Search

Cyclones are essential elements of the climate and of the water cycle in the Mediterranean. The most intense of them lead to natural disasters because of their violent winds and extreme rainfall, which can cause significant damage to the territories bordering the Mediterranean (coast and mountain ranges). Reliable forecasts of cyclones are therefore essential to better anticipate and prevent their societal impact. However, their predictability is often limited by their particularities: smaller cyclones with a shorter life cycle than in the North Atlantic, complex topography, interactions with the relatively warm sea and air masses laden with dust from the Sahara.We investigate the predictability of Mediterranean cyclones in a systematic framework using an ensemble prediction system. A reference dataset was first obtained by tracking cyclones in the ERA5 reanalysis (1979-2021), using an algorithm developped for the North Atlantic and adapted for the Mediterranean region. We then investigated the predictability using ARPEGE ensemble reforecasts in a homogeneous configuration over 22 years (2000-2021).We restricted the study on 500 cases, which were selected using a storm severity index based on wind gusts and adapted for the Mediterranean region. The cases were then divided in several categories following their dynamical context, their intensity and their geographical origin. The predictability of the reforecasts was finally quantified on each of those categories, using probabilistic scores on cyclone trajectories (along and cross track error) and on intensities (mean sea level pressure and storm severity index).While past studies have been limited by the fact that regular updates of operational forecasting systems do not allow the predictability of cases to be compared with each other, the homogeneous configuration of the ARPEGE ensemble reforecasts makes it possible to systematically identify the limitation to the predictability of Mediterranean cyclones.

Numerical atmospheric dispersion models (ADMs) are used for predicting the health and environmental consequences of nuclear accidents in order to anticipate countermeasures necessary to protect the populations. However, these simulations suffer from significant uncertainties, arising in particular from input data: weather conditions and source term. Meteorological ensembles are already used operationally to characterize uncertainties in weather predictions. Combined with dispersion models, these ensembles produce different scenarios of radionuclide dispersion, called “members”, representative of the variety of possible forecasts. In this study, the fine-scale operational weather ensemble AROME-EPS (Applications of Research to Operations at Mesoscale-Ensemble Prediction System) from Météo-France is coupled with the Gaussian puff model pX developed by the IRSN (French Institute for Radiation Protection and Nuclear Safety). The source term data are provided at 10 min resolution by the Orano La Hague reprocessing plant (RP) that regularly discharges 85Kr during the spent nuclear fuel reprocessing process. In addition, a continuous measurement campaign of 85Kr air concentration was recently conducted by the Laboratory of Radioecology in Cherbourg (LRC) of the IRSN, within 20 km of the RP in the North-Cotentin peninsula, and is used for model evaluation. This paper presents a probabilistic approach to study the meteorological uncertainties in dispersion simulations at local and medium distances (2–20 km). First, the quality of AROME-EPS forecasts is confirmed by comparison with observations from both Météo-France and the IRSN. Then, the probabilistic performance of the atmospheric dispersion simulations was evaluated by comparison to the 85Kr measurements carried out during a period of 2 months, using two probabilistic scores: relative operating characteristic (ROC) curves and Peirce skill score (PSS). The sensitivity of dispersion results to the method used for the calculation of atmospheric stability and associated Gaussian dispersion standard deviations is also discussed. A desirable feature for a model used in emergency response is the ability to correctly predict exceedance of a given value (for instance, a dose guide level). When using an ensemble of simulations, the “decision threshold” is the number of members predicting an event above which this event should be considered probable. In the case of the 16-member dispersion ensemble used here, the optimal decision threshold was found to be 3 members, above which the ensemble better predicts the observed peaks than the deterministic simulation. These results highlight the added value of ensemble forecasts compared to a single deterministic one and their potential interest in the decision process during crisis situations.

Bow echoes (BEs) are bow-shaped lines of convective cells that are often associated with swaths of damaging straight-line winds and small tornadoes. This paper describes a convolutional neural network (CNN) able to detect BEs directly from French kilometer-scale model outputs in order to facilitate and accelerate the operational forecasting of BEs. The detections are only based on the maximum pseudoreflectivity field predictor (“pseudo” because it is expressed in mm h−1 and not in dBZ). A preprocessing of the training database is carried out in order to reduce imbalance issues between the two classes (inside or outside bow echoes). A CNN sensitivity analysis against a set of hyperparameters is done. The selected CNN configuration has a hit rate of 86% and a false alarm rate of 39%. The strengths and weaknesses of this CNN are then emphasized with an object-oriented evaluation. The BE largest pseudoreflectivities are correctly detected by the CNN, which tends to underestimate the size of BEs. Detected BE objects have wind gusts similar to the hand-labeled BE. Most of the time, false alarm objects and missed objects are rather small (e.g., 2). Based on a cooperation with forecasters, synthesis plots are proposed that summarize the BE detections in French kilometer-scale models. A subjective evaluation of the CNN performances is also reported. The overall positive feedback from forecasters is in good agreement with the object-oriented evaluation. Forecasters perceive these products as relevant and potentially useful to handle the large amount of available data from numerical weather prediction models.

The use of ensemble prediction systems (EPS) is challenging because of the huge information it provides. Forecasts from ensemble prediction systems (EPS) are often summarised by statistical quantities (ie quantiles maps). Although such mathematical representation is efficient for capturing the ensemble distribution, it lacks physical consistency, which raises issues for many applications of EPS in an operational context. In order to provide a physically-consistent synthesis of the French convection-permitting AROME-EPS forecasts, we propose to automatically draw a few scenarios that are representative of the different possible outcomes. Each scenario is a reduced set of EPS members.To design a scenario synthesis, the procedure can be divided into two parts. A first step aims at extracting relevant features in each EPS member in order to reduce the problem dimensionality. Then, a clustering is done based on these features.The originality of our work is to leverage the capacities of deep learning for the features extraction. For that purpose, we use a convolutional autoencodeur (CAE) to learn an optimal low-dimensional representation (also called latent space representation) of the input forecast field. In this work, the algorithm is developed to work on 1h-accumulated rainfall from AROME-EPS, with a focus on convective cases.The CAE is trained on around 150 000 forecasts and its performance is evaluated based on the quality of the reconstructed input fields from the latent space. To examine the reconstruction quality, an object-oriented approach is used. CAE is also compared with the commonly-used principal component analysis (PCA). In a second part, different clustering methods (kmeans, HDBSCAN, …) are applied to EPS members in the latent space and evaluated using subjective and objective diagnostics.

High quality volcanic ash forecasts are crucial to minimize the economic impact of volcanic hazards on air traffic. Decision-making is usually based on numerical dispersion modeling with only one model realization. Given the inherent uncertainty of such approach, a multi-model multi-source term ensemble has been designed and evaluated for the Eyjafjallajökull eruption in May 2010. Its use for air traffic management is discussed. Two multi-model ensembles were built: the first is based on the output of four dispersion models and their own implementation of ash ejection. All a priori model source terms were constrained by observational evidence of the volcanic ash cloud top as a function of time. The second ensemble is based on the same four dispersion models, which were run with three additional source terms: (i) a source term obtained with background modeling constrained with satellite data (a posteriori source term), (ii) its lower bound estimate, and (iii) its upper bound estimate. The a priori ensemble gives valuable information about the probability of ash dispersion during the early phase of the eruption, when observational evidence is limited. However, its evaluation with observational data reveals lower quality compared to the second ensemble. While the second ensemble ash column load and ash horizontal location compare well to satellite observations, 3D ash concentrations are negatively biased. This might be caused by the vertical distribution of ash, which is too much diluted in all model runs, probably due to defaults in the a posteriori source term and vertical transport and/or diffusion processes in all models. Relevant products for the air traffic management are horizontal maps of ash concentration quantiles (median, 75 %, 99 %) at a fine-resolved flight level grid. These maps can be used for route optimization in the areas where ash does not pose a direct and urgent threat to aviation. Cost-optimized consideration of such hazards will result in much less impact on flight cancellations, reroutings, and traffic flow congestions.

Numerical dispersion models are used operationally worldwide to mitigate the effect of volcanic ash on aviation. In order to improve the representation of the horizontal dispersion of ash plumes and of the 3D concentration of ash, a study was conducted using the MOCAGE model during the European Natural Airborne Disaster Information and Coordination System for Aviation (EUNADICS-AV) project. Source term modelling and assimilation of different data were investigated. A sensitivity study of source term formulation showed that a resolved source term, using the FPLUME plume rise model in MOCAGE, instead of a parameterised source term, induces a more realistic representation of the horizontal dispersion of the ash plume. The FPLUME simulation provides more concentrated and focused ash concentrations in the horizontal and the vertical dimensions than the other source term. The assimilation of Moderate Resolution Imaging Spectroradiometer (MODIS) aerosol optical depth has an impact on the horizontal dispersion of the plume, but this effect is rather low and local compared to source term improvement. More promising results are obtained with the continuous assimilation of ground-based lidar profiles, which improves the vertical distribution of ash and helps in reaching realistic values of ash concentrations. Using this configuration, the effect of assimilation may last for several hours and it may propagate several hundred kilometres downstream of the lidar profiles.

High-quality volcanic ash forecasts are crucial to minimize the economic impact of volcanic hazards on air traffic. Decision-making is usually based on numerical dispersion modelling with only one model realization. Given the inherent uncertainty of such an approach, a multi-model multi-source term ensemble has been designed and evaluated for the Eyjafjallajökull eruption in May 2010. Its use for flight planning is discussed. Two multi-model ensembles were built: the first is based on the output of four dispersion models and their own implementation of ash ejection. All a priori model source terms were constrained by observational evidence of the volcanic ash cloud top as a function of time. The second ensemble is based on the same four dispersion models, which were run with three additional source terms: (i) a source term obtained from a model background constrained with satellite data (a posteriori source term), (ii) its lower-bound estimate and (iii) its upper-bound estimate. The a priori ensemble gives valuable information about the probability of ash dispersion during the early phase of the eruption, when observational evidence is limited. However, its evaluation with observational data reveals lower quality compared to the second ensemble. While the second ensemble ash column load and ash horizontal location compare well to satellite observations, 3D ash concentrations are negatively biased. This might be caused by the vertical distribution of ash, which is too much diluted in all model runs, probably due to defaults in the a posteriori source term and vertical transport and/or diffusion processes in all models. Relevant products for the air traffic management are horizontal maps of ash concentration quantiles (median, 75 %, 99 %) at a finely resolved flight level grid as well as cross sections. These maps enable cost-optimized consideration of volcanic hazards and could result in much fewer flight cancellations, reroutings and traffic flow congestions. In addition, they could be used for route optimization in the areas where ash does not pose a direct and urgent threat to aviation, including the aspect of aeroplane maintenance.

Volcanic eruptions comprise one of the most important airborne hazards for aviation. Although significant events are rare, they have a very high impact. The current state of tools and abilities to mitigate aviation hazards associated with an assumed volcanic cloud was tested within an international demonstration exercise. Experts in the field assembled at the Schwarzenberg barracks in Salzburg, Austria, in order to simulate the sequence of procedures for the volcanic case scenario of an artificial eruption of Etna volcano in Italy. The scope of the exercise ranged from the detection of the assumed event to the issuance of early warnings. Volcanic emission concentration charts were generated applying modern ensemble techniques. The exercise products provided an important basis for decision making for aviation traffic management during a volcanic eruption crisis. By integrating the available wealth of data, observations and modelling results directly into a widely used flight planning software, it was demonstrated that route optimization measures could be implemented effectively. With timely and rather precise warnings available, the new tools and processes tested during the exercise demonstrated vividly that a vast majority of flights could be conducted despite a volcanic plume widely dispersed within a high-traffic airspace over Europe. The resulting number of flight cancellations was minimal.

Nitrogen fertilization is the main source of amonia (NH 3 ) emissions in France. Amonia is a major atmospheric pollutant impacting environment and human health, therefore its monitoring and forecasting is of main importance. Our research deals with improving the spatial and temporal variability of ammonia emissions from agriculture, in particular by exploring how meteorological constraints influence agricultural practices of fertilization. Within this context we developped a simple model based on air temperature, precipitations, and agricultural practices to estimate possible dates of fertilizations. This approach allows to reduce fertilization temporal windows thus giving the perspective to update the NH 3 inventories used as input by Chemistry-Transport Models (CTM).

Stratosphere-to-troposphere transport (STT) is an important natural source of tropospheric ozone, which can occasionally influence ground-level ozone concentrations relevant for air quality. Here, we analyse and evaluate the Copernicus Atmosphere Monitoring Service (CAMS) global and regional forecast systems during a deep STT event over Europe for the time period from 4 to 9 January 2017. The predominant synoptic condition is described by a deep upper level trough over eastern and central Europe, favouring the formation of tropopause folding events along the jet stream axis and therefore the intrusion of stratospheric ozone into the troposphere. Both global and regional CAMS forecast products reproduce the “hook-shaped” streamer of ozone-rich and dry air in the middle troposphere depicted from the observed satellite images of water vapour. The CAMS global model successfully reproduces the folding of the tropopause at various European sites, such as Trapani (Italy), where a deep folding down to 550 hPa is seen. The stratospheric ozone intrusions into the troposphere observed by WOUDC ozonesonde and IAGOS aircraft measurements are satisfactorily forecasted up to 3 days in advance by the CAMS global model in terms of both temporal and vertical features of ozone. The fractional gross error (FGE) of CAMS ozone day 1 forecast between 300 and 500 hPa is 0.13 over Prague, while over Frankfurt it is 0.04 and 0.19, highlighting the contribution of data assimilation, which in most cases improves the model performance. Finally, the meteorological and chemical forcing of CAMS global forecast system in the CAMS regional forecast systems is found to be beneficial for predicting the enhanced ozone concentrations in the middle troposphere during a deep STT event.

Volcanic eruptions emit significant amounts of aerosols in the atmosphere and have strong impact on aviation security and air traffic. Our research deals with spatial and temporal improvements of volcanic ash aerosols within the chemistry-transport model MOCAGE [1], [2], the operational air-quality and fast-response model of Meteo-France. To better characterize the source term of volcanic eruptions, we implemented the plume rise model FPLUME [3] and compared it to the Mastin [4] empirical parametrization, which confirmed that uncertainties of the volcanic ash modelling are considerable and very sensitive to the source term. For this reason, we explore the potential of assimilating spaceborne, ground-based and in-situ observations. Assimilation of Aerosols Optical Depth (AOD) from MODIS instrument allowed to correct the extent and intensity of the plume of the 2010 Eyjafjallajokull eruption. Further experiments are ongoing to take advantage of LIDAR soundings from EARLINET network and CALIOP instrument, as well as in-situ airborne measurements.

The role of an active phase of the Madden–Julian Oscillation (MJO) on the evolution of a mesoscale convective systems (MCS) leading to a tropical depression is investigated in the South-West Indian Ocean during the Dynamics of the Madden–Julian Oscillation (DYNAMO) field experiment, with a numerical limited-area atmospheric model. A mesoscale vortex is followed in the low-troposphere from the initiation of the active MJO phase. It is shown that the interaction of the vortex with the Equatorial jet associated with the MJO plays an important role on the vortex development. As the vortex encounters the southern part of the low-level jet, it undergoes intensification that is explained by the barotropic conversion of kinetic energy from the low-level jet to the vortex.

The link between El Nino Southern Oscillation (ENSO) and tropical cyclone activity over the Southwest Indian Ocean [(0°–30°S) × (48°–90°E)] is revisited. Based on the records for 1981–2013, we show that cyclonic activity decreases dramatically during strong ENSO events: only a few cases are observed during the January-March period. This pattern is attributed to changes in large-scale thermodynamic parameters. Although El Nino is correlated with an increase in tropospheric moisture and sea-surface temperatures, it is also associated with an equatorward shift of the subtropical jet. When El Nino is sufficiently strong, the latter effect may dominate the former, inhibiting tropical cyclone activity.

International audience

Tropospheric NO2 is hazardous to human health and can lead to tropospheric ozone formation, eutrophication of ecosystems and acid rain production. It is therefore important to establish accurate data based on models and observations to understand and monitor tropospheric NO2 concentrations on a regional and global scale. In the present study, MAX-DOAS tropospheric NO2 column retrievals from four European measurement stations are compared to regional model ensemble simulations. The latter are based on regional air quality models which contribute to the European regional ensemble forecasts and reanalyses of the operational Copernicus Atmosphere Monitoring Service (CAMS). Compared to other observational data usually applied for regional model validation, MAX-DOAS data is closer to the regional model data in terms of horizontal and vertical resolution and measurements are available during daylight. In general, there is a good agreement between simulated and retrieved NO2 column values for individual MAX-DOAS measurements with correlations between 45 and 75 % for tropospheric NO2 VCDs, indicating that the model ensemble represents the emission and tropospheric chemistry of NOx (NO + NO2) well. Pollution transport towards the stations is on average well represented by the models. However, large differences are found for individual pollution plumes. Seasonal cycles are overestimated, weekly cycles are reproduced well and diurnal cycles poorly represented by the model ensemble. In particular, simulated morning rush hour peaks are not confirmed by MAX-DOAS retrievals. Our results demonstrate that a large number of validation points are available from MAX-DOAS measurements, which should therefore be used more extensively in future regional air quality modelling studies.

A 6-models strong European ensemble of Copernicus Atmospheric Monitoring Service (CAMS) was run through the season of 2014 computing the olive pollen dispersion in Europe. The simulations have been compared with observations in 6 countries, members of the European Aeroallergen Network. Analysis was performed for individual models, the ensemble mean and median, and for a dynamically optimized combination of the ensemble members obtained via fusion of the model predictions with observations. The models, generally reproducing the olive season of 2014, showed noticeable deviations from both observations and each other. In particular, the season start was reported too early, by 8 days but for some models the error mounted to almost two weeks. For the season end, the disagreement between the models and the observations varied from a nearly perfect match up to two weeks too late. A series of sensitivity studies performed to understand the origin of the disagreements revealed crucial role of ambient temperature, especially systematic biases in its representation by meteorological models. A simple correction to the heat sum threshold eliminated the season shift but its validity in other years remains to be checked. The short-term features of the concentration time series were reproduced better suggesting that the precipitation events and cold/warm spells, as well as the large-scale transport were represented rather well. Ensemble averaging led to more robust results. The best skill scores were obtained with data fusion, which used the previous-days observations to identify the optimal weighting coefficients of the individual model forecasts. Such combinations were tested for the forecasting period up to 4 days and shown to remain nearly optimal throughout the whole period.

This study is part of the efforts undertaken to resolve the “bad trough/good trough” issue for tropical cyclone (TC) intensity changes and to improve the prediction of these challenging events. Sensitivity experiments are run at 8-km resolution with vortex bogusing to extend the previous analysis of a real case of TC–trough interaction (Dora in 2007). The initial position and intensity of the TC are modified, leaving the trough unchanged to describe a realistic environment. Simulations are designed to analyze the sensitivity of TC prediction to both the variety of TC–trough configurations and the current uncertainty in model analysis of TC intensity and position. Results show that TC intensification under upper-level forcing is greater for stronger vortices. The timing and geometry of the interaction between the two cyclonic potential vorticity anomalies associated with the cutoff low and the TC also play a major role in storm intensification. The intensification rate increases when the TC (initially located 12° northwest of the trough) is displaced 1° closer. By allowing a gradual deformation and equatorward tilting of the trough, both scenarios foster an extended “inflow channel” of cyclonic vorticity at midlevels toward the vortex inner core. Conversely, unfavorable interaction is found for vortices displaced 3° or 4° east or northeast. Variations in environmental forcing relative to the reference simulation illustrate that the relationship between intensity change and the 850–200-hPa wind shear is not systematic and that the 200-hPa divergence, 335–350-K mean potential vorticity, or 200-hPa relative eddy momentum fluxes may be better predictors of TC intensification during TC–trough interactions.

disponible à l'adresse : http://ams.allenpress.com/perlserv/?request=get-abstract&doi=10.1175%2F2008JAS2618.1; International audience; Midlatitude cyclogenesis as interpreted in the framework of either baroclinic development or potential vorticity thinking heavily relies on the concept of synoptic-scale anomaly. Given the existence of potential vorticity inversion and attribution, what is at stake to provide a mathematical definition for this concept is a complete finite-amplitude alternative to the linear-based theory of cyclogenesis. The existence of a reasonably objective way to represent anomalies in both real and idealized flows would not only help understanding cyclogenesis, it would also have many other applications for both theory and in practical forecasts. Inspired by the recent theory of wavelet representation of coherent structures in two-dimensional fluid mechanics, a wavelet representation of three-dimensional potential vorticity anomalies is built. This algorithm relies on the selection of the appropriate two-dimensional wavelet coefficients from the stationary wavelet transform in order to guarantee the critical translation-invariance property. The sensitivity of the algorithm to the position, size, and shape of the structures is assessed. The wavelet extraction is then applied to the upper-level precursor of a real-case storm of December 1999 and is compared to a basic monopolar extraction. Using potential vorticity inversion and forecasts with a primitive-equation model, it is found that both anomalies have similar implications on the development of the surface cyclone. However, the coherence in time of the extracted wavelet structure in the forecast and analysis sequence is more satisfactory than the extracted monopole: this suggests that the underlying mathematical description of an anomaly proposed here does, indeed, point toward the direction of an actual physical reality of the concept.

In the framework of global atmospheric composition and air quality forecasting, accurate modeling of atmospheric chemical species redistribution and evolutions in numerical weather and air quality prediction models has become an important challenge in terms of research and developments. Such model developments are a part of the MACCII and MACCIII projects (EU-FP7) done to extend ECMWF’s integrated forecast system (IFS) by adding modules for chemistry, deposition and emission of reactive gases. It is now well known that deposition of chemical species, and especially dry deposition, is a major sink of pollutant in the atmospheric boundary layer. Several parameterizations have been developed in the last decades, while the resistances approach proposed by Wesely (1989) is the most commonly used. However, modeling of dry deposition is confronted to a lack of validation data. Indeed, this kind of parameterization is highly sensible to physical inputs like meteorological and surface fields. Currently, most of models use monthly-means values for dry deposition. It is thus impossible to take into account rapid transitions that can be observed on this inputs and can maybe leads to large biases on the short term forecasts. To investigate impacts of temporal resolution of dry deposition fields on air quality modeling, long simulations were done with the MOCAGE CTM developed at Meteo-France, four type of temporal resolution (monthly-means, daily-means, mean daily-cycle, dynamic velocities) have been used, and results have been evaluated at regional and global scale. They shows important impacts in low-levels atmospheric composition locally, but also at global scale.

MOCAGE is the 3D global off-line chemistry transport model (CTM) run at Meteo-France since 2005 for air quality operational forecasts. Three nested domains are used, with decreasing resolutions (globe, Europe, and France), and 47 vertical levels (from the surface to 5 hPa). For the global and the European domain, input meteorological forcing fields are Meteo-France ARPEGE forecasts. For France, and for the first day of forecast, MOCAGE uses the operational outputs of Meteo-France non-hydrostatic AROME model. This high-resolution (2.5 km) meteorological model is supposed to better represent urban processes (e.g., the urban heat island), which are of strong interest for air quality applications. The purpose of this study is to test the increase of resolution of the CTM MOCAGE over France from 0.1° to 0.025° (i.e. the native resolution of the input meteorological fields).

EnglishSince 1996, French regulations led France to implement an ambitious air quality monitoring infrastructure now based upon monitoring networks operated by 18 local associations in charge of air quality monitoring (AASQA). In the early 2000s, modelling capacities developed and the PREV'AIR system was set up. This pioneering air quality forecasting system in Europe was a major contributor to the MACC and MACC II European projects and now to the Copernicus Atmosphere Monitoring Service (CAMS) designed to make available to all European citizens the required information to monitor their environment and raise their awareness. francaisEn 1996, la loi sur l'air et l'utilisation rationnelle de l'energie a conduit la France a mettre en place un dispositif de surveillance de la qualite de l'air ambitieux reposant sur un reseau aujourd'hui compose de 18 associations agreees de surveillance de la qualite de l'air. Au debut des annees 2000, ce dispositif s'est enrichi de techniques offertes par la modelisation et la prevision, avec la mise en place en 2004 du systeme national Prev'air. Ce systeme pionnier en Europe a ete un contributeur majeur des projets europeens Macc puis Macc II, et aujourd'hui du service Copernicus Atmosphere (Cams) visant a developper et mettre a disposition de tous les citoyens europeens les elements utiles a la surveillance de leur environnement. Ces dispositifs permettant egalement d'attirer l'attention sur ces problematiques.

This paper describes the pre-operational analysis and forecasting system developed during MACC (Monitoring Atmospheric Composition and Climate) and continued in the MACC-II (Monitoring Atmospheric Composition and Climate: Interim Implementation) European projects to provide air quality services for the European continent. This system is based on seven state-of-the art models developed and run in Europe (CHIMERE, EMEP, EURAD-IM, LOTOS-EUROS, MATCH, MOCAGE and SILAM). These models are used to calculate multi-model ensemble products. The paper gives an overall picture of its status at the end of MACC-II (summer 2014) and analyses the performance of the multi-model ensemble. The MACC-II system provides daily 96 h forecasts with hourly outputs of 10 chemical species/aerosols (O3, NO2, SO2, CO, PM10, PM2.5, NO, NH3, total NMVOCs (non-methane volatile organic compounds) and PAN+PAN precursors) over eight vertical levels from the surface to 5 km height. The hourly analysis at the surface is done a posteriori for the past day using a selection of representative air quality data from European monitoring stations. The performance of the system is assessed daily, weekly and every 3 months (seasonally) through statistical indicators calculated using the available representative air quality data from European monitoring stations. Results for a case study show the ability of the ensemble median to forecast regional ozone pollution events. The seasonal performances of the individual models and of the multi-model ensemble have been monitored since September 2009 for ozone, NO2 and PM10. The statistical indicators for ozone in summer 2014 show that the ensemble median gives on average the best performances compared to the seven models. There is very little degradation of the scores with the forecast day but there is a marked diurnal cycle, similarly to the individual models, that can be related partly to the prescribed diurnal variations of anthropogenic emissions in the models. During summer 2014, the diurnal ozone maximum is underestimated by the ensemble median by about 4 μg m−3 on average. Locally, during the studied ozone episodes, the maxima from the ensemble median are often lower than observations by 30–50 μg m−3. Overall, ozone scores are generally good with average values for the normalised indicators of 0.14 for the modified normalised mean bias and of 0.30 for the fractional gross error. Tests have also shown that the ensemble median is robust to reduction of ensemble size by one, that is, if predictions are unavailable from one model. Scores are also discussed for PM10 for winter 2013–1014. There is an underestimation of most models leading the ensemble median to a mean bias of −4.5 μg m−3. The ensemble median fractional gross error is larger for PM10 (~ 0.52) than for ozone and the correlation is lower (~ 0.35 for PM10 and ~ 0.54 for ozone). This is related to a larger spread of the seven model scores for PM10 than for ozone linked to different levels of complexity of aerosol representation in the individual models. In parallel, a scientific analysis of the results of the seven models and of the ensemble is also done over the Mediterranean area because of the specificity of its meteorology and emissions. The system is robust in terms of the production availability. Major efforts have been done in MACC-II towards the operationalisation of all its components. Foreseen developments and research for improving its performances are discussed in the conclusion.

The assimilation of observations from teledetected images in geophysical models requires one to develop algorithms that would account for the existence of coherent structures. In the context of variational data assimilation, a method is proposed to allow the background to be translated so as to fit structure positions deduced from images. Translation occurs as a first step before assimilating all the observations using a classical assimilation procedure with specific covariances for the translated background. A simple validation is proposed using a dynamical system based on the one-dimensional complex Ginzburg–Landau equation in a regime prone to phase and amplitude errors. Assimilation of observations after background translation leads to better scores and a better representation of extremas than the method without translation.

The rapid intensification of Tropical Cyclone (TC) Dora (2007, southwest Indian Ocean) under upper-level trough forcing is investigated. TC–trough interaction is simulated using a limited-area operational numerical weather prediction model. The interaction between the storm and the trough involves a coupled evolution of vertical wind shear and binary vortex interaction in the horizontal and vertical dimensions. The three-dimensional potential vorticity structure associated with the trough undergoes strong deformation as it approaches the storm. Potential vorticity (PV) is advected toward the tropical cyclone core over a thick layer from 200 to 500 hPa while the TC upper-level flow turns cyclonic from the continuous import of angular momentum. It is found that vortex intensification first occurs inside the eyewall and results from PV superposition in the thick aforementioned layer. The main pathway to further storm intensification is associated with secondary eyewall formation triggered by external forcing. Eddy angular momentum convergence and eddy PV fluxes are responsible for spinning up an outer eyewall over the entire troposphere, while spindown is observed within the primary eyewall. The 8-km-resolution model is able to reproduce the main features of the eyewall replacement cycle observed for TC Dora. The outer eyewall intensifies further through mean vertical advection under dynamically forced upward motion. The processes are illustrated and quantified using various diagnostics.

La Réunion is a volcanic island in a tropical zone, which soil undergoes intense erosion. The possible contribution of rainfall to erosion is analyzed and quantified using one year of UHF radar profiler data located at sea level. Measurements of reflectivity, vertical and horizontal wind allow, with suitable assumptions, to determine raindrop vertical and horizontal energy fluxes, which are both essential parameters for erosion. After calibration of radar rain rates, one-year statistics between May 2009 to April 2010 allow to identify differences in rain vertical profiles depending on the season. During the cool dry season, the mean rain rate is less than 2.5 mm h−1 as high as 1.25 km and it decreases at higher altitudes due to the trade winds inversion. During the warm moist season, the mean rain rate is nearly uniform from ground up to 4 km, around 5 mm h−1. The dynamical and microphysical properties of rainfall events are investigated on three cases that are representative of meteorological events in La Réunion: summer deep convection, a cold front and a winter depression embedded in trade winds. For intense rainfall events, the rain rate deduced from the gamma function is in agreement with the rain rate deduced from the mere Marshall Palmer exponential relationship. For less intense events, the gamma function is necessary to represent rain distribution. The deep-convection event is associated to strong reflectivity reaching as high as 10 km, and strong negative vertical velocity. Wind shear is responsible for a deficiency of radar rain detection at the lower levels. During a cold front event, strong reflectivities reach the trade wind inversion (around 4 km high). The trade wind depression generates moderate rain only as high as 2 km. For all the altitudes, the horizontal kinetic energy fluxes are one order of magnitude stronger that than the vertical kinetic energy fluxes. A simple relationship between the reflectivity factor and vertical kinetic energy fluxes is found for each case study.

Several tropical cyclone forecasting centers issue uncertainty information with regard to their official track forecasts, generally using the climatological distribution of position error. However, such methods are not able to convey information that depends on the situation. The purpose of the present study is to assess the skill of the Ensemble Prediction System (EPS) from the European Centre for Medium-Range Weather Forecasts (ECMWF) at measuring the uncertainty of up to 3-day track forecasts issued by the Regional Specialized Meteorological Centre (RSMC) La Réunion in the southwestern Indian Ocean. The dispersion of cyclone positions in the EPS is extracted and translated at the RSMC forecast position. The verification relies on existing methods for probabilistic forecasts that are presently adapted to a cyclone-position metric. First, the probability distribution of forecast positions is compared to the climatological distribution using Brier scores. The probabilistic forecasts have better scores than the climatology, particularly after applying a simple calibration scheme. Second, uncertainty circles are built by fixing the probability at 75%. Their skill at detecting small and large error values is assessed. The circles have some skill for large errors up to the 3-day forecast (and maybe after); but the detection of small radii is skillful only up to 2-day forecasts. The applied methodology may be used to assess and to compare the skill of different probabilistic forecasting systems of cyclone position.

The inherent predictability of tropical cyclone tracks has received much attention since the 1980s. It is still an issue because of the recent improvement of track forecasts by numerical models. The aim of this study is to assess this predictability limit globally using an approach devised by Lorenz on several up-to-date numerical models. The differences between forecasts valid at the same instant are considered to be error values; the doubling time of these small errors leads to an estimated upper bound on predictability. This method is here applied on cyclone position forecasts obtained from three different global operational models (from ECMWF, Météo-France, and the Met Office) over the main tropical cyclone basins in the world and during three recent cyclone seasons (2006–09). The resulting estimates of predictability largely exceed the values that are commonly accepted in the literature. The doubling time of small errors is found between 30 and 50 h. An important consequence is that cyclone track forecasts have not reached their predictability limit yet. It is argued that the previous methods for computing the predictability of tropical cyclone tracks did not constrain the environment and the structure of the cyclones initially. But the Lorenz method could still underestimate the inherent predictability of tropical cyclone tracks. The sensitivity of the predictability estimates to the model characteristics is discussed. In particular, the use of wind bogus is suggested to avoid serial correlations between successive forecasts and to accelerate error growth.

International audience; This study focuses on Monsoon Depressions (MD) in the Southwest Indian Ocean (SWIO) basin. A MD is a subcategory of tropical depressions. Some aspects of MD dynamical characteristics differ from those of the classical tropical depressions. A MD displays a broader horizontal extension (>1000 km), and its structure is marked by deep convection clusters that are generally poorly organized. These convective clusters often develop as a wide convection belt in the eastern semicircle of the system. The MDnucleus is characterized by an area of weak winds about 200 km wide and by a hazy central cloud system, surrounded by a larger belt of strong winds. In the SWIO basin, between 2000 and 2008, 5 systems, out of the 100 tropical cyclones that were followed by RSMC La Réunion, comply with these criteria and then have been identified as MD. Using potential vorticity from European Centre for Medium‐Range Weather Forecasts analyses as stratospheric tracer, stratospheric signatures have been identified in the troposphere for the five events, involving different dynamical mechanisms: tropopause fold between an anticyclone and MD, tropopause fold between MD and a jet front system, and cut‐off low. The dynamical structure of MD, especially the external belt of strong winds, seems to favor tropopause disturbances and leads to stratosphere-troposphere exchange.

Mechanisms leading a synoptic surface cyclone to cross an upper-level zonal jet and its subsequent deepening are investigated using a two-layer model on a β plane. The baroclinic interaction of a low-level circular cyclonic perturbation with an upper-level one is first studied in vertical and horizontal cyclonic or anticyclonic uniform shears. A first nonlinear effect acting on the shape and energetics of the perturbations is analyzed. If the background shear is anticyclonic, the perturbations are stretched horizontally; they lose energy barotropically but gain it baroclinically by a well-maintained westward tilt with height. Conversely, if the shear is cyclonic, perturbations remain quite isotropic, but they do not keep a favorable vertical tilt with time and the baroclinic interaction is thus only transient. The latitudinal motion of the perturbations also results from a nonlinear effect. It is found to depend strongly on the background potential vorticity (PV) gradient. This effect is a baroclinic equivalent of the so-called nonlinear barotropic “β drift” and combines the nonlinear advection and vertical stretching terms. These results are confirmed when the anomalies are initially located south of a confined westerly jet. The poleward shift of the lower cyclonic anomaly occurs faster when the vertically averaged PV gradient is strongly positive, which happens when the jet has a large barotropic component. The lower anomaly crosses the jet from the warm to the cold side and deepens afterward. After a detailed description of this regeneration process with the help of an energy budget, it is shown that linear dynamics are not able to reproduce such behavior.

Disponible à l'adresse : http://ams.allenpress.com/perlserv/?request=get-abstract&doi=10.1175%2FMWR2955.1; International audience; In a nonlinear quasigeostrophic model with uniform potential vorticity, an idealized initial state sharing some features with atmospheric low-predictability situations is built. Inspired by previous work on idealized cyclogenesis, two different cyclogenesis scenarios are obtained as a result of a small change of the initial location of one structure. This behavior is interpreted by analyzing the baroclinic interaction between upper- and lower-level anomalies. The error growth mechanism is nonlinear; it does not depend on the linear stability properties of the jet, which are the same in both evolutions. The ability of ensemble forecasts to capture these two possible evolutions is then assessed given some realistic error bounds in the knowledge of the initial conditions. First, a reference statistical distribution of each of the evolutions is obtained by means of a large Monte Carlo ensemble. Smaller ensembles with size representative of what is available in current operational implementations are then built and compared to the Monte Carlo reference: several singular-vector-based ensembles, a small Monte Carlo ensemble, and a “coherent structure”-based ensemble. This new technique relies on a sampling of the errors on the precursors of the cyclogenesis: amplitude and position errors. In this context, the precursors are handled as coherent structures that may be amplified or moved within realistic error bounds. It is shown that the singular vector ensemble fails to reproduce the bimodal distribution of the variability if the ensemble is not initially constrained, whereas it is accessible at a relatively low cost to the new coherent structures initialization.

Le premier volet du projet Cyprim vise a mieux identifier les precurseurs des cyclogeneses et des precipitations intenses en region mediterraneenne. Cet article en presente les principaux resultats. Il est question des ingredients meteorologiques a l'origine de ces systemes, de la grande a la mesoechelle, des causes des fortes intensites, localisations et parfois stationnarite des recipitations associees. Le coeur de l'article presente des methodes d'identification des anomalies de tourbillon potentiel liees a l'occurrence d'evenements precipitants intenses et dresse une climatologie de ces anomalies. Les perspectives d'exploitation de cette information pour ameliorer la qualite des previsions sont evoquees. ENGLISH VERSION: The first part of the CYPRIM project (Cyclogenesis and intense precipitations in the Mediterranean region) aims at a better identification of the precursor structures of cyclogenesis and heavy precipitation in the Mediterranean region. This article presents the main results. It mentions the meteorological ingredients from which these systems originate from the large to themesoscale, the reasons for large intensities, the localization and sometimes stationarity of the associated precipitations. The heart of the article shows methods for identifiying potential vorticity anomalies involved in the occurrence of heavy precipitation events and sets up a climatology of these anomalies.The perspectives of the usefulness of this information for improving forecast quality are discussed.

Weather forecasting, based on the modelling of atmospheric chaotic system, has a degree of uncertainty. This uncertainty can be assessed by an ensemble prediction system. This PhD-thesis focuses on the post-processing of the Météo-France convection-permitting ensemble prediction system (AROME-EPS). The aim is to develop new tools to extract a relevant information from models for forecasting. These tools are based on deep learning methods. This manuscript starts with the presentation of ensemble forecasts, existing post-processing methods and work methods of forecasters. In order to enhance the forecasting tools currently available, this thesis proposes two complementary approaches to extract new and relevant information from the AROME-EPS.Firstly, an approach to deal with the prediction of a specific type of damaging thunderstorms, called bow echoes, is developed. The objective is to detect them in output of each AROME-EPS member. This detection is made by a convolutional neural network (inspired from the U-Net architecture). The selected neural network demonstrates a good ability to distinguish bow echoes from other convective systems. The skill of this U-Net has led to a daily research mode production summarising the risk of bow echoes. The objective evaluation of forecasts shows the added value of the AROME-EPS compared to the deterministic AROME model beyond 15 lead times.The second approach developed during this thesis aims at implementing a rainfall scenario synthesis. A scenario can be defined as a group of members predicting similar weather events. Since this approach doesn't focus on a particular event, it can be useful on a daily basis to quickly examine ensemble forecasts. This synthesis is based on a dimension reduction of rainfall fields using another convolutional neural network (autoencoder). Each member is expressed and classified in the latent space of the autoencoder which has tens of dimensions. The relevance of this scenario approach is discussed using scores and case studies. The scores show complementarity between the deterministic AROME model and the AROME-EPS. They show it is more interesting to watch the two biggest scenarios in the PE-AROME than two AROME runs. But they also show in case of two equally probable scenarios in the PE-AROME, the scenario that is also the one of an AROME forecast is more likely correct than the other.; La prévision météorologique, basée sur la modélisation du système chaotique atmosphérique, comporte une part d'incertitude. Celle-ci peut être évaluée à l'aide d'ensembles de prévisions. La thèse s'intéresse au post-traitement de la prévision d'ensemble à échelle kilométrique de Météo-France (PE-AROME). L'objectif est de développer de nouveaux outils pour extraire des informations essentielles des modèles pour les prévisions. Ces outils s'appuient sur des méthodes d'apprentissage profond. Ce manuscrit commence par la présentation des prévisions d'ensemble, des méthodes de post-traitement existantes et des méthodes employées aujourd'hui pour expertiser les modèles météorologiques. Cette thèse propose d'enrichir la palette d'outils disponibles avec deux approches complémentaires pour extraire de nouvelles informations de la PE-AROME.Dans un premier temps, une approche pour aider à la prévision de structures orageuses particulières et à fort impact, qu'on nomme les échos arqués, est développée. L'objectif est de les détecter dans les sorties de chaque membre PE-AROME. Cette détection est faite par un réseau de neurones convolutif (de type U-Net). Le réseau de neurones retenu montre une bonne capacité à discerner les échos arqués d'autres organisation orageuses. Les performances de ce U-Net ont permis de déboucher sur une production en mode recherche quotidienne synthétisant le risque d'échos arqués. L'évaluation objective des prévisions montre la plus-value de la PE-AROME comparée au modèle déterministe AROME au-delà de 15 heures d'échéance.La deuxième approche développée durant cette thèse vise à mettre en œuvre une synthèse par scénarios des prévisions de pluies. Un scénario peut être défini comme un groupe de membres prévoyant des évènements météorologiques similaires. Cette approche ne s'intéresse pas à un évènement particulier et est donc moins spécialisée que la détection des échos arqués. Cette synthèse repose sur une réduction de dimension du champ de pluies à l'aide d'un autre réseau de neurones convolutif (de type autoencodeur). Chaque membre est ainsi exprimé et classé dans l'espace latent de l'autoencodeur qui possède quelques dizaines de dimensions. La pertinence de cette représentation sous forme de scénarios est discutée à l'aide de scores et d'études de cas. Les scores montrent une réelle complémentarité entre le modèle déterministe AROME et la PE-AROME. Ils montrent qu'il est plus intéressant de suivre les deux scénarios les plus peuplés de la PE-AROME que deux réseaux AROME. Mais ils montrent aussi qu'en cas de deux scénarios équiprobables dans la PE-AROME, le scénario qui est aussi celui d'une prévision AROME a plus de chance d'être correct que l'autre.

Despite significant improvements in Tropical Cyclone (TC) track forecasts over the past few decades, anticipating the sudden intensity changes of TCs remains a major operational issue. The main purpose of this thesis is to analyze TC rapid intensification processes in relation with external forcing induced by upper-level troughs originating from the mid-latitudes. The impact of initial storm structure on storm evolution and prediction is also documented. An objective definition for rapid intensification in the southwest Indian Ocean is first proposed. The location and frequency of TC-trough interactions are identified, as well as TC-trough arrangements conducive to TC intensification. An interesting study case, TC Dora (2007), is chosen to run numerical simulations initialized with synthetic TC observations blended in a global analysis. The simulated TC-trough interaction is intricate with potential vorticity (PV) advection from the trough into the TC core at mid and upper levels. Vortex intensification first occurs inside the eyewall and results from PV superposition. Further intensification is associated with a subsequent secondary eyewall formation triggered by external forcing from the trough. The numerical model is able to reproduce the main features associated with outer eyewall spin-up, inner eyewall spin-down, and their effects on vortex intensity changes. Another numerical study examines typhoons in the northwest Pacific and demonstrates the critical role played by initial vortex structure in TC track and intensity prediction. Upgrading the initial specification of a TC inner-core structure in numerical models is recommended for future TC prediction improvements.; Dans un contexte international, la prévision d'intensité des cyclones tropicaux connaît encore de graves déficiences tandis que la prévision de trajectoire de ces phénomènes météorologiques extrêmes s'est grandement améliorée ces dernières décennies. Une source d'erreur pour la prévision d'intensité est le manque de connaissance des processus physiques qui régissent l'évolution de la structure et de l'intensité des cyclones. Cette thèse, proposée dans le cadre des responsabilités du Centre Météorologique Régional Spécialisé (CMRS) de la Réunion et des axes de recherche du LACy et du CNRM, a pour but d'améliorer la prévision numérique et la compréhension des mécanismes de changement de structure et d'intensité des cyclones dans le sud-ouest de l'océan Indien. On observe statistiquement dans le bassin de fréquents déferlements d'ondes de Rossby qui correspondent à une intrusion des talwegs d'altitude depuis les moyennes latitudes vers les régions où évoluent les cyclones. Ces déferlements advectent dans la troposphère tropicale de l'air d'origine stratosphérique à fort tourbillon potentiel (PV). Le cœur d'un cyclone tropical étant caractérisé par un vortex cyclonique de fort PV, il est donc légitime de se demander si de tels talwegs sont capables de « nourrir » un cyclone en déferlant jusqu'à lui, et l'intensifier par superposition de PV. D'un autre côté, l'approche d'un talweg est associée à d'autres facteurs pouvant jouer en défaveur d'une intensification, comme un fort cisaillement vertical de vent. L'étude de processus est réalisée sur le cyclone Dora (2007) avec le modèle opérationnel du CMRS sur le bassin, Aladin-Réunion. Ce modèle hydrostatique à aire limitée bénéficie d'une résolution horizontale de 8 km et de son propre schéma d'assimilation 3Dvar avec bogus de vent. Un tel bogus permet d'affiner la structure du cyclone à l'instant initial en ajoutant des observations de vent déduites d'un profil analytique et des paramètres de structure du cyclone estimés par les images satellites. Des diagnostiques sur les variables thermodynamiques en sortie de modèle montrent que la phase d'intensification rapide de Dora est bien associée à l'advection de tourbillon potentiel (PV) en provenance du talweg. Bien que fortement cisaillé, le système parvient à s'intensifier grâce à la forte inclinaison du talweg qui advecte du PV au cœur du cyclone en 2 temps et à 2 niveaux (haute et moyenne troposphère). Lorsque le talweg est au plus proche du cyclone, il force un processus dynamique interne appelé « cycle de remplacement du mur de l'œil ». On observe une inclinaison et un renforcement des vitesses verticales à l'extérieur du mur de l'œil principal, associé à une accélération de la circulation cyclonique tangentielle par advection de moment angulaire sur toute l'épaisseur de la troposphère dans cette zone annulaire (mis en évidence par les flux d'Eliassen-Palm). Un second maximum de vent relatif apparaît alors et une deuxième phase d'intensification rapide s'ensuit avec la contraction du mur secondaire. Le forçage de processus internes par une influence externe (un talweg) semble donc être le moteur de l'intensification rapide de Dora dans un environnement cisaillé, et potentiellement celui d'autres cyclones dans le bassin qui sont approchés par des talwegs d'altitude. Les prévisionnistes du CMRS sont invités à surveiller les champs de PV de tels systèmes, en attendant que de plus amples diagnostiques soient réalisés avec l'outil d'inversion du tourbillon potentiel développé sur le modèle global Arpège.