Environmental processes are highly complex and their understanding involves the analysis of various quantitative and qualitative parameters (physical, chemical, geographical etc), which are more or less correlated. Appropriate environmental knowledge can deal with this complexity in a tractable way. Such knowledge is essential for solving particular environmental problems. Generating valuable environmental knowledge is a challenging research topic, especially for environmental data science, as efficient knowledge can lie behind data. Integrated environmental modelling uses a holistic view and can provide a possible better solution to environmental problems understanding. The paper presents a knowledge modelling framework for intelligent environmental decision support systems (IEDSS) by following such a holistic perspective. Thus, the proposed framework integrates an ontological approach and two data analysis approaches (data mining and Bayesian networks), which are applied for the generation of a knowledge base that is used by an IEDSS for decision making. The application of the framework is illustrated on three case studies from different environmental domains: (1) water (river resource management, river water pollution analysis), (2) air (air pollution analysis, ozone prediction), and (3) soil (soil pollution analysis).

环境过程非常复杂，其理解涉及对各种定量和定性参数（物理，化学，地理等）的分析，这些参数或多或少是相关的。适当的环境知识可以轻松应对这种复杂性。这些知识对于解决特定的环境问题至关重要。生成宝贵的环境知识是一个具有挑战性的研究课题，尤其是对于环境数据科学而言，因为高效的知识可能会隐藏在数据背后。集成环境建模使用整体视图，可以为理解环境问题提供更好的解决方案。通过遵循这种整体观点，本文提出了一种智能环境决策支持系统（IEDSS）的知识建模框架。因此，所提出的框架集成了本体论方法和两种数据分析方法（数据挖掘和贝叶斯网络），这两种方法被用于生成知识库，该知识库由IEDSS进行决策。在来自不同环境领域的三个案例研究中说明了该框架的应用：（1）水（河流资源管理，河水污染分析），（2）空气（空气污染分析，臭氧预测）和（3）土壤（土壤污染分析）。