A Hybrid Model Tree (MT) — Genetic Algorithm (GA) Scheme for Toxic Cyanobacteria Predictions in Lake Kinneret

This paper presents a hybrid model tree (MT) — genetic algorithm (GA) scheme for toxic Cyanobacteria predictions in Lake Kinneret. Lake Kinneret (the Sea of Galilee) is the most important surface water resource in Israel providing approximately 35% of its annual drinking water, a proportion that is constantly increasing. For more than 30 years there have been no major problems with respect to the water quality of the lake. However, the appearance of toxic Cyanobacteria blooms in 1994 suggests that the future water quality of the lake might be at risk. A full physical understanding of the reasons for the toxic Cyanobacteria blooms is lacking. This study suggests a data driven modeling approach, relying on the vast existing data base of the lake, to explore the possible major factors causing the toxic Cyanobacteria to bloom, and to predict their possible appearance.