You are here:
A short introduction to the application of Bayesian networks in ecological risk assessment and decision making
Citation:
Landis, W. AND John F. Carriger. A short introduction to the application of Bayesian networks in ecological risk assessment and decision making. To be Presented at SETAC North America 40th Annual Meeting, TorontoC, November 03 - 07, 2019.
Impact/Purpose:
This poster will be the first poster in a Society of Environmental Toxicology and Chemistry annual North America meeting poster session on Bayesian networks and environmental assessment and management. The purpose is to introduce the sessions and Bayesian networks to visitors to the poster session.
Description:
Summary Bayesian networks (BNs, also known as Bayesian belief networks) are a means of describing cause and effect using directed acyclic graphs where the interactions between the nodes are described using conditional probability tables. That is a way of saying that that there are nodes (steps in a cause-effect model),which are connected by lines of influence, and the interactions are described using probability tables that take into account all possible combinations of inputs to generate the probability distributions for each possible outcome. Bayesian networks inherently incorporate uncertainty and can use wide varieties of data, from precise determinations of molecular interactions to expert elicitation. Since 2010 a number of papers have been published on the use of BNs to evaluate management options, to evaluate risk, to provide a more quantitative framework for the relative risk model, and to become a part of adaptive management. At the 2017 SETAC NA (Sacramento) and the 2018 SETAC EU (Helsinki) These sessions included presentations by developers and users of Bayesian networks to estimate risk, calculate the outcomes of management alternatives and the use of such techniques to make policy decisions. The topics included large scale risk assessments, BNs in QSAR models, the use of BNs to build quantitative adverse outcome pathways, sustainable management of large scale systems, and the management of environmental resources. This session will include presentations integrating the use of pesticides and synthetic biology, multiple stressors, and other applications. The question will be where do we go now that the use of Bayesian networks is part of the mainstream of risk assessment and decision making.