A central European program pursues regional extensive scientific and technical collaboration in the field of operational limited-area numerical modeling for regional and local weather research, forecasting, and applications.
Limited-area numerical weather prediction (NWP) models are essential tools for regional and local weather prediction and atmospheric research. The development, implementation, and operational execution of a limited-area model (LAM) are demanding on both human and computer resources, and it is scientifically and technically challenging as well. An operational LAM system should add value to the global NWP forecast; therefore, its model core (physics, dynamics, and physics–dynamics interface) needs to be appropriately designed for its higher resolution and be adapted to local characteristics, such as orography and soil features (Mesinger 2001; Kalnay 2003). Further, an operational LAM system should utilize data assimilation for better use of dense observations available locally and for routine procedures, such as preprocessing of observations and data quality control. Code management, a scripting system, and library cycle/version management are vital for an operational NWP system. Many small- to medium-sized national (hydro-) meteorological services (NHMSs) do not have the capacity to develop and run an operational LAM system alone, because of a lack of human resources, NWP knowledge, and technical facilities. In Europe, many NHMSs have recognized the need for international cooperation in the development of an operational LAM system since the 1980s. The success of the European Centre for Medium-Range Weather Forecasts (ECMWF) has shown the benefit of joining forces to improve operational NWP systems. It has strongly encouraged many NHMSs to cooperate on operational LAM systems to improve short-range forecasts. In the last 30 years, 5 NWP consortia among more than 30 European NHMSs have been established for the development of operational LAM systems: Aire Limitée Adaptation Dynamique Développement International (ALADIN), the Consortium for Small-Scale Modeling (COSMO), the High Resolution Limited Area Model (HIRLAM) group, the Regional Cooperation for Limited Area Modelling in Central Europe (RC LACE), and very recently the South-East European Consortium for Operational Weather Prediction (SEECOP).
RC LACE was initiated in 1990. It is a collaborative program in central Europe and is currently supported by the NHMSs of Austria, Croatia, the Czech Republic, Hungary, Romania, Slovakia, and Slovenia, which are listed in Table 1. The principle objectives of RC LACE are as follows:
to develop and operate a state-of-the-art LAM and data assimilation system in the member states and
to conduct joint scientific and technical research toward improving the quality of the forecasts.
NHMSs of RC LACE.
The aim of this paper is to give an overview of RC LACE: its 27-yr journey of cooperation, the organizational concept, the LAM and data assimilation (DA) system implemented in the member states and the consortium’s common operational actions on the real-time exchange of dense national observations, data preprocessing for DA, verification, and ensemble forecasting. We highlight the RC LACE research and development (R&D) contribution to operational NWP. We also discuss the RC LACE experiences and challenges and its collaboration with the ALADIN and HIRLAM consortia.
HISTORY AND ORGANIZATION OF RC LACE.
In the late 1980s, profound political changes in Europe through the fall of the so-called Iron Curtain dividing Europe brought great opportunities for cooperation in central Europe. These opportunities were soon recognized by central European NHMSs, as they realized the necessity, importance, and benefits of international cooperation on LAMs.
In 1990, a meeting was convened at ZAMG in Vienna, Austria, to explore the feasibility of the establishment of a central European center for limited-area modeling, following the example of ECMWF. This was the beginning of RC LACE, where the “RC” stood for regional center but was later changed to regional cooperation. The official structure was established by the end of 1994 with the NHMSs of Austria, Croatia, Czech Republic, Hungary, Slovakia, and Slovenia. The NHMS of Romania joined RC LACE in 2007.
The French government also quickly spotted these opportunities and provided substantial financial support to the cooperation with central European countries. With this backing, Météo-France proposed to develop an LAM version of its global model Action de Recherche Petite Echelle Grande Echelle (ARPEGE; Courtier and Geleyn 1988), desirable for operational NWP application in France and the central European countries. This proposal included training, research, and development components. The LAM-ARPEGE Project, which later became ALADIN, started in Toulouse, France, in September 1991, with a majority of participating scientists coming from future RC LACE member countries. In this context, an efficient synergy developed between the central European and Météo-France initiatives. In fact, the support of Météo-France has been fundamental for the development of RC LACE.
The first operational application was made possible when RC LACE concluded an official agreement with Météo-France in November 1994. By spring 1996, the ALADIN/LACE model was run on the supercomputer at Météo-France, Toulouse, under the responsibility of the core RC LACE operational team.
The ALADIN/LACE operations were transferred from Toulouse to CHMI, Prague, Czech Republic, in July 1998, while ZAMG in Vienna became the telecommunications and archival center in this period. The Prague center also acted as the backup for the reference ALADIN system software, thereby providing a valuable service to the ALADIN community. The application covered the whole central European domain (Fig. 1) with 12.2-km horizontal resolution and 31 vertical levels, and it was run twice per day at 0000 and 1200 UTC up to 36 h, with lateral boundary conditions provided by the global model ARPEGE.
The model domain and topography of ALADIN/LACE as of Jul 1998. The horizontal resolution was 12.2 km.
Citation: Bulletin of the American Meteorological Society 99, 7; 10.1175/BAMS-D-16-0321.1
The regional center covered not only NWP operations, but it was also the focal point of R&D. Scientists from RC LACE member countries frequently visited the regional center to perform research studies and technical development on the ALADIN system. Over time, as technology became more accessible, the ALADIN system has been implemented by local teams in the member NHMSs.
Increasing demands for new ALADIN products and thus a growing number of applications in member countries forced the transformation of RC LACE into decentralized operations, which was implemented in January 2003. Since then, “RC” stands for regional cooperation instead of regional center.
In the current RC LACE organization, each NHMS is responsible for its own operational NWP system; however, scientific research and developments are coordinated and supported by a common budget. The RC LACE Memorandum of Understanding defines the cooperation, stating the objectives, rules, and terms of reference of the relevant bodies, which include the council for governance, the management group for execution, and the steering committee for an advisory role.
THE MODEL SYSTEM.
As a result of the historical development, RC LACE exploits and contributes to the ALADIN system, which shares its model code with the global NWP systems ARPEGE and the Integrated Forecast System (IFS) of the ECMWF. Currently, there are two canonical model configurations (CMCs; Termonia et al. 2018), named Application of Research to Operations at Mesoscale (AROME; Seity et al. 2011) and ALADIN–AROME (ALARO). They are unified to a maximum possible extent by sharing dynamical kernel and data assimilation, differing mainly by the used set of physical parameterizations. In short, the concept of AROME is to enable a fast operational implementation of physical package upgrades coming from the research model Meso-NH (Lafore et al. 1998). It is designed for scales resolving the moist deep convection (typical horizontal mesh sizes of 2.5 km and finer). The concept of ALARO, relying on the multiscale physics, provides the missing bridge where this phenomenon becomes partly resolved, that is, for horizontal mesh sizes roughly between 7 and 3 km (the so-called convection-permitting scales). ALARO makes it possible to go progressively from the larger-scale applications, such as climate or ensemble prediction systems, down to convection-resolving NWP. Running ALARO at a wide range of resolutions is an important option, enabling a gradual quality increase of operational weather forecasts pending available computing resources. The multiscale character of ALARO physics enables the use of the ALADIN system in a large variety of applications.
Data assimilation is applied to improve atmosphere and land surface initial conditions (ICs). To initialize soil parameters, all members use the optimal interpolation (OI) method (Giard and Bazile 2000) to assimilate screen-level synoptic observations. For the upper-air ICs, ALADIN three-dimensional variational data assimilation (3D-Var; Fischer et al. 2005; Bölöni 2006) is used by most of the members. The implementations differ, in particular in cycling frequency (6 or 3 h), observation usage, and the background error covariance specification. Furthermore, a pseudoassimilation method called digital filter (DF) blending (Brožková et al. 2001) has been used. Recently, a combination of the DF blending with the 3D-Var—the BlendVar configuration—was operationally implemented within RC LACE.
The ALADIN 3D-Var system allows assimilation of both in situ and remote sensing observations: synoptic measurements, wind profilers, radiosondes, aircrafts, spaceborne wind measurements, Doppler wind and reflectivity from ground-based radars, and satellite radiances, as well as ground-based Global Navigation Satellite System (GNSS) measurements. The preprocessing of observations for DA is handled in common by the Observation Preprocessing System for RC LACE (OPLACE) system. In addition to the OPLACE data, some other national observations, such as radar and GNSS, are assimilated locally.
The ALADIN system offers enough flexibility for individual implementation. Within RC LACE, members run one or more CMCs, adapted to their specific needs and computer capacities. Details on the settings are summarized in Table 2, and the operational model domains are presented in Fig. 2.
Summary of operational ALADIN configurations at RC LACE NHMSs. Colors correspond to the colored boxes in Fig. 2.
The operational ALADIN model domains at RC LACE NHMSs. The colors correspond to the configurations of ALADIN shown in Table 2.
Citation: Bulletin of the American Meteorological Society 99, 7; 10.1175/BAMS-D-16-0321.1
VERIFICATION.
Successful development and improvement of an NWP model requires an extensive analysis of its strengths and weaknesses. Large variability in forecast quality is usually visible over time, reflecting the type of verification statistics being used and the predictability of different flow regimes. To overcome this difficulty, a combined statistical score for the ZAMG NWP index (ZNI) has been implemented at ZAMG, where it is used to monitor the overall long-term evolution of the operational ALADIN models. The ZNI basically follows the U.K. NWP index (UKMO 2010) with some minor differences concerning the choice of meteorological parameters and scores. The ZNI includes statistical measures such as mean error, mean-square error, or equitable threat score for seven parameters, giving slightly more weight to those showing higher potential for improvement (e.g., global radiation, precipitation). Figure 3 shows the evolution of the ZNI for the period 2005–16. To reduce the variability of the presented data and to extract an overall trend, the ZNI is presented as a 1-yr running average (blue curve), a 1-yr running 90th percentile (P90; green), and a 1-yr running 10th percentile (P10; red). An overall increase of the forecast quality is evident for the period shown. The improvement is most obvious in the evolution of P90, indicating a reduction in forecasts with large errors. Some of the changes in forecast quality can be directly correlated to operational changes in the model setup. The change from ALADIN (9.6 km) to ALARO (9.6 km) at ZAMG in 2008 is accompanied by a significant rise in the ZNI, while the change to ALARO (4.8 km) in 2011 brought a decrease in the ZNI. This drop in forecast quality ends with the introduction of several upgrades of the ALARO physics in 2012 and 2013. The next significant change in ZNI is visible by the end of 2014, which corresponds to the period when the AROME (2.5 km) forecast data starts to be used.
ZNI shown as a running 30-day average (black points), a running 1-yr average (blue), a running 1-yr 10th percentile (P10; red), and a running 1-yr 90th percentile (P90; green) for the period 2005–16.
Citation: Bulletin of the American Meteorological Society 99, 7; 10.1175/BAMS-D-16-0321.1
Besides the verification scores computed by each member for its applications, RC LACE runs a common verification allowing a comparison of different model versions. Members send their operational forecasts for a predefined set of locations to ARSO in Ljubljana, Slovenia, where they are verified in a unified way. Similarly, monthly averages of selected scores are visualized on the RC LACE web portal.
COMMON OPERATIONS.
Even with decentralized model operational suites, RC LACE runs several common applications for the benefit of the whole consortium, which reduces costs and improves the use of staff resources: 1) an observation preprocessing system for data assimilation and verification, 2) the mesoscale ensemble system ALADIN–Limited Area Ensemble Forecasting (ALADIN-LAEF), and 3) a web portal to make available forecast products of all RC LACE NHMSs to the whole consortium.
OPLACE.
Observational data1 handling is a demanding subject, and collaboration is desirable. Observation preprocessing may comprise simple quality checks and data format conversions but also more advanced processing. To avoid duplication of work and to ensure the provision of observations in an appropriate format to all RC LACE members, OPLACE was built in 2009.
OPLACE is hosted by OMSZ and currently provides surface synoptic data, upper-air sounding, wind profiler, and aircraft observations. Besides conventional observations, there are also various remote sensing data, such as AMSU-A, AMSU-B, HIRS, and IASI radiances, provided by the Meteorological Operational (MetOp) and National Oceanic and Atmospheric Administration (NOAA) polar-orbiting satellites, atmospheric motion winds, and SEVIRI radiances from Meteorological Satellite-10 (Meteosat-10). OPLACE is updated every 20 min; the observations are split into hourly time slots and are separated by observation type. The separation allows users to download only what they need.
RC LACE NHMSs exchange their dense national surface synoptic measurements and high-resolution aircraft Mode-S MRAR data in real time in OPLACE, as illustrated in Fig. 4. RC LACE might have the first transnational agreement to exchange observations in real time for operational NWP among so many countries.
The coverage of surface observations available in the WMO Global Telecommunication System (blue squares) and of denser national observations (red dots) exchanged by OPLACE.
Citation: Bulletin of the American Meteorological Society 99, 7; 10.1175/BAMS-D-16-0321.1
OPLACE has facilitated the implementation of data assimilation in RC LACE countries as a result of its stable and reliable framework. It allows an easier handling of local and/or international observation network upgrades like changes to World Meteorological Organization (WMO) data formats. In the near future, OPLACE will use observations from new satellite sensors, such as CrIS and ATMS, as well as GNSS, and radar data.
Regional ensemble ALADIN-LAEF.
The mesoscale ensemble prediction system ALADIN-LAEF has been in development since 2008 (Wang et al. 2011). It provides a reliable short-range probabilistic forecast for members and probabilistic information for downstream applications such as civil warnings, hydrology, transportation, and the energy industry. It became operational in 2011 and was upgraded in 2013.
The current operational ALADIN-LAEF consists of 16 perturbed ensemble members and the control run. It runs at a resolution of 11 km in the horizontal and 45 levels in the vertical and is integrated up to 72 h at 0000 and 1200 UTC each day. An upgrade to 5-km horizontal resolution and 60 vertical levels is already being prepared. The ALADIN-LAEF domain covers all of Europe and a large part of the Atlantic as shown in Fig. 5. The IC perturbations are generated by a blending method (Wang et al. 2014), and lateral boundary condition (LBC) perturbations are provided by the first 16 ECWMF Ensemble Prediction System (EPS) members (Leutbecher and Palmer 2008). An ensemble of surface data assimilations (Belluš et al. 2016) is implemented in ALADIN-LAEF for surface IC perturbations. The model uncertainties are simulated by several combinations of different microphysics, radiation, deep and shallow convection, and turbulence schemes in the ALADIN system.
ALADIN-LAEF domain (blue line). The dots represent the Global Telecommunication System observation sites (red dots) and additional RC LACE national observations (green dots) processed in the ALADIN-LAEF ensemble surface assimilation. The black inner line denotes the postprocessing domain.
Citation: Bulletin of the American Meteorological Society 99, 7; 10.1175/BAMS-D-16-0321.1
ALADIN-LAEF runs operationally at the high-performance computing facility at ECMWF. Its analyses and forecasts over the RC LACE domain are archived in the Meteorological Archival and Retrieval System (MARS) at ECMWF and are distributed in real time to members. A basic set of probabilistic products are displayed on the RC LACE web portal.
RC LACE web portal.
The RC LACE web portal (www.rclace.eu) provides members with 1) observation monitoring, 2) operational products and verification scores of operational ALADIN suites from members, and 3) probabilistic products of ALADIN-LAEF. The online operational products are available daily in a uniform graphical environment, updated regularly four times per day. The verification scores of the main meteorological parameters from all RC LACE operational suites are postprocessed on a monthly basis and are available in graphical and digital form.
Further, the web portal maintains a forum for interactive discussion and information exchange in the community. It is mostly used by colleagues of the ALADIN, RC LACE, and HIRLAM consortia, and it currently has 131 registered members who have contributed to more than 1,200 active posts. Other RC LACE relevant information, such as R&D projects, documents and publications, and meeting events, is also available on the web portal.
R&D HIGHLIGHTS.
From the beginning, RC LACE has worked closely with the ALADIN consortium on the development of the ALADIN system. RC LACE has contributed roughly one-third of the manpower used for the ALADIN system development. In recent years, the major effort of RC LACE has been put toward the development of ALADIN CMCs ALARO and AROME, and the implementation of member applications at high resolutions, up to convection-resolving scales. RC LACE R&D contribution to state-of-the-art NWP is being made in dynamics, physics, data assimilation, and ensemble prediction, which are described below.
Model dynamics.
RC LACE has been involved in several crucial developments of the ALADIN dynamical core. First, coupling to the global model was addressed by Radnóti et al. (1995). Then, time and advection schemes were developed, and a nonhydrostatic equation system was designed as the necessary difference with respect to the proven hydrostatic version; see Bubnová et al. (1995) and Bénard et al. (2010) for details. The ALADIN dynamical core was adopted in ALARO and AROME. We describe it briefly here with an emphasis on the contributions from RC LACE. An example of a characteristic situation that is captured well with a nonhydrostatic version of the ALARO model configuration, an orographic lee wave, is shown in Fig. 6.
An orographic wave over the Krušné Mountains on the western border of the Czech Republic. (left) The wave is captured by the Aqua satellite Moderate Resolution Imaging Spectroradiometer (MODIS) 250-m band 1 at 1150 UTC 27 Jan 2008 [data source: NOAA Comprehensive Large Array-Data Stewardship System (CLASS) archive, processed by Martin Setvák, CHMI]. (right) Vertical velocity (m s−1) at 850 hPa from the nonhydrostatic ALARO simulation at 1-km horizontal resolution with 87 vertical levels. The orographic wave with the correct wavelength is clearly visible close to the state border.
Citation: Bulletin of the American Meteorological Society 99, 7; 10.1175/BAMS-D-16-0321.1
Either hydrostatic primitive equations (HPE) or nonhydrostatic fully elastic Euler equations (EE) are solved using semi-Lagrangian advection and semi-implicit time marching schemes that allow for long time steps (∼3 min at 5-km horizontal mesh size). Spectral field representation with double Fourier decomposition is used to calculate horizontal derivatives, to solve the Helmholtz equation, and to apply linear diffusion. Then the fields are transformed to the collocation grid to perform all nonlinear calculations, including semi-Lagrangian advection and physical processes. The horizontal grid is unstaggered, since the precision of derivatives is guaranteed by the spectral method.
In the vertical, a mass-based hybrid terrain-following coordinate of Laprise (1992) is used, making the concept of hydrostatic pressure applicable also in the nonhydrostatic EE core. The vertical grid is staggered to suit the finite-difference discretization of Simmons and Burridge (1981), which was adjusted and completed for the nonhydrostatic EE core. However, since this discretization is only first-order accurate for nonuniform spacing of vertical levels, an alternative finite-element vertical discretization was progressively implemented for both hydrostatic and nonhydrostatic options. In the HPE case, it uses cubic B-splines as basis functions (Untch and Hortal 2004). In the nonhydrostatic EE core, a more general scheme was introduced, enhancing accuracy to an arbitrary order (Vivoda and Smolíková 2013).
As stated above, the time marching scheme is semi-implicit. The reference linear operator used for its derivation has constant coefficients. For the EE core, two additional nonhydrostatic prognostic variables are introduced. They differ in spectral and nonlinear computations. The choice in the spectral part is determined by stability requirements (Bénard at al. 2004, 2005), while that in the nonlinear part is determined by the accuracy of numerical approximation (Ch. Smith 2003, unpublished manuscript; Bénard at al. 2010). Using only two time levels requires the extrapolation of nonlinear source terms with care for stability (Hortal 2002). Alternatively, the iterative centered-implicit scheme, developed namely for the nonhydrostatic EE core, allows the extrapolation to be avoided (Bénard 2003).
In addition, RC LACE is a main contributor to the development of the semi-Lagrangian horizontal diffusion (SLHD). This nonlinear diffusion scheme was proposed and implemented by Váňa et al. (2008) as an option to the existing linear spectral diffusion. The damping properties of semi-Lagrangian interpolators are used to represent diffusive processes with the strength being controlled by local flow deformation. It better reflects the nature of turbulent dissipation, which is strongly flow dependent and thus nonlinear. Furthermore, it may be applied on all the advected fields (even those not being spectrally represented), it can be enhanced to have a full 3D character, and it is numerically efficient.
Model physics.
The major RC LACE contribution to the ALADIN physics is the development of ALARO. The success of the ALARO concept arises from the following design principles:
multiphase governing equations cast in the flux form, ensuring conservation of mass, momentum, and energy via the consistent set of approximations used;
multiscale character, allowing for a progressive increase in resolution and wide range in application;
modularity of the schemes, enabling flexible time step organization and progressive increases in complexity; and
use of prognostic variables where possible (water species, turbulent kinetic energy, updraft–downdraft velocity, mesh area fraction, etc.), ensuring memory and interactions of the schemes.
First, the mass-weighted equations for moist physics of Catry et al. (2007) became the general thermodynamic basis for the development of microphysical computations. To safely afford the long time steps allowed by the ALADIN dynamics, Geleyn et al. (2008) implemented a statistical approach to the sedimentation of precipitation.
Second, the convection parameterization problem is addressed by introducing the microphysical treatment of convective precipitation separately from the convective transport of enthalpy, momentum, and nonprecipitating water species. The key design feature is a single call of the microphysics scheme, using joint input provided by subgrid- and grid-scale condensation. This, together with the vertical geometry of clouds and precipitation, ensures a smooth transition between the resolved and unresolved origin of precipitation. The resulting Modular Multiscale Microphysics and Transport (3MT) scheme of Gerard et al. (2009) enables convection-permitting scales to be accessed, and therefore it became an ALARO flagship. The Cold Air Outbreak case study of Field et al. (2017) confirmed the independence of the domain-averaged ALARO precipitation forecast on horizontal resolution; its seamless character is qualitatively demonstrated in Fig. 7.
Seamless ALARO prediction: 1-h precipitation accumulated from a 30–31-h forecast, starting at 1200 UTC 30 Jan 2010. The computational domain covers the relatively small area between the Faroe and Orkney Islands. The large-scale structure of the precipitation pattern is roughly kept as more details appear progressively with finer model resolutions.
Citation: Bulletin of the American Meteorological Society 99, 7; 10.1175/BAMS-D-16-0321.1
Third, the turbulence scheme Third Order Moments Unified Condensation Accounting and N-Dependent Solver (TOUCANS) of Bašták Ďurán et al. (2014) is based on a unified treatment of stability functions, applicable in both stable and unstable regimes. TOUCANS can emulate several turbulence models compliant with observational evidence that there is no stability limit suppressing the turbulent mixing completely. The basic version of TOUCANS uses a prognostic equation for turbulent kinetic energy, and it includes the effects of water phase changes consistently. A more advanced version adds the prognostic treatment of total turbulent energy, and it can parameterize the higher-order terms responsible for countergradient transport, which is not permitted by simple eddy diffusivity schemes.
Fourth, the radiation scheme Actif Calcul de Rayonement et Nébulosité, version 2 (ACRANEB2), of Mašek et al. (2016) and Geleyn et al. (2017) elaborated the broadband approach, reaching its accuracy limits in both shortwave and longwave spectra. The key design feature is a full cloud–radiation interaction, made affordable via a broadband solver applied at every model grid point and time step, using actual cloud and intermittently updated gaseous optical properties.
Data assimilation.
RC LACE started LAM DA with the implementation of a poor man’s data assimilation proxy, the DF blending method, proposed by Brožková et al. (2001). It combines the global analysis and LAM background in a way that the large-scale information from the global model is incrementally added to the high-resolution LAM background. The DF-blending 6-h assimilation cycle became operational at RC LACE in Prague in 2001. The idea of combining large-scale analyses in LAM data assimilation has been widely explored since then (Guidard et al. 2006; Derková and Belluš 2007; Guidard and Fischer 2008; Brožková et al. 2006; Yang 2005; Wang et al. 2014). The hybrid method BlendVar was proposed by Široká et al. (2003), and it became operational at CHMI in 2015 (Bučánek et al. 2015).
In the 1990s it was decided to develop an LAM version of the variational assimilation algorithms available in ARPEGE/IFS. The adaptation started with the plane geometry aspects, then a successful algorithmic implementation took place for sensitivity tests of the initial conditions and for four-dimensional variational data assimilation (4D-Var) experiments, and was finalized with extensive validation work on an observational part of the variational algorithm around 2000. The first RC LACE upper-air data assimilation, based on ALADIN 3D-Var, became operational at OMSZ in Budapest, Hungary, in 2005 (Bölöni 2005). To improve the representation of forecast errors in DA, different error sampling strategies and flow-dependent components, such as spatially varying error structures (Strajnar 2008), were studied. RC LACE also participated in the development and testing of the first version of the ALADIN 4D-Var and assembled an experimental ensemble transform Kalman filter (ETKF) system as well (Csomós and Bölöni 2009). At the same time, other methods accounting for observations with reduced representativity errors, such as the Rapid Update Cycle (RUC) and the 3D first guess at appropriate time (3D-FGAT), were investigated.
Mesoscale data assimilation is one topic of RC LACE research. The 3D-Var RUC technique is recognized as an affordable and successful approach at many NHMSs (Mile et al. 2015). Very recently, an hourly rapid-refresh AROME assimilation system using radar wind and reflectivity measurements was implemented at ZAMG with 2.5- and 1.2-km horizontal resolution and 90 vertical levels for nowcasting purposes. In this AROME nowcasting framework, incremental analysis updates (Bloom et al. 1996) and latent heat nudging (Jones and Macpherson 1997; Meier 2015) are incorporated, based on the precipitation analysis and nowcasting of Integrated Nowcasting through Comprehensive Analysis (INCA; Haiden et al. 2011; Wang et al. 2017). The results are quite encouraging, particularly from a forecaster’s perspective.
Studies have been conducted on the assimilation of nonconventional data at the highest possible resolution. Satellite observations, such as AMSU-A, AMSU-B, SEVIRI, and AMV (Randriamampianina 2006), have been the components of many RC LACE DA systems. Furthermore, zenith total delay measurements from GNSS (Yan et al. 2008) have been studied in mesoscale AROME and ALARO. Recently, the implementation of new observations has been investigated, for instance, the aircraft-derived Mode-S MRAR (Strajnar 2012; Strajnar et al. 2015) and the High Resolution Winds (HRW) AMV software in ALARO and in AROME 3D-Var (Mile and Randriamampianina 2012). These observations have an important influence on analyses. An example of absolute and relative degree of freedom for signal (DFS) scores is shown in Fig. 8; it can further increase the benefit of local DA systems within RC LACE.
(top) Absolute and (bottom) relative DFS: that is, information content for the main conventional [surface synoptic (SYNOP), aircraft reports (AIREP), upper air (TEMP)] and nonconventional (GNSS, HRW AMV, radar reflectivity, and Doppler wind) observation types assimilated by AROME 3D-Var.
Citation: Bulletin of the American Meteorological Society 99, 7; 10.1175/BAMS-D-16-0321.1
The assimilation of ASCAT soil moisture (Wagner et al. 2013) by using an extended Kalman filter approach has been extensively investigated; its impact on regional precipitation forecasts shows a slight improvement for ALADIN precipitation forecasts (Schneider et al. 2014). Currently, ZAMG is testing it with the soil water index (SWI) and land surface temperature assimilation.
Regional ensemble prediction.
Most of the NHMSs in Europe already run their deterministic LAMs on scales of around 5 km or smaller. However, in unstable situations with limited predictability, the small-scale uncertainties may grow quickly to the synoptic scale. This could lead to different weather scenarios even in the short-range forecast. The sensitivity of local weather to ICs becomes more pronounced with increasing resolution and therefore should be considered in NWP. Moreover, sensitivity to the input information and to model accuracy is more important in cases of high-impact weather. Ensemble methods that appropriately simulate uncertainties in the forecast chain are thus very important.
A strategy for perturbing surface ICs in a regional EPS, called noncycling surface breeding (NCSB), was proposed by Wang et al. (2010). It uses short-range surface forecasts driven by perturbed atmospheric forcing. Belluš et al. (2016) implemented ensemble surface data assimilation (ESDA) to perturb the surface ICs in ALADIN-LAEF and compared it to NCSB. They showed that ensemble surface data assimilation is more skillful than NCSB, particularly for screen-level parameters.
The simulation of the ICs’ uncertainty in the upper air by blending in a regional EPS was proposed by Wang et al. (2014). It combines the large-scale initial perturbations provided by the global ensemble with the small-scale perturbations generated by the regional ensemble. The upper-air blending (Derková and Belluš 2007) also benefits from the spectral character of the ALADIN model. The IC perturbations generated by blending can represent well both large-scale and small-scale uncertainties in the analysis and are more consistent with the LBC perturbations provided by a global EPS.
Weidle et al. (2016) investigated the impact of large-scale perturbations of the ICs and LBCs obtained from different global EPSs on the forecast quality of a regional EPS. They found that the best global EPS might not always provide the best perturbations of IC and LBC for a regional ensemble system.
Model uncertainty is currently simulated by 16 combinations of different microphysics, deep and shallow convection, radiation, and turbulence schemes. Because of maintenance issues, these will be reduced to only four different sets in the future. An additional spread will be added by stochastic perturbation of physics tendencies (SPPT; Palmer et al. 2009), which was already tested in ALADIN-LAEF and AROME EPS. The perturbation of partial physics tendencies is being developed and is showing rather promising results. It contains an interaction of uncertainties among the radiation, turbulence, shallow convection, and microphysics schemes. Energy imbalance, an often-mentioned shortcoming of the SPPT approach, is reduced, and inherent numerical instabilities are avoided. Publications about this topic are currently in preparation.
Recently, more effort has been put into convection-permitting ensemble AROME EPS with 2.5-km horizontal resolution. AROME EPS was tested and evaluated in detail and was found to outperform its mesoscale ALADIN-LAEF counterpart in cases of precipitation forecasts, especially for the mountains, with a smaller impact observed for the lowlands as well (Schellander-Gorgas et al. 2017).
ALADIN-LAEF was compared with the global ECMWF EPS to investigate the added value of regional to global EPS. Wang et al. (2012) concluded that a regional single-model-based ensemble with fewer members could have better skill scores for screen-level weather variables than the global EPS with a larger ensemble size, whereas it may have limitations when applied to upper-air weather variables. The performance of ALADIN-LAEF was also compared to other regional ensembles during the Beijing 2008 Olympic Summer Games and the Sochi 2014 Olympic and Paralympic Winter Games in the framework of the WMO research demonstration projects (Duan et al. 2012; Kiktev et al. 2017). It was shown that ALADIN-LAEF compares favorably with the regional ensemble systems developed by other NWP centers.
Figure 9 illustrates the added value of ALADIN-LAEF over downscaling of the corresponding 16 global ECMWF EPS members. For screen-level temperature and relative humidity, the ensemble spread is increased, being closer to the slightly decreased rmse of the ensemble mean. Reduced precipitation yields a smaller spread and bias. For all three fields, the diurnal cycle of errors is weaker, and the percentage of outliers is significantly reduced. The impact on upper-air fields is rather neutral (not shown).
Scores of screen-level temperature, relative humidity, and 6-h total precipitation vs forecast lead time, calculated against SYNOP observations for ECMWF EPS downscaling (black) and ALADIN-LAEF (red). Thin lines denote the 10% and 90% confidence intervals. Both ensembles use the same domain with a horizontal mesh size of 5 km, which is intended for future operational use. Results from a 17-day period in May 2016 are shown, with forecasts starting at 1200 UTC. The verification area is the black domain in Fig. 5.
Citation: Bulletin of the American Meteorological Society 99, 7; 10.1175/BAMS-D-16-0321.1
CONLUSIONS AND CHALLENGES.
The journey of RC LACE since its conception has led to many fruitful experiences. RC LACE has established an extensive regional scientific and technical collaboration in the field of NWP for weather research, forecasting, and applications. It has demonstrated that the demands of developing and maintaining a complex operational LAM system are best met not by one small- or medium-sized country but by several partners working together on both technical infrastructure and scientific issues. RC LACE was developed as a result of small teams strongly focused on various aspects of NWP.
In the last 27 years, RC LACE has contributed to the development of the ALARO and AROME canonical model configurations of the ALADIN system in the areas of data assimilation, model dynamics and physics, regional ensemble prediction, and verification. RC LACE members exchange their national observations in real time, run a common preprocessing observation system for data assimilation and verification, and operate a common mesoscale ensemble system over all of Europe for early warning of potential severe weather. The RC LACE objectives have been successfully achieved by cooperation: sharing data, expertise, and technical capacities; creating synergy; combining and optimizing human and financial resources; increasing efficiencies; and avoiding duplicate efforts.
RC LACE clearly brings significant benefits to its members. However, it is not without its challenges. First, the limited manpower in RC LACE is spread thinly over many tasks mostly representing individual national interests and operational demands. Moreover, national technical capacities and computer resources vary among the individual NHMSs depending on the economic status of the respective countries. Working on scientifically promising but resource-intensive issues, such as convection-permitting data assimilation and ensembles, may become unaffordable for some RC LACE NHMSs.
Second, to ensure the healthy development of the consortium, it is necessary to make compromises between individual national interests and consortium benefits. It requires good planning and commitment of the individual teams. Efforts invested in common actions must be efficiently organized.
Third, data policy is needed to negotiate and agree on a scheme that satisfies all partner NHMSs, because several NHMSs have to protect their commercial revenues from the NWP products provided by others. Further, the intellectual property rights (IPR) of the NWP system among the members is another complex issue that should be fairly established for the NHMSs involved. RC LACE shares the intellectual property rights of the ALADIN system with other ALADIN partners based on the workforce contribution from the beginning of the ALADIN cooperation.
Fourth, there is always an active discussion on the benefits and drawbacks of decentralizing operational tasks. There is no doubt that decentralization can bring significant advantages, for example, that national interests and requirements are first met. However, some NHMSs are running similar configurations on comparable domains.
Last but not least, the ALADIN and HIRLAM consortia plan to merge into a single consortium consisting of 26 NHMSs until 2020. Effective cooperation in the merged consortium becomes a new challenge for RC LACE.
By addressing these challenges, the 27 years of experience of RC LACE have clearly demonstrated the value of regional cooperation among small- and medium-sized NHMSs in its successful development of a modern numerical forecasting system and in knowledge transfer and capacity building.
ACKNOWLEDGMENTS
RC LACE is grateful to Météo-France for its generous support from the beginning, and we thank Jean-François Geleyn for his great help on RC LACE. Further, we acknowledge all the RC LACE colleagues of Austria, Croatia, the Czech Republic, Hungary, Romania, Slovakia, and Slovenia for their contributions to the consortium. We give special thanks to the ALADIN and HIRLAM colleagues, in particular the program managers, Piet Termonia and Jeanette Onvlee-Hooimeijer, for the successful cooperation over the years. This study was funded by Zentralanstalt für Meteorologie und Geodynamik, the Meteorological and Hydrological Service of the Republic of Croatia, the Czech Hydrometeorological Institute, the Hungarian Meteorological Service, the National Meteorological Administration of Romania, the Slovak Hydrometeorological Institute, and the Slovenian Environment Agency.
REFERENCES
Bašták Ďurán, I., J.-F. Geleyn, and F. Váňa, 2014: A compact model for the stability dependency of TKE production–destruction–conversion terms valid for the whole range of Richardson numbers. J. Atmos. Sci., 71, 3004–3026, https://doi.org/10.1175/JAS-D-13-0203.1.
Belluš, M., Y. Wang, and F. Meier, 2016: Perturbing surface initial conditions in a regional ensemble prediction system. Mon. Wea. Rev., 144, 3377–3390, https://doi.org/10.1175/MWR-D-16-0038.1.
Bénard, P., 2003: Stability of semi-implicit and iterative centered-implicit time discretizations for various equation systems used in NWP. Mon. Wea. Rev., 131, 2479–2491, https://doi.org/10.1175/1520-0493(2003)131<2479:SOSAIC>2.0.CO;2.
Bénard, P., R. Laprise, J. Vivoda, and P. Smolíková, 2004: Stability of leapfrog constant-coefficients semi-implicit schemes for the fully elastic system of Euler equations: Flat-terrain case. Mon. Wea. Rev., 132, 1306–1318, https://doi.org/10.1175/1520-0493(2004)132<1306:SOLCSS>2.0.CO;2.
Bénard, P., J. Mašek, and P. Smolíková, 2005: Stability of leapfrog constant-coefficients semi-implicit schemes for the fully elastic system of Euler equations: Case with orography. Mon. Wea. Rev., 133, 1065–1075, https://doi.org/10.1175/MWR2907.1.
Bénard, P., J. Vivoda, J. Mašek, P. Smolíková, K. Yessad, Ch. Smith, R. Brožková, and J.-F. Geleyn, 2010: Dynamical kernel of the Aladin–NH spectral limited-area model: Revised formulation and sensitivity experiments. Quart. J. Roy. Meteor. Soc., 136A, 155–169, https://doi.org/10.1002/qj.522.
Bloom, S. C., L. L. Takacs, A. M. da Silva, and D. Ledvina, 1996: Data assimilation using incremental analysis updates. Mon. Wea. Rev., 124, 1256–1271, https://doi.org/10.1175/1520-0493(1996)124<1256:DAUIAU>2.0.CO;2.
Bölöni, G., 2005: Operational implementation of ALADIN 3DVAR at the Hungarian Meteorological Service. ALADIN Newsletter, No. 28, CNRM, Toulouse, France, 60–65, www.umr-cnrm.fr/aladin-old/newsletters/news28/PAPERS/BOLONI.pdf.
Bölöni, G., 2006: Development of a variational data assimilation system for a limited-area model at the Hungarian Meteorological Service. Idojaras, 110, 309–327.
Brožková, R., and Coauthors, 2001: DFI blending: An alternative tool for preparation of the initial conditions for LAM. PWRP Rep. 31 (CAS/JSC WGNE Rep.), WMO/TD-1064, 1–7.
Brožková, R., M. Derková, M. Belluš, and A. Farda, 2006: Atmospheric forcing by ALADIN/MFSTEP and MFSTEP oriented tunings. Ocean Sci., 2, 113–121, https://doi.org/10.5194/os-2-113-2006.
Bubnová, R., G. Hello, P. Bénard, and J.-F. Geleyn, 1995: Integration of the fully elastic equations cast in the hydrostatic pressure terrain-following coordinate in the framework of the ARPEGE/Aladin NWP system. Mon. Wea. Rev., 123, 515–535, https://doi.org/10.1175/1520-0493(1995)123<0515:IOTFEE>2.0.CO;2.
Bučánek, A., A. Trojáková, and R. Brožková, 2015: The BlendVar assimilation scheme at CHMI. Meteor. Zpr., 68, 180–185.
Catry, B., J.-F. Geleyn, M. Tudor, P. Bénard, and A. Trojáková, 2007: Flux-conservative thermodynamic equations in a mass-weighted framework. Tellus, 59A, 71–79, https://doi.org/10.1111/j.1600-0870.2006.00212.x.
Courtier, P., and J.-F. Geleyn, 1988: A global numerical weather prediction model with variable resolution: Application to the shallow water equations. Quart. J. Roy. Meteor. Soc., 114, 1321–1346, https://doi.org/10.1002/qj.49711448309.
Csomós, P., and G. Bölöni, 2009: First steps towards the application of the ensemble transform Kalman filter technique at the Hungarian Meteorological Service. HIRLAM Newsletter, No. 54, KNMI, De Bilt, Netherlands, 9–19, http://hirlam.org/index.php/publications-54/hirlam-newsletters-a/doc_download/127-hirlam-newsletter-no-54-paper-02-csomos.
Derková, M., and M. Belluš, 2007: Various applications of the blending by digital filter technique in the ALADIN numerical weather prediction system. Meteor. Cas., 10, 27–36.
Duan, Y., and Coauthors, 2012: An overview of the Beijing 2008 Olympics Research and Development Project (B08RDP). Bull. Amer. Meteor. Soc., 93, 381–403, https://doi.org/10.1175/BAMS-D-11-00115.1.
Field, P., and Coauthors, 2017: Exploring the convective grey zone with regional simulations of a cold air outbreak. Quart. J. Roy. Meteor. Soc., 143, 2537–2555, https://doi.org/10.1002/qj.3105.
Fischer, C., T. Montmerle, L. Berre, L. Auger, and S. Stefanescu, 2005: An overview of the variational assimilation in the ALADIN/France numerical weather prediction system. Quart. J. Roy. Meteor. Soc., 131, 3477–3492, https://doi.org/10.1256/qj.05.115.
Geleyn, J.-F., B. Catry, Y. Bouteloup, and R. Brožková, 2008: A statistical approach for sedimentation inside a micro-physical precipitation scheme. Tellus, 60A, 649–662, https://doi.org/10.1111/j.1600-0870.2008.00323.x.
Geleyn, J.-F., J. Mašek, R. Brožková, P. Kuma, D. Degrauwe, G. Hello, and N. Pristov, 2017: Single interval longwave radiation scheme based on the net exchanged rate decomposition with bracketing. Quart. J. Roy. Meteor. Soc., 143, 1313–1335, https://doi.org/10.1002/qj.3006.
Gerard, L., J.-M. Piriou, R. Brožková, J.-F. Geleyn, and D. Banciu, 2009: Cloud and precipitation parameterization in a meso-gamma-scale operational weather prediction model. Mon. Wea. Rev., 137, 3960–3977, https://doi.org/10.1175/2009MWR2750.1.
Giard, D., and E. Bazile, 2000: Implementation of a new assimilation scheme for soil and surface variables in a global NWP model. Mon. Wea. Rev., 128, 997–1015, https://doi.org/10.1175/1520-0493(2000)128<0997:IOANAS>2.0.CO;2.
Guidard, V., and C. Fischer, 2008: Introducing the coupling information in a limited-area variational assimilation. Quart. J. Roy. Meteor. Soc., 134, 723–735, https://doi.org/10.1002/qj.215.
Guidard, V., C. Fischer, M. Nuret, and A. Džiedžic, 2006: Evaluation of the ALADIN 3D-VAR with observations of the MAP campaign. Meteor. Atmos. Phys., 92, 161–173, https://doi.org/10.1007/s00703-005-0156-5.
Haiden, T., A. Kann, C. Wittmann, G. Pistotnik, B. Bica, and C. Gruber, 2011: The Integrated Nowcasting through Comprehensive Analysis (INCA) system and its validation over the eastern Alpine region. Wea. Forecasting, 26, 166–183, https://doi.org/10.1175/2010WAF2222451.1.
Hortal, M., 2002: The development and testing of a new two-time-level semi-Lagrangian scheme (SETTLS) in the ECMWF forecast model. Quart. J. Roy. Meteor. Soc., 128, 1671–1687, https://doi.org/10.1002/qj.200212858314.
Jones, C. D., and B. Macpherson, 1997: A latent heat nudging scheme for the assimilation of precipitation data into an operational mesoscale model. Meteor. Appl., 4, 269–277, https://doi.org/10.1017/S1350482797000522.
Kalnay, E., 2003: Atmospheric Modeling, Data Assimilation and Predictability. Cambridge University Press, 341 pp.
Kiktev, D., and Coauthors, 2017: FROST-2014: The Sochi Winter Olympics international project. Bull. Amer. Meteor. Soc., 98, 1908–1929, https://doi.org/10.1175/BAMS-D-15-00307.1.
Lafore, J. P., and Coauthors, 1998: The Meso-NH atmospheric simulation system. Part I: Adiabatic formulation and control simulations. Ann. Geophys., 16, 90–109, https://doi.org/10.1007/s00585-997-0090-6.
Laprise, R., 1992: The Euler equations of motion with hydrostatic pressure as an independent variable. Mon. Wea. Rev., 120, 197–207, https://doi.org/10.1175/1520-0493(1992)120<0197:TEEOMW>2.0.CO;2.
Leutbecher, M., and T. N. Palmer, 2008: Ensemble forecasting. J. Comput. Phys., 227, 3515–3539, https://doi.org/10.1016/j.jcp.2007.02.014.
Mašek, J., J.-F. Geleyn, R. Brožková, O. Giot, H. O. Achom, and P. Kuma, 2016: Single interval shortwave radiation scheme with parameterized optical saturation and spectral overlaps. Quart. J. Roy. Meteor. Soc., 142, 304–326, https://doi.org/10.1002/qj.2653.
Meier, F., 2015: Radar assimilation and latent heat nudging in AROME Nowcasting/AROME-PI. RC LACE Stay Rep., 22 pp., www.rclace.eu/File/Data_Assimilation/2015/reportLACEstaymeierautumn2015.pdf.
Mesinger, F., 2001: Limited area modeling: Beginnings, state of the art, outlook. 50th Anniversary of Numerical Weather Prediction Commemorative Symposium, A. Spekat, Ed., Deutsche Meteorologische Gesellschaft, 91–118.
Mile, M., and R. Randriamampianina, 2012: The impact of high resolution winds on the ALADIN Hungary model. Proc. 2012 EUMETSAT Meteorological Satellite Conf., Sopot, Poland, EUMETSAT, 8 pp., www.eumetsat.int/website/wcm/idc/idcplg?IdcService=GET_FILE&dDocName=PDF_CONF_P61_S7_03_MILE_V&RevisionSelectionMethod=LatestReleased&Rendition=Web.
Mile, M., G. Bölöni, R. Randriamampianina, R. Steib, and E. Kucukkaraca, 2015: Overview of mesoscale data assimilation developments at the Hungarian Meteorological Service. Idojaras, 119, 215–239.
Palmer, T., R. Buizza, F. Doblas-Reyes, T. Jung, M. Leutbecher, G. Shutts, M. Steinheimer, and A. Weisheimer, 2009: Stochastic parametrization and model uncertainty. ECMWF Tech. Memo. 598, 42 pp., www.ecmwf.int/sites/default/files/elibrary/2009/11577-stochastic-parametrization-and-model-uncertainty.pdf.
Radnóti, G., and Coauthors, 1995: The spectral limited area model ARPEGE/ALADIN. PWRP Rep. 7, 111–117.
Randriamampianina, R., 2006: Investigation of the AMV data derived from Meteosat-8 in the ALADIN/HU data assimilation system. Proc. Eighth Int. Winds Workshop, Bejing, China, EUMETSAT, 8 pp., www.eumetsat.int/website/wcm/idc/idcplg?IdcService=GET_FILE&dDocName=PDF_CONF_P47_S2_06_RANDRIAMA_V&RevisionSelectionMethod=LatestReleased&Rendition=Web.
Schellander-Gorgas, T., Y. Wang, F. Meier, F. Weidle, Ch. Wittmann, and A. Kann, 2017: On the forecast skills of a convection-permitting ensemble. Geosci. Model Dev., 10, 35–56, https://doi.org/10.5194/gmd-10-35-2017.
Schneider, S., Y. Wang, W. Wagner, and J. F. Mahfouf, 2014: Impact of ASCAT soil moisture assimilation on regional precipitation forecasts: A case study for Austria. Mon. Wea. Rev., 142, 1525–1541, https://doi.org/10.1175/MWR-D-12-00311.1.
Seity, Y., P. Brousseau, S. Malardel, G. Hello, P. Bénard, F. Bouttier, C. Lac, and V. Masson, 2011: The AROME-France convective-scale operational model. Mon. Wea. Rev., 139, 976–991, https://doi.org/10.1175/2010MWR3425.1.
Simmons, A. J., and D. M. Burridge, 1981: An energy and angular-momentum conserving vertical finite-difference scheme and hybrid vertical coordinates. Mon. Wea. Rev., 109, 758–766, https://doi.org/10.1175/1520-0493(1981)109<0758:AEAAMC>2.0.CO;2.
Široká, M., C. Fischer, V. Casse, R. Brozkova, and J.-F. Geleyn, 2003: The definition of mesoscale selective forecast error covariances for a limited area variational analysis. Meteor. Atmos. Phys., 82, 227–244, https://doi.org/10.1007/s00703-001-0588-5.
Strajnar, B., 2008: Implementation and test of spatially varying (climatological and flow-dependent) background error variances in ALADIN. Météo-France Stay Rep., 19 pp., www.umr-cnrm.fr/aladin/IMG/pdf/pdf_STRAJNAR.pdf.
Strajnar, B., 2012: Validation of Mode-S Meteorological Routine Air Report aircraft observations. J. Geophys. Res., 117, D23110, https://doi.org/10.1029/2012JD018315.
Strajnar, B., N. Žagar, and L. Berre, 2015: Impact of new aircraft observations Mode-S MRAR in a mesoscale NWP model. J. Geophys. Res. Atmos., 120, 3920–3938, https://doi.org/10.1002/2014JD022654.
Termonia, P., and Coauthors, 2018: The ALADIN System and its canonical model configurations AROME CY41T1 and ALARO CY40T1. Geosci. Model Dev., 11, 257–281, https://doi.org/10.5194/gmd-11-257-2018.
UKMO, 2010: UK index documentation. Met Office Rep., 7 pp., www.wmo.int/pages/prog/www/DPFS/Meetings/CG-FV_Montreal2011/Doc5-5-Annex_UK_index_documentation_2010-1.pdf.
Untch, A., and M. Hortal, 2004: A finite-element scheme for the vertical discretization of the semi-Lagrangian version of the ECMWF forecast model. Quart. J. Roy. Meteor. Soc., 130, 1505–1530, https://doi.org/10.1256/qj.03.173.
Váňa, F., P. Bénard, J.-F. Geleyn, A. Simon, and Y. Seity, 2008: Semi-Lagrangian advection scheme with controlled damping: An alternative to nonlinear horizontal diffusion in a numerical weather prediction model. Quart. J. Roy. Meteor. Soc., 134, 523–537, https://doi.org/10.1002/qj.220.
Vivoda, J., and P. Smolíková, 2013: Finite elements used in the vertical discretization of the fully compressible forecast model ALADIN-NH. ALADIN–HIRLAM Newsletter, No. 1, KNMI, De Bilt, Netherlands, and IRM, Bruxelles, Belgium, 31–46, www.umr-cnrm.fr/aladin/IMG/pdf/aladin_hirlam_newsletter_1_september2013.pdf.
Wagner, W., and Coauthors, 2013: The ASCAT soil moisture product: A review of its specifications, validation results, and emerging applications. Meteor. Z., 22, 5–33, https://doi.org/10.1127/0941-2948/2013/0399.
Wang, Y., A. Kann, M. Belluš, J. Pailleux, and C. Wittmann, 2010: A strategy for perturbing surface initial conditions in LAMEPS. Atmos. Sci. Lett., 11, 108–113, https://doi.org/10.1002/asl.260.
Wang, Y., and Coauthors, 2011: The central European limited-area ensemble forecasting system: ALADIN-LAEF. Quart. J. Roy. Meteor. Soc., 137, 483–502, https://doi.org/10.1002/qj.751.
Wang, Y., S. Tascu, F. Weidle, and K. Schmeisser, 2012: Evaluation of the added values of regional ensemble forecast on global ensemble forecast. Wea. Forecasting, 27, 972–987, https://doi.org/10.1175/WAF-D-11-00102.1.
Wang, Y., M. Belluš, J. Geleyn, X. Ma, W. Tian, and F. Weidle, 2014: A new method for generating initial perturbations in regional ensemble prediction system: Blending. Mon. Wea. Rev., 142, 2043–2059, https://doi.org/10.1175/MWR-D-12-00354.1.
Wang, Y., and Coauthors, 2017: Integrating nowcasting with crisis management and risk prevention in a transnational and interdisciplinary framework. Meteor. Z., 26, 459–473, https://doi.org/10.1127/metz/2017/0843.
Weidle, F., Y. Wang, and G. Smet, 2016: On the impact of the choice of global ensemble in forcing a regional ensemble system. Wea. Forecasting, 31, 515–530, https://doi.org/10.1175/WAF-D-15-0102.1.
Yan, X., V. Ducrocq, P. Poli, G. Jaubert, and A. Walpersdorf, 2008: Mesoscale GPS zenith delay assimilation during a Mediterranean heavy precipitation event. Adv. Geosci., 17, 71–77, https://doi.org/10.5194/adgeo-17-71-2008.
Yang, X., 2005: Analysis blending using a spatial filter in grid-point model coupling. HIRLAM Newsletter, No. 48, KNMI, De Bilt, Netherlands, 49–55.
Advanced Microwave Sounding Unit-A (AMSU-A); Advanced Microwave Sounding Unit-B (AMSU-B); atmospheric motion vectors (AMVs); Advanced Scatterometer (ASCAT); Advanced Technology Microwave Sounder (ATMS); Cross-Track Infrared Sounder (CrIS); High Resolution Infrared Radiation Sounder (HIRS); Infrared Atmospheric Sounding Interferometer (IASI); Mode-S meteorological routine air report (MRAR); and Spinning Enhanced Visible and Infrared Imager (SEVIRI).