWP model ARPEGE

Pourret

Šavli

Mahfouf

et al. 2022

Quart J Royal Meteoro Soc

View full text Add to dashboard Cite

show abstract

“…Model dynamics are non-hydrostatic, semi-implicit, and semi-Lagrangian. Lateral boundary conditions (LBCs) are extracted from the global model ARPEGE (Bouyssel et al, 2022). AROME-France features a Davies relaxation (Davies, 1976) coupling and ARPEGE synchronization.…”

mentioning

confidence: 99%

Assimilation of Meteosat Third Generation (MTG) Lightning Imager (LI) observations in AROME-France – Proof of Concept

Erdmann

Caumont²,

Defer³

2022

Preprint

View full text Add to dashboard Cite

Abstract. This study develops a Lightning Data Assimilation (LDA) scheme for the regional, convection-permitting NWP model AROME-France. The LDA scheme intends to assimilate total lightning, i.e., cloud-to-ground (CG) and inter- and intra-cloud (IC), of the future Meteosat Third Generation (MTG) Lightning Imager (LI). MTG-LI proxy data are created and Flash Extent Density (FED) fields are derived. An FED forward observation operator (FFO) is trained based on modeled, column integrated graupel mass from 24 storm days in 2018. The FFO is successfully verified for 2 independent storm days. With the FFO, the LDA adapts a 1-dimensional Bayesian (1DBay) retrieval followed by a 3-dimensional variational (3DVar) assimilation approach that is currently run operationally in AROME-France for radar reflectivity data. The 1DBay retrieval derives relative humidity profiles from the background by comparing the FED observations to the FED inferred from the background. Retrieved relative humidity profiles are assimilated as sounding data. The evaluation of the LDA comprises different LDA experiments and four case studies. It is found that all LDA experiments can increase the background integrated water vapor (IWV) in regions where the observed FED exceeds the FED inferred from AROME-France outputs. In addition, IWV can be reduced where spurious FED is modeled. A qualitative analysis of 6-hour accumulated rainfall fields reveals that the LDA is capable of locating and initiating some local precipitation fields better than a radar data assimilation (RDA) experiment. However, the LDA also leads to too high rainfall accumulations at some locations. Fractions Skill Scores (FSSs) of 6-hour accumulated rainfall are overall similar for the developed LDA and RDA experiments. An approach aiming at mitigating effects due to differences in the optical extents of lightning flashes and the area of the corresponding cloud was developed and included in the LDA, however, it does not always improve the FSS.

show abstract

Simulation of analysis error contributions arising from uncertainties in the parametrisation of deep convection

Hubans

Berre

Bouteloup

et al. 2022

Quart J Royal Meteoro Soc

View full text Add to dashboard Cite

In order to simulate analysis error contributions resulting from observation errors and model errors, a possible methodology is to run a data assimilation ensemble that includes observation perturbations and a multiphysics approach. This method is examined here formally and experimentally, in order to investigate the effects of deep convection parametrisation uncertainties on analysis errors. A linearised expansion of initial condition errors is first employed to decompose them into contributions arising from observation errors and model errors. This is compared formally with a similar expansion for analysis differences resulting from the use of two different model versions, as in a multiphysics approach. Experimental results with the Action de Recherche Petite Echelle Grande Echelle global model are then examined. Observation error contributions are simulated by analysis differences between an unperturbed experiment and another experiment that includes observation perturbations that are drawn from specified observation error covariances. This is compared with analysis differences resulting from the use of two different parametrisation schemes of deep convection. The results indicate that the contribution of convection uncertainties is relatively large in the Tropics, with similar amplitudes to the contribution of observation errors. Simulated model errors are also found to be pronounced in the boundary layer and in convective areas where the analysis constraint by observations is relatively small, due to lower observation density.

show abstract

The 2020 Global Operational NWP Data Assimilation System at Météo-France

Cited by 18 publications

References 45 publications

Operational assimilation of Aeolus winds in the Météo‐France global NWP model ARPEGE

Operational assimilation of Aeolus winds in the Météo‐France global NWP model ARPEGE

Assimilation of Meteosat Third Generation (MTG) Lightning Imager (LI) observations in AROME-France – Proof of Concept

Simulation of analysis error contributions arising from uncertainties in the parametrisation of deep convection

Contact Info

Product

Resources

About