Developing best-practice Bayesian Belief Networks in ecological risk assessments for freshwater and estuarine ecosystems: A quantitative review

McDonald, Karlie; Ryder, Darren; Tighe, Matthew

doi:10.1016/j.jenvman.2015.02.031

Cited by 40 publications

(35 citation statements)

References 72 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Then, the spatio-temperal distribution of ecological risk assessment at a national scale can be made for scientific decision-makings. However, a specific threshold index for ecological risk assessment should be aimed to a specific region, and can be improved by other methods, e.g., Bayesian network approaches [36], artificial neural network [37], and ecosystem services [38,39].…”

Section: Conclusion and Discussionmentioning

confidence: 99%

Regional Ecological Risk Assessment in the Huai River Watershed during 2010–2015

Qin

Chang

et al. 2017

Sustainability

View full text Add to dashboard Cite

Ecosystem deterioration has been and is still a serious threat to human survival and regional economic development. Theoretical and methodological challenges exist in assessing ecological risk of watershed ecosystem that is imposed by natural changes or human activities. To fill this research gap, this research proposes an interdisciplinary and quantitative methodology based on some techniques such as the Procedure for Ecological Tiered Assessment of Risk (PETAR), the Entropy, and the Celluar Automata Markov (CA-Markov). We focused on six vulnerable environmental variables, namely land-use change, water quantity, water quality, gross domestic product (GDP), environmental pollutants, and soil erosion in the Huai River watershed in the Henan Province in order to build multi-dimensional quantitative method. Further, the Coupling Coordination Degree Model is constructed, and the "threshold index" is also addressed to reflect the limitation of ecological risk. Our results show that the spatio-temperal distribution of the eco-environmental quality has greatly varied across this study area during different time spans. Natural eco-environmental quality has moderately degraded in 70% of this study area (mainly agricultural region), at a prefectural level from 2000 to 2010, and has slightly improved over the agricultural region (<170 m above sea level) during 2010-2015. However, when considering negative stressors from human social system on the natural ecosystem, the extent and distribution of the ecological risk varied across the whole area during 2000-2015. The results show that there was almost 90.40% of this region under the ecological risk, with varying extents over the study time, e.g., Kaifeng, Shangqiu, Xuchang, and Xinyang, with a moderate deterioration in the eco-environmental quality, and Zhengzhou with a slight deterioration in the eco-environmental quality. This paper provides a valuable perspective for governments at all levels to manage watershed environment resources.

show abstract

Section: Conclusion and Discussionmentioning

confidence: 99%

Regional Ecological Risk Assessment in the Huai River Watershed during 2010–2015

Qin

Chang

et al. 2017

Sustainability

View full text Add to dashboard Cite

show abstract

“…The outcomes of each node are known as states (McDonald, Ryder, & Tighe, 2015). Probability distributions are usually defined for each node in terms of the states (i.e., conditional probability tables [CPTs]; McCann et al, 2006).…”

Section: Bayesian Belief Networkmentioning

confidence: 99%

“…It is important to note that causal links should be carefully defined in order to prevent high levels of uncertainty in the model's outputs (McDonald et al, 2015). It is important to note that causal links should be carefully defined in order to prevent high levels of uncertainty in the model's outputs (McDonald et al, 2015).…”

Section: Bayesian Belief Networkmentioning

confidence: 99%

“…CPTs, which are marginal probability distributions, are often parameterized and calculated using approaches based on Monte Carlo simulation (Phan et al, 2016), Gibbs sampling or dynamic discretization (Nojavan, Qian, & Stow, 2017;Pérez-Miñana, 2016), maximum likelihood or the Laplace correction (Aguilera et al, 2011), or the expectation maximization algorithm for small and incomplete datasets and the gradient learning algorithm for large incomplete datasets and continuous data (McDonald et al, 2015). CPTs, which are marginal probability distributions, are often parameterized and calculated using approaches based on Monte Carlo simulation (Phan et al, 2016), Gibbs sampling or dynamic discretization (Nojavan, Qian, & Stow, 2017;Pérez-Miñana, 2016), maximum likelihood or the Laplace correction (Aguilera et al, 2011), or the expectation maximization algorithm for small and incomplete datasets and the gradient learning algorithm for large incomplete datasets and continuous data (McDonald et al, 2015).…”

Section: Bayesian Belief Networkmentioning

confidence: 99%

“…There are recent reviews covering BBN applications in areas such as environmental modelling (Aguilera et al, 2011), ecosystem services modelling (Landuyt et al, 2013;Pérez-Miñana, 2016), ecological risk assessment (McDonald et al, 2015), and water resources management (Phan et al, 2016), which also include some early studies related to stream health modelling Marcot, Holthausen, Raphael, Rowland, & Wisdom, 2001). The aforementioned reviews addressed different aspects of BBN model development/application such as data preprocessing, complexity, training, optimization, validation methods, variations and extensions (e.g., dynamic Bayesian networks for time series modelling and hidden Markov models for higher order relationships; Tucker & Duplisea, 2012), integration with other modelling techniques, and software comparison.…”

Section: Bayesian Belief Networkmentioning

confidence: 99%

See 2 more Smart Citations

A review of macroinvertebrate‐ and fish‐based stream health modelling techniques

Hernandez‐Suarez

Nejadhashemi

2018

Ecohydrology

View full text Add to dashboard Cite

During the last three decades, explaining cause-effect relationships between natural and anthropogenic disturbances with measures of stream health have motivated the growing application of statistical, machine learning, and soft computing methods.The aim of this review is to provide insight into the most widely used methods for predicting biological variables based on macroinvertebrate and fish species in riverine ecosystems. Therefore, we describe several methods including multiple linear regression, generalized linear models, generalized additive models, boosted regression trees, random forests, artificial neural networks, fuzzy logic-based, and Bayesian belief networks along with recent applications of these. Moreover, issues regarding variable selection, model interpretability, ensemble modelling, and model evaluation and overfitting are discussed. Recent advances have suggested the need for integrated modelling systems to enhance the predictive ability and improve interpretability.However, trade-offs between model complexity and accuracy demand research efforts in uncertainty quantification/propagation in model ensembles. Additionally, models should be perceived as complementary tools that require further validation with field measurements. Therefore, a consensus regarding monitoring and modelling practices for stream health applications is recommended.

show abstract

Discussion of ‘Environmental decision‐making using bayesian networks: Creating an environmental report card’

Bodini

2017

Appl Stoch Models Bus & Ind

View full text Add to dashboard Cite

Developing best-practice Bayesian Belief Networks in ecological risk assessments for freshwater and estuarine ecosystems: A quantitative review

Cited by 40 publications

References 72 publications

Regional Ecological Risk Assessment in the Huai River Watershed during 2010–2015

Regional Ecological Risk Assessment in the Huai River Watershed during 2010–2015

A review of macroinvertebrate‐ and fish‐based stream health modelling techniques

Discussion of ‘Environmental decision‐making using bayesian networks: Creating an environmental report card’

Contact Info

Product

Resources

About