Integrating Science Through Bayesian Belief Networks: Case Study of Lyngbya in Moreton Bay
Eva G. Abal, Clair Alston‐Knox, Tony Chiffings, Grant S. Hamilton, Barry T. Hart, Kerrie Lee Mengersen · 2005
Bayesian Belief Networks (BBNs) are emerging as valuable tools for investigating complex ecological problems. In a BBN, the important variables in a problem are identified and causal relationships are represented graphically. Underpinning this is the probabilistic framework in which variables can take on a finite range of mutually exclusive states. Associated with each variable is a conditional probability table (CPT), showing the probability of a variable attaining each of its possible states conditioned on all possible combinations of it parents. Whilst the variables (nodes) are connected, the CPT attached to each node can be quantified independently. This allows each variable to be populated with the best data available, including expert opinion, simulation results or observed data. It also allows the information to be easily updated as better data become available This paper reports on the process of developing a BBN to better understand the initial rapid growth phase (initiation) of a marine cyanobacterium, Lyngbya majuscula, in Moreton Bay, Queensland. Anecdotal evidence suggests that Lyngbya blooms in this region have increased in severity and extent over the past decade. Lyngbya has been associated with acute dermatitis and a range of other health problems in humans. Blooms have been linked to ecosystem degradation and have also damaged commercial and recreational fisheries. However, the causes of blooms are as yet poorly understood. The initial phase of model development consisted of a series of workshops and individual meetings with participants, in which important hydrologic, nutrient, light and temperature variables were identified and their causal relationships defined. This systematic approach allowed for the construction of a conceptual model describing factors and their interactions that contribute to the initiation of Lyngbya. This early model was subsequently refined, balancing model complexity against research costs. The CPT were then populated by the group based on their scientific knowledge. In some cases this was via a process, simulation, mathematical or statistical model; in others it was via expert opinion and round-table agreement. The software package Netica® was used to construct the model. Note that further refinements to the model are expected. There have been a number of important outcomes from the process of building the preliminary BBN. The interchange among scientists and other stakeholders during workshops led to: a greater understanding of important variables and their interaction by building on the experience of the group, in turn leading to a greater maturity in reasoning about the problem during successive iterations of the model; a greater awareness of the role of individual scientists' research in the larger picture; and a recognition of the need to coordinate data and models arising from the different research projects. Significant outcomes are that the BBN has provided a framework to prioritise and integrate the ongoing research contributions and to quantify the data needs. At this stage, a range of longer term goals have been identified. The model will be further refined as new research or better information becomes available, and serve as a focus for the coordination of data and models arising from the various research projects on Lyngbya . This BBN will also be used to contribute both to the design of monitoring programmes for Lyngbya in Moreton Bay, and the analysis of data from these and other programmes. The model should be validated by applying the Bayesian network to other locations in order to identify local and global characteristics and assess the robustness of both the conceptual model and the network inputs. Importantly, it is planned to develop a similar conceptual model and BBN for the management aspects of Lyngbya, and integrate this into the scientific model with the aim of identifying possible management strategies, predicting their potential impact and developing scenarios relevant to future planning.