洪涝灾害是我国主要自然灾害之一。淮河流域因其独特的地理位置、复杂的气候条件使得该流域的洪涝灾害发生频繁且带来的影响巨大。为了风险防范减少损失，科学的风险评估是重要的研究内容。本研究基于洪涝灾害多由多个致灾因子共同作用产生的，引入金融领域中应用成熟的Copula函数，对淮河流域洪涝灾害发生的两个关键致灾因子建立双变量联合概率分布模型，进行双变量的联合重现期研究，并结合灾害产生的损失评估洪涝灾害风险。通过基于联合双变量致灾因子的洪涝灾害风险分析，有利于更加合理准确地表达洪涝灾害的复杂性，使风险评估的结果更科学，从而采取更加及时、合理的防灾减灾措施以降低该流域的洪涝灾害风险、减少经济损失以及灾害带来的影响。所得主要结论如下：

1、得出了淮河流域洪涝灾害致灾因子的最优边缘分布及Copula函数模型。

不管是淮河流域，还是两个典型地级市，Gumbel Copula函数对淮河地区的拟合效果是更好的。Gumbel Copula函数能够更好的描述淮河流域不同尺度、不同地区基于降水量、径流量这二者间的洪涝灾害。

对于淮河流域整体的洪涝灾害来说，Weibull分布对降水量的拟合效果最好；GEV分布对径流量的拟合是最优的。对于典型地级市，驻马店市的降水量和径流量的最优边缘分布分别是Weibull分布和GEV分布，信阳市的最优边缘分布均是GEV分布。

2、建立了双变量联合分布模型并推算联合重现期。

将降水量和径流量进行联合分布分析是非常必要的，且进行双变量联合分析的结果明显比考虑单变量的分析结果更符合实际，基于联合分布计算所得的联合重现期的结果更符合洪涝灾害发生的机理及实际情况，不同类型的重现期满足：双变量联合重现期≤任意特征变量的单变量重现期≤双变量同现重现期。基于Copula函数计算的联合重现期可以更客观、更全面地反映洪涝灾害的实际情况。 当计算了联合重现期的分布后，给定任意重现期的值，就能快速计算对应变量的取值范围；当给定变量值，亦能迅速得出联合重现期。

3、评估了淮河流域不同尺度基于联合重现期及损失的洪涝灾害风险。

淮河流域洪涝灾害发生非常频繁。流域的平均联合重现期不足0.02年，超过74%的灾害事件的联合重现期小于0.02年，且造成的损失在2亿元以下；驻马店市及信阳市的平均联合重现期均小于1年，且信阳市较驻马店市更小；驻马店市灾害损失在1亿元以下且联合重现期小于1年占比近50%，信阳市更是超过66%。不管淮河流域还是任意一个市的洪涝灾害基本符合小灾小损失、大灾大损失的特点，即小灾平均单次洪涝灾害损失值很小，对应发生的频次极高，大灾发生次数少，单次损失值较大，由此小灾产生的总损失也是不容忽视的，大灾造成的影响和损失也是可观的。

淮河流域的洪涝灾害并非只发生在变量值均超过一定阈值时，在降水量或是径流量任意一个量值或是两个变量的值较小的情况下，亦会产生洪涝灾害损失，且在致灾变量值较小的情况下，洪涝灾害发生的可能性也非常高。因此灾害的防治不仅要注重对大灾、巨灾的防治，还要注重对小灾、轻灾的防治。这也为城市或是流域防治洪涝灾害、提高防抗洪涝灾害的韧性提出更大的挑战，海绵城市、流域的建设就是极为有效的提高城市、流域防抗洪涝灾害能力里的一种方式。

Flood disaster is one of the main natural disasters in China. Due to its unique geographical location and complex climatic conditions, the flood disasters in basin of huaihe river occur frequently and have great impact. In order to prevent risks and reduce losses, scientific risk assessment is an important research content. This study is based on the occurrence characteristics of flood and the impact mechanism of multiple hazard factors. Then introduce the Copula function which applied maturely in the field of financial. A bivariate joint probability distribution was established for two main hazard factors in the basin of huaihe river, and the return period of two hazard factors was carried out, to analyze the flood disaster risk combine with loss statistics. It is helpful to express the complexity of flood disaster more reasonably and accurately. To take measures of disaster prevention and mitigation more timely and reasonable to reduce the risk of flood disasters, reduce economic losses and the impact of disasters. The main conclusions of this paper are as follows:

1. The optimal edge distribution and Copula function of flood hazard factors in basin of Huaihe river were obtained.

For the flood disaster in basin of Huaihe river, Weibull distribution has the best fitting effect on precipitation, and the GEV distribution is optimal for runoff. For typical cities, in Zhumadian, the optimal marginal distributions of precipitation and runoff are Weibull distribution and GEV distribution respectively, while the optimal marginal distributions in Xinyang are all GEV distributions.

Whether it is the basin of huaihe river or two typical cities, the Gumbel Copula function has a better fitting effect on these data of flood disaster. Gumbel Copula function can better describe the flood disasters in different scales in basin of huaihe river based on precipitation and runoff.

2. Establish the bivariate joint distribution model and calculate the joint return period.

It is very necessary to do joint distribution analysis with precipitation and runoff, and the results of considering multiple -formative factors to do multivariate joint analysis are superior to the results of traditional methods considering single feature variables analysis. The results of joint return period which calculated based on the joint distribution are more accord with the result of the flood disaster and the actual situation. The different types of return period meet: the bivariate joint return period ≤ the single variable return period of any characteristic variable ≤ the bivariate co-occurrence return period. The joint return period calculated based on the Copula function can reflect the real characteristics of flood disaster more objectively and comprehensively. The joint return period based on the joint distribution can quickly calculate the value ranges of the corresponding two variables in a given joint return period, and when the values of hazard factors were given, the joint return period can also be quickly obtained.

3. The risks of flood disasters of different scales in basin of huaihe river were evaluated according to the joint return period and the losses of flood disasters.

In basin of huaihe river, floods occur very frequently. In the basin, the average joint return period is less than 0.02 years. More than 74% disasters of joint return periods are less than 0.02 years, and the losses are less than 200 million. The average joint return period of Zhumadian and Xinyang all were less than 1 year, and Xinyang was smaller than Zhumadian. The loss in Zhumadian is less than 100 million and the joint return period is less than 1 year, accounting for nearly 50%, and Xinyang is more than 66%. Whether it is the basin of huaihe river or any city, almost matched the characteristics of minor disasters correspond to minor losses, while major disasters correspond to major losses. Although the average loss of minor floods is very small, the corresponding frequency is very high, and the number of large disasters is few, and the single loss is large. So the total loss caused by minor disasters cannot be ignored, and the impact and loss caused by large disasters are considerable.

Flood disasters do not only occur when two hazard factors exceed the certain thresholds. In the case that the value of precipitation or runoff is small, the flood disaster loss will also occur, and in the case of smaller values of the variables, the possibility of flood disaster is higher. Therefore, the prevention and control for flood disasters should not only pay attention to the large disasters, but also pay attention to the minor disasters. This also presents a greater challenge for cities and river basins to prevent and control flood disasters and improve their resilience against flood disasters. The construction of sponge city and river basin is one of very effective ways to improve their abilities to prevent and withstand floods.