基于自组织映射神经网络的水库洪水过程分类

针对传统洪水分类方法中洪水特征提取时存在信息损失和主观性强的问题, 本文基于洪水全过程构建自组织映射神经网络(Self-Organizing Map, SOM), 综合考虑代表性和拓扑性等评价指标以优选网络规模, 实现洪水全过程的拓扑逻辑关系挖掘及分类。以三峡水库洪水过程为研究对象, 研究结果表明: ① 2×3维SOM覆盖率达到56.7%, 与3×3维SOM相比, 仅有约2%的覆盖率差距, 具有良好代表性; 2×3维SOM输出层仅有1处翻转, 拓扑结构比3×3维SOM更优, 更适合三峡水库洪水过程分类。② 2×3维SOM将洪水过程划分为6类, 其神经元拓扑结构可有效刻画各分类的差异与联系, 说明SOM可基于可视化拓扑逻辑关系实现高维洪水数据的可靠客观分类。③与传统方法的历史典型洪水分类结果相比, SOM能提供可靠且丰富的分类信息。     

基于PCP-C耦合模型的流域洪水分类研究

洪水过程受多要素综合影响,因此多指标的综合分类方法成为目前洪水分类研究的主要趋势.针对目前多指标洪水综合分类方法中存在的诸多不足,可采用主成分投影-聚类(PCP-C)耦合模型进行洪水分类.该方法首先对原始指标数据进行无量纲化(均值化)处理,再对处理后的数据矩阵进行正交变换,由此将原指标转换成彼此正交的综合指标,并利用各主成分设计一个理想决策向量,以各被评价对象相应的决策向量在理想决策方向上的投影值作为一维的综合分类指标.最后通过对各分类样本的一维投影值的聚类分析,得到分类结果.实例分析表明,建议方法简单,模型构建容易,计算简便,分类直观简洁,可行性强.     

地下水质量指标分类综合评价方法研究

近年来随着污染加剧,地下水中检出的污染指标种类不断增加,新修订的地下水质量标准中水质指标由GB/T 14848-1993的39项增加至93项,其中大部分为有机毒理学指标。原有的地下水质量综合评价方法因不能区分毒理学指标和感官性状与一般化学指标的差异,评价结果容易产生歧义,且评价结果不符合地学统计特征,已不再适用。为此在反复摸索计算的基础上提出了指标分类地下水质量综合评价方法。该方法利用单指标质量分类统计法评价感官性状与一般化学指标,利用单指标最大分类确定法评价毒理学指标,最终结果分别给出感官性状与一般化学指标级别和毒理学指标级别,同时给出毒理学指标的饮水途径健康风险以供参考。该方法在淮河流域某地区的地下水质量综合评价中应用结果显示,评价结果简单明了,符合地学统计特征,与内梅罗指数法相比,评价结果更为客观。地下水质量指标分类综合评价方法,可有效解决水质综合评价结果容易产生歧义和误导的问题,满足常规指标和毒理学指标的综合评价,能够更好地综合体现地下水水质信息。     

基于投影寻踪和粒子群优化算法的洪水分类研究

洪水分类实际上是洪水强度大小辨别的优化问题。洪水分类不仅影响着水库的实时调度，而且也影响着洪水灾害危险评估。对利用降水预报进行洪水资源利用的水库来说，洪水分类对水库实时调度规则的建立有着重要的作用。因此，洪水分类是一个重要的理论和实践问题。本文以长江三峡水库代表性水文站——宜昌站为研究对象．基于投影寻踪方法建立了洪水分类的优化模型，并利用粒子群优化算法对所建模型进行求解。结果表明了投影寻踪方法和粒子群优化算法在洪水分类研究中的有效性和合理性．     

洪水灾情SVM评估模型

为解决单项洪水灾情等级评价指标结果的不相容问题,提高综合评价的精度,本文利用支持向量机分类算法,将洪水灾情非线性、高维评价样本指标通过映射转换为高维空间的线性问题,选取径向基函数作为模型的核函数,建立洪水灾情的支持向量机分类模型,实例证明该方法简便、适用性强,为洪水灾情评价提供了新的途径.     

软弱夹层强度参数的主要影响因素分析

在大量试验研究以及总结前人研究成果的基础上,全面分析研究了物质组成、组构特征、粒度成分、含水量和地应力等主要因素对软弱夹层抗剪强度的影响,这些影响因素对夹层强度的影响是相互制约、相互影响的,并不是孤立的。因此,要正确、合理地给定软弱夹层强度参数,必须综合考虑各因素的影响,并抓住其主控因素。天然性状指标(w/wP)是一个能反映夹泥的含水量、密实度,黏土矿物成分、黏粒含量以及夹泥稠度状态的指标。以向家坝为例研究了性状指标(w/wp)与强度参数间的相关性,其相关性良好,并建立了性状指标和强度参数间的相关方程,这对容易获得天然性状指标的夹层的强度参数取值是非常方便的。     

变雨强单位线分析法在浙江地区的应用

一、前言我们在编制“暴雨径流查算图表”工作中,在分析我省暴雨洪水汇流特性基础上,应用纳希瞬时单位线概念性模式,并吸取我国应用单位线非线性改正的经验,同时结合原有变雨强单位线非线性分析,考虑各场暴雨洪水 NK 及对应 I 分析法,共用中小流域34个站387场暴雨洪水资料,进行了分析研究和综合定线。在分析计算中首先分析单位线的最优参数,然后结合地形条件进行单站和多站的参数综合。在上述基础上,我们还用292场洪水资料进行参数综合成果精度的验证计算。结果表     

沉积型软弱夹层成因分类及强度特征

研究区为一沉积岩覆盖区。由于区内软弱夹层分布广泛，构成了影响坝肩、坝基抗滑稳定性的王要工程地质问题之一。文章从软弱夹层的成因入手，认为区内特定的沉积环境和构造运动是夹层形成的基础，确定了夹层的成因类型；根据夹层分布、夹泥粒度组成，依据工程分类方法，将坝区夹层分为3种不同类型：粘泥型(Ⅰ1类)，泥含粉粒碎屑型(Ⅰ2类)及碎屑夹泥型(Ⅱ类)；分析了区内夹层夹泥的物理性质及表征区内夹层物理性质指标；通过分析软弱夹层强度特征与其物理性质的关系，确定了能够表征夹层强度特征的物理性状指标W／Wp；在软弱夹层工程分类基础上，研究各类夹层的强度特征，表明综合物理性状指标W／Wp与强度指标之间具有较好的相关性，确定了不同类型夹层的W／Wp和强度指标的相关关系式。在上述相关关系式基础上，结合坝区内控制性夹层的物理指标，给定了其强度建议值。文章这种“成因-物理性质-强度特征”的研究方法及以此确定的不同类型软弱夹层物理性状指标和强度指标间的相关关系，为沉积型软弱夹层的强度取值提供了一条捷径，对其它类型软弱夹层的强度取值也具一定的参考价值。     

南宁第三系浅表层风化泥岩物理力学及膨胀特性指标分析

通过对南宁市201个工程第三系泥岩资料进行的分析,对各物理、力学及膨胀性指标进行统计研究,为南宁盆地地下工程的参数选取提供参考。研究表明,南宁第三系浅表层风化泥岩物理指标变异性较小,尤其在95%置信区间取值范围偏差较小,在参数分析时可作为常量考虑; 而力学指标和膨胀特性指标数据变异性较大,需要考虑其差异性。南宁第三系泥岩具有抗压强度高,压缩性低,抗剪强度高,黏聚力大的特点。膨胀特性指标方面,南宁第三系泥岩的自由膨胀率与其他城市膨胀岩土相比偏低,为弱膨胀性岩土; 膨胀力因岩组成因和风化程度不同差异较大,建议在参数选取时针对具体工程考虑。     

南宁第三系浅表层风化泥岩物理力学及膨胀特性指标分析

通过对南宁市201个工程第三系泥岩资料进行的分析,对各物理、力学及膨胀性指标进行统计研究,为南宁盆地地下工程的参数选取提供参考。研究表明,南宁第三系浅表层风化泥岩物理指标变异性较小,尤其在95%置信区间取值范围偏差较小,在参数分析时可作为常量考虑;而力学指标和膨胀特性指标数据变异性较大,需要考虑其差异性。南宁第三系泥岩具有抗压强度高,压缩性低,抗剪强度高,黏聚力大的特点。膨胀特性指标方面,南宁第三系泥岩的自由膨胀率与其他城市膨胀岩土相比偏低,为弱膨胀性岩土;膨胀力因岩组成因和风化程度不同差异较大,建议在参数选取时针对具体工程考虑。     

Hydrological hazards in Russia: origin,classification, changes and risk assessment

Hydrological extreme events pose an imminent risk to society and economics. In this paper, various aspects of hydrological hazards in Russia are analysed at different scales of risk assessment. It is shown that the number of hydrological and meteorological hazards in Russia has been growing every year. The frequency of economic losses associated with extreme low flow in this century has increased by factor five compared to the last decade of the previous century. With regard to floods, an interesting spatial patter can be observed. On the one hand, the number of floods in the Asian part of the country has increased, whereas on the other hand, the number and intensity of floods in estuarine areas in the European part of Russia have significantly reduced since the middle of the twentieth century, especially in the 2000s. This decrease can be attributed to runoff flooding in the mouths of regulated rivers, with an effective system of flood and ice jam protection. The analysis shows that there is an 8–12-year periodicity in the number of flood occurrences and that flood surges have intensified over the last 110 years, especially on the European territory of Russia. An integrated index that accounts for flood hazards and socio-economic vulnerability was calculated for each region of Russia. A classification of flood risk was also developed, taking into account more than 20 hydrological and social–economic characteristics. Based on these characteristics, hazard and vulnerability maps for entire Russia were generated which can be used for water management and the development of future water resources plans.

     

Spatially distributed rainfall information and its potential for regional landslide early warning systems

Hydrological extreme events pose an imminent risk to society and economics. In this paper, various aspects of hydrological hazards in Russia are analysed at different scales of risk assessment. It is shown that the number of hydrological and meteorological hazards in Russia has been growing every year. The frequency of economic losses associated with extreme low flow in this century has increased by factor five compared to the last decade of the previous century. With regard to floods, an interesting spatial patter can be observed. On the one hand, the number of floods in the Asian part of the country has increased, whereas on the other hand, the number and intensity of floods in estuarine areas in the European part of Russia have significantly reduced since the middle of the twentieth century, especially in the 2000s. This decrease can be attributed to runoff flooding in the mouths of regulated rivers, with an effective system of flood and ice jam protection. The analysis shows that there is an 8–12-year periodicity in the number of flood occurrences and that flood surges have intensified over the last 110 years, especially on the European territory of Russia. An integrated index that accounts for flood hazards and socio-economic vulnerability was calculated for each region of Russia. A classification of flood risk was also developed, taking into account more than 20 hydrological and social–economic characteristics. Based on these characteristics, hazard and vulnerability maps for entire Russia were generated which can be used for water management and the development of future water resources plans.     

青海省玉树旱涝态势变化规律分析

选取青海省玉树1953~2013年月降水数据,基于标准化降水指数SPI,利用频率分析、小波周期分析等方法对玉树县近61年来旱涝强度、频率分布、周期性变化等旱涝态势演变特征进行了研究。研究结果表明:(1)SPI-3、SPI-6和SPI-12三种尺度标准化降水指数对旱涝指示程度存在差别,相比大时间尺度,小时间尺度的标准化降水指数值更为分散,波动幅度更大,对干旱和洪涝的识别更为敏感。近年来,随着玉树县水土流失和沙化,当地土壤持水力程下降趋势,对干旱和洪涝较为敏感,因此玉树县可采用三种尺度标准化降水指数。(2)近61年,玉树县重旱平均发生概率为2.1%,重涝平均发生概率为1.3%。(3)未来几年,预计玉树县仍呈现偏涝趋势。(4)玉树县SPI-12以18a为主周期进行变化。     

Global snowmelt flood disasters and their impact from 1900 to 2020北大核心CSCD

Under the background of climate warming, the occurrence time, frequency, intensity, and impact of snowmelt flood disasters have changed significantly. Thus, establishing a global snowmelt flood disaster database is particularly important for disaster risk management. With the help of a web crawler, and based on multiple data sources such as natural disaster databases, documents, books, government agency websites, and news media, this study collected relevant information of snowmelt floods and mixed floods and established standards for identifying snowmelt flood events and their disaster impacts based on data from the different sources. Following the screening, sorting, fusion, and integration of snowmelt flood events, a global snowmelt flood disaster dataset containing 579 pieces of data with strong pertinence and reliability was constructed. The temporal and spatial distribution characteristics of global snowmelt flood disasters from 1900 to 2020 were preliminarily analyzed. The results showed that the snowmelt floods were mainly distributed between 30° N and 60° N, with more mixed floods south of 50° N and more snowmelt floods north of 50° N. Spring was the period of highest incidence of snowmelt flood disasters, followed by winter, summer, and autumn, respectively. The snowmelt floods that occurred in spring, autumn, and winter were mainly at 40°~50° N, and the snowmelt floods that occurred in summer were mainly at 30°~40° N. Compared with the snowmelt floods, the mixed floods were more frequent and more destructive, and their frequency increased with climate warming. The results provide a scientific basis for risk prevention and loss assessment of global snowmelt flood disasters. © 2022 Science Press (China).     

西江流域梧州站干支流洪水组成及遭遇规律分析

为分析西江流域灾害性洪水的风险,采用水文学分析法从洪水组成、遭遇时间以及洪峰重现期等方面,重点分析了梧州站干支流洪水遭遇规律。研究表明:梧州15场洪水中红水河、柳江、郁江及桂江的日平均流量平均比重分别为32.31%、40.57%、13.19%及9.13%;其中西江干流洪水发生时间集中在6～8月,柳江洪水发生时间集中在6～7月,桂江洪水发生时间集中在5～7月,郁江洪水发生时间集中在7～9月;西江上游红水河与柳江洪水遭遇频繁,而郁江、桂江遭遇洪水量级较小,其中全流域洪水、红水河—柳江—桂江洪水、柳江—郁江—桂江洪水三种遭遇类型所构成的梧州大洪水重现期依次约为100年、50年、20年一遇,揭示了西江干支流洪水遭遇后致使梧州站洪水重现期显著增大。研究结果可为西江流域防洪减灾提供水文分析参考。     

动态实时洪水风险分析及管理系统

当前洪水风险分析按照典型设计标准洪水进行计算的模式难以满足实际防洪管理需要,为了提高洪水风险分析的实时性以及适应洪水演进的动态性,设计了动态实时洪水风险分析框架。在本框架中,先采用一维和二维动态耦合水动力学数值方法耦合溃堤模型,然后在樵桑联围防洪保护区建立洪水演进模拟模型,通过灵活处理模型计算边界条件以及动态设置溃堤功能,计算不同设计标准洪水发生时,堤防出现单一溃口或者组合溃口后保护区内洪水演进过程。按照上述框架开发了樵桑联围动态实时洪水风险图编制与管理应用系统,并利用历史洪水资料开展模型验证,验证结果表明,2008-06洪水马口站、三水站、大熬站、甘竹（一）站的实测最高水位和模型计算最高水位的绝对误差分别为-0.10、0.10、0.09、0.04 m,均满足洪水模拟精度要求。利用模型计算了西江发生200年一遇的洪水情况下,江根堤防出现溃口后的洪水流量及溃口内外洪水水位变化过程,模拟溃口宽度168 m,最大溃口洪水流量达到5 190 m³,分析了堤防溃决后3、6和24 h洪水漫延导致村落淹没情况,结果表明其满足合理性分析。     

Damaging flood severity assessment in Northern Portugal over more than 150 years (1865–2016)

Floods are a major natural hazard, with vast implications over a wide range of socio-economic activities. A harmonized post-flood classification is critical for a better understanding of this hazard, by providing homogeneous flood catalogues for future research on triggering mechanisms. We apply a flood severity index (FSI) to damaging floods in Northern Portugal over a 152-year period (1865–2016) and identify the most critical areas to flood occurrences. The index is a damage-based post-event assessment tool, which includes five categories ranging from minor flooding (1) to catastrophic flooding (5). FSI is applied to a historical damaging flood database with 2318 occurrences. In Northern Portugal, serious floods (3) are the most frequent typology, while catastrophic floods are typically river floods occurring in the Douro basin. Overall, damaging flood occurrences are favoured by the positive phase of the East Atlantic pattern and by the negative phase of the North Atlantic Oscillation. Furthermore, the north-western areas reveal higher concentrations of damaging flood occurrences, mainly due to higher population density, higher precipitation values and more flood plain areas. In particular, 48% of all occurrences are concentrated in the Porto Metropolitan Area, mainly the Porto city centre and nearby riverside areas of the Douro River. High-population density and heavily urbanized areas lead to greater exposure to flood risk, whereas the most peripheral municipalities, with large agricultural/forested areas, show much lower numbers of damaging floods. FSI is tool to communicate the magnitude of the flood risk and is, therefore, of foremost relevance to civil protection and risk management.     

Scaling property of regional floods in New South Wales Australia

Regional flood frequency analysis (RFFA) is often used in hydrology to estimate flood quantiles when there is a limitation of at-site recorded flood data. One of the commonly used RFFA methods is the index flood method, which is based on the assumptions that a region satisfies criterion of simple scaling and it can be treated homogeneous. Another RFFA method is quantile regression technique where prediction equations are developed for flood quantiles of interest as function of catchment characteristics. In this paper, the scaling property of regional floods in New South Wales (NSW) State in Australia is investigated. The results indicate that the annual maximum floods in NSW satisfy a simple scaling assumption. The application of a heterogeneity test, however, reveals that NSW flood data set does not satisfy the criteria for a homogeneous region. Finally, a set of prediction equations are developed for NSW using quantile regression technique; an independent test shows that these equations can provide reasonably accurate design flood estimates with a median relative error of about 27%.