Cyanobacterial risk prevention under global warming using an extended Bayesian network

Abstract

Cyanobacterial blooms under global warming are increasing worldwide, producing emerging contaminants, which threaten the health of human beings and aquatic ecosystems. The health burdens warrant the development of a useful risk-assessment tool and a holistic preventive-control scheme to prevent cyanobacterial blooms. This paper aims to integrate cyanobacterial risk assessment and risk preventive control by investigating the relationships amongst cyanobacterial blooms and multi-dimensional influencing variables. Two challenges hinder such a task. First, the time-series variations in cyanobacteria and influencing variables are uncertain and nonlinear. Second, there rarely exists an explicit modelling framework for integrating cyanobacterial risk assessment and risk preventive control. This study builds an extended Bayesian network model and proposes an integrated framework with functions of assessment, inference, preventive control, and visualisation of the risk of cyanobacterial blooms. Field data from a tropical lake are used to evaluate the model and framework. The proposed model achieves better performance than the seven models in comparison. The cyanobacterial risk is anticipated to increase by 38.5% under global warming. On the contrary, guided by the model and framework, the risk could be reduced by about 60% by taking the identified risk preventive control scheme. The cyanobacterial risk prevention would reduce aquatic emerging contaminants in drinking and recreational water sources.