文章基本信息

标题：A Bayesian network approach to study hydromorphological modifications over space and time in the framework of a sustainable river restoration project: the “Lac des Gaves” case study
本地全文：下载
作者：Rabab Yassine ; François Pérès ; Hélène Roux 等
期刊名称：IFAC PapersOnLine
印刷版ISSN：2405-8963
出版年度：2018
卷号：51
期号：5
页码：7-12
DOI：10.1016/j.ifacol.2018.06.191
语种：English
出版社：Elsevier
摘要：AbstractThe “Lac des Gaves”, an artificial lake located in the main stream of the “Gave de Pau” river in the high Pyrenees, has gone through very intensive sediment extractions over the past century. This resulted in the lake acting like a sediment trap causing a brutal longitudinal profile discontinuity that leads to increasing safety and environmental risks. Considering the multi-criteria character of this study, a deep analysis of the lake’s historical evolution until its current situation is needed to be able to propose sustainable restoration solutions. In order to understand the current situation and predict the future behavior of the study area after its restoration, we decided to analyze, in a complementary way, its historical and experimental hydromorphological characteristics to help design well suited solutions.To be able to include uncertainties in the qualification of the morphological trajectory from a period to another, identify the causes of a given modification, analyze their potential impact on the studied system and fill the data gaps by expert knowledge, a probabilistic approach supported by Bayesian Belief Networks (BBNs) has been chosen. BBNs are increasingly being used as tools for decision-making in river management because they can adjust to complex multi-criteria systems with multiple interactions like the ones we consider in this project. They can be divided in two components: a causal graph that illustrates the qualitative relationships between the variables and the quantitative description of these relationships thanks to Conditional Probability Tables (CPTs). The approach considered in this study will rely on the capacity to feed the models with information collected on the ground combined with physical knowledge on the phenomena. This part of the project is hence consistent with the approach presented above. The final goal would be to extrapolate this method to other watercourses going through the same kinds of pressures.
关键词：KeywordsBayesian Belief Networksmulti-criteriariver restorationhydromorphologydiachronic analysis