基于水文多分形特征的流域站网空间配置研究

Basin station network can provide basic data of hydrology for the watershed planning, analysis and management of water resources. However, since it is difficult to take account of the Long-range dependency, such as the space coupling of the characteristics among points and information transmission, the spatial structure of the station network shows a low confidence in the efficient data collection and collaborative evaluation of water resources and it is also one of the unavoidable bottlenecks in digital basin construction. Accordingly, in order to adapt to the needs of overall function analysis and dynamic reasonable adjustments of the station network, research is taken on analysis of multi-fractal characteristic (Long memory and the Change point) in the hydrological characteristics of gauging station, more than on the evaluation of the spatial topology of the hydrological station network applying the spatial information processing technology. By taking the hydrometric station of Qingjiang River as sample data, the project structures a new analysis model based on multi-fractal theory which is suitable to analyze Long-range dependency among stations and the characteristic of change point, and optimize particle swarm so as to solve the model.It composites multi-agent technology to take the initiative to examine the space state of the station site and self-organization law of the combination; Besides, the particle swarm optimization is embedded in the GIS platform to build the weighted Voronoi diagram, to achieve the digestion of space conflict, such as the difference overlay on the inter-site information, and to analyze the spatial stability of the station network. The prospective achievements will have a balanced effect on evaluation and optimization of network layout and will provide possible ways for communication net organization in basin.

为流域规划、水资源分析和管理提供水文基础数据的流域站网，由于布设时难以顾及点间特性量的空间耦合、信息传递等特性，其站网全域空间结构对数据高效采集和水资源协同评价呈现出低置信度，更是数字流域构建不可回避的瓶颈之一。拟开展以水文多分形特征（长记忆性和变点）分析为基础的流域站网空间配置研究，以适应站网整体功能状态分析和动态合理调整的需要。项目以清江流域水文站点为实验对象、GIS为信息处理平台，在地形因子分析基础上，基于多重分形理论构建适于站间特性量长记忆和变点分析模型，并改进微粒群算法来解算模型；复合多智能体技术，以期主动考察站点的空间状态与点间组合的自组织特征；把改进的微粒群算法嵌入GIS平台，探索适于流域站网的Voronoi图分析，实现消解站点间信息的差异叠合等空间冲突，分析站网的空间结构稳定性。成果具有评价和优化水文站点空间配置的均衡作用，为流域通讯组网提供方法储备。

随着长江流域经济带的保护性开发，对其支流—湖北西部清江的水资源预测、水文监测的需求凸显。项目针对清江流域站网空间优化问题，开展研究并取得如下成果：（1）针对平行位移优化与重叠优化改进的ATRIC算法实现了清江流域集水面积阈值的确定，自动提取的清江河网与实际河网的拟合程度高。（2）基于盒维数法计算了清江流域水系分形维数值，得出汇流累积阈值、分维值和河网密度均有着良好的对数拟合关系，方法可实现快速分析地貌特征。（3）首次把回溯搜索优化算法(BSA)用于估计非线性马斯京根模型的参数取值，并用洪水演算过程对优化得到的模型参数进行校准，证明了BSA在水文行为及水资源混沌系统控制及参数辨识方面的有效性。（4）研究了站网评价和空间重构的信息熵多目标方法。一方面通过改进联合熵最大和冗余度最小的双熵多目标优化模型，拟定联合熵比、冗余度比和NSE三个目标函数，对站网空间位置进行重构，有效地对原站网进行了约简；另一方面建立基于信息熵的多目标组合优化模型，通过考察信息提供量与信息冗余量的平衡来确定站网组合，分别从信息量百分比、信息冗余度、信息传递能力对水文站点进行评价和优化，最终基于信息距离传输模型对站点空间控制范围进行评估与可视化表达。（5）运用主成分法构建了新型遥感生态指数(RSEI)，克服了以往生态环境评价指标难以空间化和主观性确定权重的缺点，对清江水电开发区域的生态环境进行评价；定量揭示了清江中下游地区水资源利用与生态环境的空间变化关系。（6）引入分数阶算子，结合人工鱼群算法(SAFSA),建立了改进灰色预测水量模型，用于用水量的预测和水资源量时空变异特性研究，改进模型在拟合精度及预测精度均起到积极作用。（7）利用IDRISI工具的CA-Markov模型，探究了清江流域土地利用动态变化趋势与流域水文特性的时空关系。项目成果对清江流域水资源调度、水电开发和生态保护有借鉴。

内容获取失败，请点击重试