omsz
>>
  Hungarian Meteorological Service  founded: 1870
Research and development | Numerical Weather Prediction  | Analysis of the Atmospheric Environment | 
Felbontás: KicsiFelbontás: NormálFelbontás: KözepesFelbontás: NagyobbFelbontás: Nagy  

ALADIN

The ALADIN project is one of the most important international cooperation of the Hungarian Meteorological Service (HMS) in the field of numerical weather prediction (NWP). This project was initiated in 1990 by Meteo-France with the aim of developing a limited area model (LAM) for numerical weather prediction purposes and today 15 countries are participating in the common work. The HMS has been actively taking part in the ALADIN cooperation since its beginning. Detailed information on the ALADIN model can be found on the ALADIN web page. The history of the ALADIN project can be read on this web page. Hereafter we describe the basis of limited area modelling and the main characteristics of the ALADIN model version run at HMS. Finally we give an overview of the developments done at the NWP department.

Today, the prediction of local, mesoscale meteorological processes for the short range (up to 48 hours) is a basic task for the most of the meteorological services. Such predictions are provided by running limited area models (LAM). Limited area models give detailed, high resolution, forecasts for a relatively small geographical area. Joining the ALADIN project, HMS took the opportunity to take part in the development of the LAM ALADIN and also to run this model operationally for a convenient domain in Budapest. The local installation and operation is essential, because it gives freedom and flexibility, which is needed in order to satisfy the different requirements of the users.

The local operation gives the following advantages:

Freedom in the choice of the geographical domain, horizontal and vertical resolution

Freedom in the choice of the frequency of the model runs

It is possible to define several model domains nested in each other

Large choice of possible products with high frequency in time

Freedom in modifying the local system

Local practice and experience in modelling, possibility for education in this field


The LAM outputs are widely used. We list below the most important applications:

Prediction of extreme weather events, warnings with high spatial resolution and high accuracy in time (accurate forecasts are highly beneficial in economy)

Agricultural applications (frosts, hale, precipitation, etc.)

Traffic applications (freezing, snow)

Pollution prediction (tropospherical ozone concentration, diffusion of chemicals in case of industrial accidents)

Aviation (thunderstorms, freezing, turbulence, fog, etc.)

Hydrology (input for hydrological models, flood prediction)

Energy sector (prediction of needed energy for heating and electricity in the function of the temperature)

For those who are interested in limited area modelling, it is worth to visit the HIRLAM web page, which presents a project similar to ALADIN considering its goals. It is important to mention that the leader committees of both projects (HIRLAM and ALADIN) decided to cooperate in the future and coordinate together their research and development activity.

The ALADIN products are more and more used in the operational forecasting at HMS. At the same time there are more and more products that are generated automatically from the raw model outputs. The number of users is increasing too. Consequently, the quality of the NWP products is a key issue in order to satisfy the demands with high quality information.


The main characteristics of the ALADIN model

Spectral, hydrostatic limited area model (the non-hydrostatic version with efficient numerics is already available for testing and serves as a basis for the AROME model, which is planned to put into operations at 2008).

Application of fully harmonic functions (2 dimensional Fourier representation) for the horizontal spectral representation of the meteorological fields. The periodic extension of the functions is done using a so-called extension zone (see the figure below).

Advanced parametrizations of atmospheric physical processes (radiation, convection, boundary layer processes, turbulence, microphysics, etc.)

Lateral boundary conditions provided by the ARPEGE global model or a larger ALADIN version. The lateral boundary conditions ensure the propagation of meteorological information into the limited area domain from outside (coupling, relaxation). The relaxation zone (see the figure below) ensures the smooth transition of the meteorological parameters towards the lateral boundaries.

The initial conditions of the model can be provided by running local data assimilation (3D-VAR, OI) or by interpolating the ARPEGE initial conditions onto the LAM grid.

The initialization (the filtering of gravity waves from the initial conditions which are not consistent with the simplified model equations) is done by digital filtering.

The ALADIN models domain
 The ALADIN models domain
Further interesting information on the ALADIN model can be picked up from the following article (in Hungarian):

Kullmann L., 2003: Az ALADIN mezoskálájú numerikus modell-rendszer, Meteorológiai Tudományos Napok 29. kiadványa, pp. 104-118 (pdf)


Operative practice


In the every day operations a model run consists of the following steps:

Interpolation of the lateral boundary conditions (and the initial conditions if there is no local data assimilation) onto the target LAM geometry.

Subtraction of the orography and climatological information from an external "climatology file", which can be created by a configuration of the ALADIN model using raw information from US NAVY database. The climatology files have to be created only once for one domain for each month of the year.

Initialisation of the initial conditions (digital filter).

Run the forecast.

Interpolation of the output fields to a geographical coordinate system and perform file conversions according to the given application.

The ALADIN model is run twice a day starting from 00 and 12 UTC. It is important, that the forecast starts as soon as the lateral boundary conditions are available, in order to provide the information in time for the forecasters. The model is run by an automatic system (shell scripts, cron daemon), so basically it does not require any human intervention. However, in case of any technical problems or missing lateral boundary conditions, the model has to be run by direct intervention. Consequently, a control system has been set up which consists of 24 hour supervision of the model runs by operators. If the forecast fails, the operator calls the actual colleague from the NWP department being in shift who should connect the machine of HMS using a mobile phone and a laptop and should solve the problem.

The presently operational ALADIN version used at HMS:

Geographical domainContinental Europe
Horizontal resolution8 km
Number of vertical levels49
Initial conditions 3D-VAR local assimilation
Lateral boundary conditionsARPEGE global model
Coupling frequency3 hours
Production forecast range48 hours
  


The operational model run is performed
on an


Visualization and post processing

Forecasters use a complex visualization system developed at HMS (HAWK-2) in order to interpret the NWP model's outputs. They also can combine the model fields with other information like radar or satellite observations with this system. Modern visualization systems make possible to modify the NWP prediction fields as well. Model outputs are post processed by statistical and dynamical adaptation methods as well for those meteorological fields, which are rather poorly represented by NWP models generally, due to the fact that such model's orography and surface characteristics are not perfect.

The ALADIN/HU domain and orography
The ALADIN/HU domain and orography

The visualization of ALADIN fields in HAWK2
The visualization of ALADIN fields in HAWK2

Forecasted pseudo satellite image
Forecasted pseudo satellite image


RC LACE research project

It is important to mention the RC LACE (Regional Cooperation for Limited Area Modeling in Central Europe) project, which is a sub-cooperation inside the ALADIN consortia including HMS. The aim of this cooperation is a commonly coordinated research activity between the member services. The HMS plays an important role in this cooperation while coordinating the research in the field of data assimilation and the short range probabilistic forecasting (LAMEPS). Our colleagues are also taking part in the development of the model physiscs and dynamics as well. Further information on RC LACE can be found on the RC LACE website.

The most important scientific and strategic plans for the ALADIN project

The further development of the variational data assimilation (3d-var) scheme with special emphasis on the inclusion of new observational types and background error modelling. In the more distinct future the extension of the three-dimensional scheme towards four dimension is expected (3dfgat: 3d first guess at appropriate time, then 4d-var).

Active participation on the development of AROME model (initiated by Météo France). The AROME model as successor of the ALADIN model is based on the non-hydrostatic dynamics and the 3d-var data assimilation system of the ALADIN model and the physical parametrisation package of the meso-NH model (developed by the French research community). The AROME non-hydrostatic model is planned to put into operations at 2008 for the precise and reliable prediction of small scale dangerous weather events (also for nowcasting purposes).

The development of short range ensemble system. The project started at the end of 2003 is aiming at the extension of ensemble approach for short range with the capabilities of providing probabilistic weather forecasts for poorly predictable high impact weather situations. The Hungarian Meteorological Service has a leading role in the coordination and execution of the LAMEPS activities inside the ALADIN project.

Increased (code) collaboration with the HIRLAM project in mesoscale numerical modelling. The common work with HIRLAM is aiming a better distribution of resources with efficient coordination in order to achieve the objectives described above.