• María Jesús Souto Alvedro. CESGA.
• Francisco Landeira Vega. CESGA.
• Enrique Roca Bordello. CMA

The aim of this project is to adapt different meteorological modelation solutions to the needs of the  General Management of the Centre of Environmental Information and Technology of  the Government of Galicia (Xunta de Galicia) to CESGA concerning meteorology and air quality.

This project includes the study, adaptation and rehersal of a meteorological prediction model for Galicia, which can be used to elaborate its own meteorological prediction, becoming the Numerical Meteorological Prediction System of Galicia.

1.Analysis and model selection

In order to analyze and select the models, it was necessary to study Galicia's environment features to determine the most adequate numerical solution for this region. Later, a search for the existing solutions was prepared ( taking into account the ones offered by the General Management ). These solutions should be safe and have no additional cost.

Galicia's orography, variety of oil usages and geographic location , as well as the entrance of high associated energy fronts coming from the Atlantic side, favour the existence of important pressure gradients, which cannot be represented if:

a) The model chosen has an inferior resolution to the one corresponding to the spatial variability of Galicia's meteorological phenomena.

b)The equations incorporate the hydrostatic hypothesis, which, theoretically, includes a lineal pressure gradient, if additional approximate solutions which allow to describe the spatial variability of the gradient are not incorporated.
 
Considering this, the model chosen to rehearse the prediction and the model ARPS (Advanced Regional Prediction System), developed in the Centre for Analysis and Prediction of Storms (CAPS) from the University of Oklahoma, which incorporates  non-hydrostatic advanced solutions applicable to extreme meteorological conditions (including cyclons). On top of this, it has code strenght and is freely available with no limit of usage, which makes it an interesting choice to be applied in Galicia. It also has a lot of computing options to be chosen by the user, and which include, for instance, many environment conditions treatment possibilities, initial data, estimation of the advection terms, cloud parametrization schemes, turbulence proximity or the possibility to parametrize water in the solid, liquid and gaseous states.

2. Development of digital cartographical information

To execute the model ARPS, the digital cartographical information needed was developed. The model itself specified the format, which covers the whole autonomous region of Galicia, with a 5 Km horizontal resolution. The information desvribes topographic heights, and makes a classification of the covers (soil, sea and inland aquifers).

This information, developed from the MDT200 has to provide enough resolution to cover the micro-mesoscale meteorological phenomena of Galicia. However, it should not require prohivitive computing times, for, that way, the application of the model would be unfeasible. Both hypotheses will have to be checked in the rehearsals during this project.

Figure 1 represents Galicia's topography incorporated to the model ARPS. Several kilometres of sea were included in west and east zones to simulate the effects of the sea in coastal areas.

3. Adaptation of the model ARPS

To adapt the model ARPS to CESGA's computers, different stages were followed: firstly, the code of the model itself was compiled, and secondly, another compilation was made for the complementary programmes and the necessary graphic packages to represent and analyze the results.

After the changes in the Makefile and the code ARPS, the programme was compiled in the paralell vector computer Fujitsu VPP300E and the paralell scalar computer AP3000 (both owned by CESGA) and the first prediction test was executed with a simple example supplied by the User Manual. Although this was a simple example (the variables were iniciated through just one probing and with no orography) it was useful to check that the computing time in the scalar compute is unfeasible when such a complex model as ARPS is processed. Therefore, from that moment on, the successive tests were made in the vectorial computer.

4. Rehearsing the model ARPS

4.1. Validation

The validation of the model is essential to check out if the results obtained in our computers are similar to those obtained, in this case, in the CAPS, and to make sure that no errors are being made in the prediction just because they are processed in CESGA's computers. Thus, the manual ARPS offers several results for standard cases which were compared to with the response of the computer Fujitsu VPP300E, getting satisfactory results.

4.2 Rehearsing in Galicia

Several application rehearsals were made for the community of Galicia, which include both an study of the response of the meteorological prediction model ARPS  to Galicia's topography and the adaptation of the graphic interfaces ZXPLOT and Vis5d to the new environment.

The computing domain has 50'50 points and it includes all of Galicia's land cover and approximately 30 km of North and West sea areas. The horizontal space covers 5 kms in direction North-South and East-West. Therefore, the predictions will be made in an area of 250x 250 km², as figure 3.1 shows.