“Karelian lakes” expert system: 1. Ordinal and nominal characteristics of lakes

An expert system is proposed for the assessment of characteristic and the trophic status of lakes for which little or no direct observational data are available. The system is based on the mathematical apparatus of fuzzy logic and applied to the best studied in Russia Karelian lake region. The created database on the morphometry, hydrology, hydrochemistry, and hydrobiology for 100 best studied lakes is described. The lakes are classified by methods of multidimensional statistics, factor, cluster, and logical-information analysis with the use of multidimensional scaling technique. Serious drawbacks of different classifications are demonstrated, in particular, because of the lack of a constructive algorithm for the recovery of unknown characteristics of a lake once it is referred to a certain class. In such cases, the mathematical apparatus of fuzzy sets and fussy logic should be used.