Heavy precipitation (HP) constitutes a major meteorological threat in the western Mediterranean (WMed). Every year, recurrent events affect the area with fatal consequences for infrastructure and personal losses. Despite this being a well-known issue widely investigated in the past, open questions still remain. Particularly, the understanding of the underlying mechanisms and the modeling representation of the events must be improved. One of the major goals of the Hydrological Cycle in the Mediterranean Experiment (HyMeX; 2010–2020) has been to advance knowledge on this topic. In this article, we present an overview of the most recent lessons learned from HyMeX towards an improved understanding of the mechanisms leading to HP in the WMed. The unique network of instruments deployed as well as the use of finer model resolutions and coupled models provided an unprecedented opportunity to validate numerical model simulations, develop improved parameterizations, and design high-resolution ensemble modeling approaches and sophisticated assimilation techniques across scales. All in all, HyMeX, and particularly the science team heavy precipitation, favored the evidencing of theoretical results, the enrichment of our knowledge on the genesis and evolution of convection in a complex topography environment, and the improvement of precipitation forecasts. Illustratively, the intervention of cyclones and warm conveyor belts in the occurrence of heavy precipitation has been pointed out, and the crucial role of the spatiotemporal distribution of atmospheric water vapor for the understanding and accurate forecast of the timing and location of deep convection has been evidenced, as has the complex interaction among processes across scales. The importance of soil and ocean conditions and the interactions among systems were highlighted, and such systems were specifically developed in the framework of HyMeX to improve the realism of weather forecasts. Furthermore, the benefits of cross-disciplinary efforts within HyMeX have been a key asset in bringing our knowledge about heavy precipitation in the Mediterranean region a step forward.

During autumn 2012 and winter 2013, two special observation periods (SOPs) of the HYdrological cycle in the Mediterranean EXperiment (HyMeX) took place. For the preparatory studies and to support the instrument deployment during the field campaign, a dedicated version of the operational convective-scale Application of Research to Operations at Mesoscale (AROME)-France model was developed: the AROME-WMED (West Mediterranean Sea) model. It covers the western Mediterranean basin with a 48 h forecast range. It provided real-time analyses and forecasts which were sent daily to the HyMeX operational centre to forecast high-precipitation events and to help decision makers on the deployment of meteorological instruments. This paper presents the main features of this numerical weather prediction system in terms of data assimilation and forecast. Some specific data of the HyMeX SOP were assimilated in real time. The forecast skill of AROME-WMED is then assessed with objective scores and compared to the operational AROME-France model, for both autumn 2012 (05 September to 06 November 2012) and winter 2013 (01 February to 15 March 2013) SOPs. The overall performance of AROME-WMED is good for the first HyMeX special observation period (SOP1) (i.e. mean 2 m temperature root mean square error (RMSE) of 1.7 °C and mean 2 m relative humidity RMSE of 10 % for the 0–30 h forecast ranges) and similar to those of AROME-France for the 0–30 h common forecast range (maximal absolute difference of 2 m temperature RMSE of 0.2 °C and 0.21 % for the 2 m relative humidity); conversely, for the 24–48 h forecast range it is less accurate (relative loss between 10 and 12 % in 2 m temperature and relative humidity RMSE, and equitable threat score (ETS) for 24 h accumulated rainfall), but it remains useful for scheduling observation deployment. The characteristics of parameters, such as precipitation, temperature or humidity, are illustrated by one heavy precipitation case study that occurred over the south of Spain.

The Infrared Atmospheric Sounding Interferometer (IASI) is an essential instrument for numerical weather prediction (NWP). It measures radiances at the top of the atmosphere using 8461 channels. The huge amount of observations provided by IASI has led the community to develop techniques to reduce observations while conserving as much information as possible. Thus, a selection of the 300 most informative channels was made for NWP based on the concept of information theory. One of the main limitations of this method was to neglect the covariances between the observation errors of the different channels. However, many centres have shown a significant benefit for weather forecasting to use them. Currently, the observation-error covariances are only estimated on the current IASI channel selection, but no studies to make a new selection of IASI channels taking into account the observation-error covariances have yet been carried out. The objective of this paper was therefore to perform a new selection of IASI channels by taking into account the observation-error covariances. The results show that with an equivalent number of channels, accounting for the observation-error covariances, a new selection of IASI channels can reduce the analysis error on average in temperature by 3 %, humidity by 1.8 % and ozone by 0.9 % compared to the current selection. Finally, we go one step further by proposing a robust new selection of 400 IASI channels to further reduce the analysis error for NWP.

This article investigates the potential of W-band radar reflectivity to improve the quality of analyses and forecasts of heavy precipitation events in the Mediterranean area. The “1D+3DVar” assimilation method, operationally employed to assimilate ground-based precipitation radar data in the Météo-France kilometre-scale numerical weather prediction (NWP) model AROME, has been adapted to assimilate the W-band reflectivity measured by the airborne cloud radar RASTA (Radar Airborne System Tool for Atmosphere) during a 2-month period over the Mediterranean area. After applying a bias correction, vertical profiles of relative humidity are first derived via a 1-D Bayesian retrieval, and then used as relative humidity pseudo-observations in the 3DVar assimilation system of AROME. The efficiency of the 1-D Bayesian method in retrieving humidity fields is assessed using independent in-flight humidity measurements. To complement this study, the benefit brought by consistent thermodynamic and dynamic cloud conditions has been investigated by separately and jointly assimilating the W-band reflectivity and horizontal wind measurements collected by RASTA in the 3 h 3DVar assimilation system of AROME. The data assimilation experiments are conducted for a single heavy precipitation event and then also for 32 cases. Results indicate that the W-band reflectivity has a larger impact on the humidity, temperature and pressure fields in the analyses compared to the assimilation of RASTA wind data alone. Besides, the analyses get closer to independent humidity observations if the W-band reflectivity is assimilated alone or jointly with RASTA wind data. Nonetheless, the impact of the W-band reflectivity decreases more rapidly as the forecast range increases when compared to the assimilation of RASTA wind data alone. Generally, the joint assimilation of the W-band reflectivity with wind data results in the best improvement in the rainfall precipitation forecasts. Consequently, results of this study indicate that consistent thermodynamic and dynamic cloud conditions in the analysis leads to an improvement of both model initial conditions and forecasts. Even though to a lesser extent, the assimilation of the W-band reflectivity alone also results in a slight improvement of the rainfall precipitation forecasts.