基于贝叶斯网络的生态环境脆弱性评估模型与应用

发布时间：2018-01-12 15:28

本文关键词：基于贝叶斯网络的生态环境脆弱性评估模型与应用　出处：《中国科学院烟台海岸带研究所》2017年硕士论文　论文类型：学位论文

【摘要】：黄河三角洲湿地是我国北方典型的滨海湿地,蕴藏着极其丰富的生物资源、土地资源、油气资源以及旅游资源等,也是重要的鸟类迁徙中转站、越冬栖息地和繁殖地。近年来,由于人为干扰以及自然因素的共同作用,使得当地的生态环境脆弱性日益凸显。因此,开展黄河三角洲湿地生态环境脆弱性评估,识别造成研究区脆弱性的主要驱动因子,具有重要的科学意义。本文针对黄河三角洲生态系统复杂性和特异性,基于贝叶斯网络模型的方法进行生态环境脆弱性评估和预测。首先,在总结前人研究的基础上,筛选了代表研究区生态环境特征的指标,构建了包含8个因子的准则层、11个因子的指标层的脆弱性评估体系。其次,构建了基于贝叶斯网络的黄河三角洲生态环境脆弱性预测模型,利用交叉验证的方法进行模型筛选。在此基础之上,对黄河三角洲生态环境现状进行了评估,系统分析了研究区的脆弱性分布及原因,从多角度探讨指标与脆弱性之间的关系。模型敏感性分析表明:对研究区脆弱性影响较大的三个指标分别为:NDVI、土地利用和植被类型。最后,利用构建的贝叶斯网络模型,设置了与人类活动、气候、NDVI等相关情景,预测不同情景下的脆弱性状态,并提出了相应的管理建议,为黄河三角洲资源开发利用与生态环境保护提供了一定的科学依据。分析表明,研究区以轻微脆弱和轻度脆弱为主,约占总面积2/3,中度脆弱约占2/15,重度脆弱和极度脆弱约占总面积的1/5。通过模型性能分析,筛选一个错误率仅为17.35%的优良模型,为情景分析奠定了基础。设置的一系列情景分析表明,基于贝叶斯网络以不确定知识表达形式进行生态脆弱性预测是一种有效的方法。综上,贝叶斯网络模型可以综合各类生态环境信息,并可以有效融合,从而推断各因子之间的因果关系,为湿地生态环境脆弱性的预测提供了一种新的思路。
[Abstract]:The Yellow River delta wetland is a typical coastal wetland in northern China, has abundant biological resources, land resources, oil and gas resources and tourism resources, is also an important bird migration station, wintering habitat and breeding grounds. In recent years, due to the combined effect of human disturbance and natural factors, the local ecological environment the vulnerability has become increasingly prominent. Therefore, the vulnerability assessment on ecological environment of the Yellow River delta wetland, identify the main driving factors caused by the vulnerability of the study area, it has important scientific significance. In view of the ecological system of the Yellow River delta complexity and specificity method based on Bayesian network model for the ecological environment vulnerability assessment and prediction. First of all, on the basis of previous studies, selected representative of the ecological environment of the study area indexes, constructs the criterion layer includes 8 factors, 11 factors that The evaluation system of the vulnerability of the target layer. Secondly, the construction of the ecological environment of the Yellow River Delta Based on Bias network vulnerability prediction model, using the method of cross validation of the model selected. On this basis, the current situation of ecological environment of the Yellow River Delta were evaluated, systematic analysis of the study area to investigate the vulnerability distribution and reason. The relationship between the index and the vulnerability from various angles. The model sensitivity analysis shows that: the study area of vulnerability affects three of the high index were: NDVI, land use and vegetation types. Finally, using the Bias network model construction, set up and human activities, climate, NDVI and other related scenarios, vulnerability state prediction under different scenarios, and put forward the corresponding management recommendations for the development and utilization of resources in the Yellow River delta provides certain scientific basis and the protection of the ecological environment. The analysis shows that the study area is minor Fragile fragile and mild, accounting for a total area of 2/3, moderate vulnerability accounted for about 2/15, severe vulnerability and extreme fragility of the total area of about 1/5. through the model performance analysis, screening an error rate of only 17.35% of the fine model, laid the foundation for scenario analysis. A series of scene setting analysis shows that in to determine the knowledge expression form of ecological vulnerability prediction is an effective method based on Bayesian networks. Therefore, Bayesian network model can integrate various ecological environment information, and can effective integration so as to infer the relationship between causal factors, provides a new idea for wetland ecological environment vulnerability prediction.

【学位授予单位】：中国科学院烟台海岸带研究所
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：X826

【参考文献】