Qualitative Modeling and Data Mining In the Monitoring of the Selected Ecosystem

The Trebon region in the South Bohemia is known with its picturesque landscape and with its ponds. However the Trebon ecosystem is violated by continuous parasitic dehumidifying and dehydrating which attacks the Small Water Cycle (SWC). In our mentioned ecosystem is evaporated water quickly brought up in the zone in which does not condense yet and in this height zone is transported outside the ecosystem to distanced mountains where spontaneously condenses in rising air streams. As a basic reason of these facts are usually introduced two phenomena: • The overheating of the air above landscape surface as a result of the solar radiation phenomena on the extensive reflex areas (corn fields, roofs of towns, etc.). • The increased transport velocity in non condensation zone. Seemed to be, that the extension of vegetation fields (that makes evaporation slower) and the decrease of reflex areas surface (that will be rather difficult in towns), could be sufficient for improvement the situation. The proposed paper that belongs to the project (Pokorný et al. 2008), deals with another possible ways yet. (The goals of the project (Pokorný et al. 2008) are concentrated on the description of the energy flows and on the development of biodiversity in the ecosystem.) The paper proposes two data mining methods used for modeling and monitoring of the selected ecosystems. The first method works with qualitative state diagram and leads to the detection of unexpected situations (Bíla and Jura 2007). The second method extracts the matroid structure on the set of states and works with implications from matroid bases. The methodology is developed with data support of a real ecosystem. The frame structure of the measurement centers network and the associated structure of the system database is introduced. Key references: [1] Pokorný, J., Jirka, V., Pechar, L., Bíla, J., Jura, J., Hofreiter, M., Petrová, R., Zicha, J., Kobrzek F. and Mareček, J. (2008). Partial Report to project 2B 06023, MSMT, Czech Republic. [2] Bila J. and Jura, J. (2007). Fuzzy Concepts in the Detection of Unexpected Situations. Acta Polytechnica. Vol. 47, No.1. pp. 5-8.