INTEGRATION OF BP ARTIFICIAL NEURAL NETWORK AND FUZZY THEORY ON EVALUATING CYANOBACTERIA-DOMINANT BLOOM OCCURRENCE RISK OF LAKE TAIHU

Cyanobacteria-dominant bloom is one of the most serious environmental problems both in China and all the world.However,the algal bloom outbreak mechanisms are complex with obvious uncertainty.This study aimed to develop an approach to evaluate algal bloom occurrence risk in Lake Taihu.Firstly,BP artificial neural network was applied to reveal the relations of algal bloom and the impact factors in order to evaluate the status of algal bloom.Then,according to these results,the fuzzy theory was integrated to deeply evaluate algal bloom occurrence probability with different possibility,which would greatly quantify the uncertainty of algal bloom outbreak.Finally,the composite algal bloom occurrence risk of Lake Taihu was assessed based on the recent fine years' monitoring data with 65 sites.The results showed that the west coast area and MeiLiang Bay are of the highest risk,Lake Zhushan and Lake Wuli are of moderate risk,the southern coast area,Lake Gonghu and the central region of Lake Taihu are of low risk,Middle East area and East Lake are of little risk.The methodology developed and results can be utilized to support further study on algal bloom forecast and water resource management.