INSTANTANEOUS AND AUTOMATIC IDENTIFICATION OF WEATHER TYPES FOR A FOREST-FIRE DECISION-SUPPORT SYSTEM

The objective of this study, carried out as part of a European Union C ontract, was to elaborate an aid package that enables automatic and '' real time'' identification of the instantaneous weather situation duri ng seasons with high forest fire hazards. The weather types are define d on the basis of discrimination between the breeze and synoptic wind regimes established from a network of 21 automatic sensors installed i n the southeastern extremity of France. The establishment of rules suc h as: if at station X the wind is from a direction of 100-140 degrees and blows at 2 to 5 m/s; and if, etc..., then the situation is such - enables, in a logic of artificial intelligence, the proposal of a ''sy mbolical'', ''analog'' model of the weather behaviours to be expected over a given area.