Diagnosing reservoir water quality using self-organizing maps and fuzzy theory

Abstract

Since trophic status assessment of water quality is very important for the water resources management, the assessment results obtained from using only one parameter may easily mislead or bias the decision makers or managers. Even when using a multivariable index, how to determine the weights of all factors is debatable. In this research, one complementary evaluation method, self-organizing map (SOM), for diagnosing water quality has been used to develop a trophic state classifier and is illustrated with a case study of trophic status assessment for Fei-Tsui Reservoir in Taiwan. The historical database was collected from the management agency of Fei-Tsui Reservoir from 1987 to 1995. The results of SOM are compared with those of the Carlson index and Fuzzy synthetic evaluation, showing that the inconsistent records can be mapped to the conflicting data zone of the SOM output map. In addition, SOM creates a diagnostic axis on the map to express the trophic status of the water body. As long as the SOM model is well-trained, new records can be assessed and classified as either one of three trophic levels or speciarcases. If special water quality conditions are expressed on the SOM output, those data can reveal that either total phosphorus (TP) or chlorophyll A (chl a) is higher than usual.