基于时变Copula的干旱与水资源短缺关系综合模型体系及应用

Given that there exist intrinsical distinctions between drought and water scarcity, we are the first in China to explore the relationship between drought and water scarcity from theory and methodology, based on what we have done before. Focusing on the relationship between drought and water scarcity, the interaction mechanism with other relating fields, such as water resources management, drought evaluation and drought management, could be analyzed and indicator system for comparing them could be set-up. Based on risk theory, the process of risk forming could be revealed. On the basis of regionalization analysis, Mann-Kendall test, improved Hilbert-Huang transform, dynamic time warping (DTW), dynamic comprehensive evaluation model and GIS technique could be used to analyze the catastrophe point, period, trend, similarity, change speed state, change speed trend and spatial characteristics of drought and water scarcity indicators. Then we could like to set the time period into natural period and interrupted period. The Water Rights Analysis Packages (WRAP) model could be used to establish a framework in order to distinguish drought and water scarcity. Base on the empirical distribution we established for drought and water scarcity indicators, time-varying copula could be used to calculate all kinds of return periods in order to reveal the dependency between them. Applied in Beijing, Tianjin and Hebei Province of Haihe River basin, we could like to draw risk maps of the relationship between them and make a calculation software based on time-varying copula. This research may promote the development and improvement of drought and water scarcity theory and methodology. Its results could provide a reference for other areas with a shortage of water resources in China.

鉴于干旱与水资源短缺（二者）存在本质的差别，在申请者前期工作的基础上，在国内率先从理论和方法上系统探讨二者的关系。以二者关系为核心，研究其与相关学科的相互作用机制、构建二者对比分析表征指标体系，以风险理论为基础、揭示其相互作用的机理与风险形成过程；在区划分析的基础上，采用变点分析、改进的HHT、DTW、动态综合评价模型、GIS技术等分析二者表征指标的突变点、变化周期与趋势、时变制约关系、变化速度状态和趋势、空间分布特征等；在天然时期与干扰时期划分的基础上，应用WRAP模型和游程理论构建二者关系的区分模型；通过构建二者表征指标的经验分布，进而构造时变Copula函数并作出最佳选择，计算各类重现期，研究二者相依关系；在海河流域京、津、冀地区开展应用研究，绘制二者关系风险图并做叠图分析，编制基于时变Copula的二者关系模型计算软件；该研究成果可为其他地区开展类似研究提供示范。

干旱和水资源短缺关系紧密，可二者是不同的现象，存在本质的区别。项目组对二者关系和风险问题做了深入的探讨。干旱风险侧重于农业损失，水资源短缺风险不仅对农业，对其他产业都可能带来损失，两者同时发生后存在着相互促进的作用，提升了风险发生后的等级程度。.构建了基于时变COPULA 的干旱与水资源短缺相依关系模型，分区揭示京津冀地区在干旱事件发生条件下水资源短缺事件发生的概率。研究表明，Clayton Copula函数能够较好的拟合逐月降水与逐月SPEI的二维联合分布；Gumbel Copula函数能够较好的拟合研究区逐年WEI+与逐年SPEI的二维联合分布。在相同的干旱程度下，w越大，WEI+≥w的极端水资源短缺发生的概率越小；干旱程度越严重，WEI+≥w的概率越大，WEI+≤w的概率越小。.开展了基于Bayesian-MCMC方法的时空嵌套Copula模型构建与应用的研究，创新性地提出干旱强度中心和干旱强度聚集区域的定义，这可直观的表现干旱强度在地理位置上的聚集和转移情况。在不影响信度的前提下，Bayesian MCMC估计得到的预测区间精度更高。.针对一个地区由二者引起的水分亏缺经常同时发生而且难以区分的情况，项目组提出基于SWAT水文模型定量区分干旱和水资源短缺对水文系统影响的模型体系，研究表明：人为因素是引起漳卫南运河流域水文系统发生变化的主要原因，并且人为因素影响造成的径流损失量是自然因素造成径流损失量的4倍。.采用GIS绘制干旱频率空间分布图，利用前期的研究成果绘制了京津冀地区水资源短缺风险图，最后采用叠图技术将干旱风险图和水资源短缺风险图做叠加和分类。.研究成果将干旱与水资源短缺关系研究推向了一个新的阶段，丰富了水资源风险管理的理论与方法，也可为京津冀地区实施可持续发展战略提供支撑。