BP神经网络洪水预报模型在洪水预报系统中的应用

采用相关分析法,在区域降水、观测断面流量(或水位)因子中识别出影响预报断面径流过程的主要变量,在多个观测断面的数据均为流量情况下,采用基于时延组合的合成流量为影响预报断面径流过程的变量,采用自相关分析法,识别出影响预报断面径流过程的前期流量(或水位),以这些变量为BP神经网络模型的输入,以预报断面的流量(或水位)为模型的输出,在BP神经网络隐层节点数自动优选的基础上,构建了基于BP神经网络的洪水预报模型。将模型载入中国洪水预报系统中,应用结果表明:模型在历史洪水训练样本具有一定代表性的情况下,可获得较高的预报精度。     

水位演算模型及其在水位预报中的应用

将单一河道洪水演算系统视为单输入一单输出系统,基于时间序列分析中的ＡＲＭＡＶ（２,１）模型,应用时间序列分析最优控制理论建立了水位演算模型和水位预报方法。证明若采用水位作变量,则可采用无约束的最优化方法确定参数;若采用水深作变量,则可采用以水量平衡为约束的寻优方法确定参数。本文所建立的水位预报方法用于黄河下游水位预报,得到了满意的预报精度。     

考虑回水顶托影响的水位预报研究

将有回水顶托影响的河段视为多输入-单输出系统,把水位过程视为时间序列,用时间序列分析最优控制理论,建立了可考虑回水顶托影响的水位过程预报模型。提出了在棱柱形河道情况下,以时段水位涨差作变量时,模型中各参数应满足的物理性约束条件;较好地解决了长江荆江河段中受回水顶托影响的石首站的水位过程预报问题,为水位过程预报提供了一条不同于水文学途径和水力学途径的建模途径。     

河道洪水实时概率预报模型与应用

通过数据同化方法合理地将实时水文观测数据融入到洪水预报模型中,可提高洪水预报模型的实时性和精确度。选取沿程断面流量、水位和糙率系数作为代表水流状态的基本粒子,以监测断面实测水位数据作为观测信息,建立了基于粒子滤波数据同化算法的河道洪水实时概率预报模型。模型应用于黄河中下游河道洪水预报计算的结果表明,采用粒子滤波方法同化观测水位后,不仅可以直接校正水位,同时也可以有效地校正流量和糙率,为未来时刻模型预报计算提供更准确的水流初始条件和糙率取值区间,进而有效地提高模型预报结果的精度,给出合理的概率预报区间。不同预报期的预报结果表明,随着预报期的增长,同化效果减弱,模型预报结果的精度会有所降低,水位概率预报结果受粒子间糙率不同的影响不确定性增加,而流量概率预报结果受给定模型边界条件的影响不确定性降低。所提出模型可以有效同化真实水位观测数据,适合应用于实际的洪水预报工作中。     

双向线性回归法在椒江临海站水位预报中的应用

将感潮河段预报断面的洪水过程视作上游洪水波和下游潮水波双向传播后的叠加,两个传播过程分别应用水位演算方法进行模拟计算,再结合多元线性回归法,以此建立同时受多个主要上游洪水及下游潮汐影响的水位预报模型.该模型方法简单.物理概念清楚.本模型在椒江感潮河段临海站的水位预报过程中取得了良好的效果.     

长江中游洞庭湖防洪系统水流模拟——Ⅱ．模型实现和率定检验

在长江中游洞庭湖防洪系统水流模拟模型的建模思路和基本算法的基础上,侧重讨论建模过程中遇到的内、外边界条件的特殊处理、洪水实时预报校正模型的应用、资料处理和模型数据的拟定等重要环节,强调了有关处理及其与模型基本算法的磨合在复杂水流模型中的重要作用.采用80年代两场实测洪水对所建模型进行了率定和检验.结果表明,所建模型清晰地重现了长江中游--洞庭湖区的洪水动态演进过程,得到了总体趋势合理的水位分布和流速或流量分布;模拟的控制站水位、流量过程与实况比较,峰谷完全对应,过程吻合良好,精度满足防洪预报和分析的要求,能够较好地模拟洪水波在这样复杂水系中的水流特征.     

GIS-PModflow联合系统在松嫩平原西部潜水环境预警中的应用

针对松嫩平原西部地下水潜水位变化对环境产生影响的现状,运用GIS-PModflow联合系统确定了地下水环境预警警戒线,采用以不至于发生土壤盐碱化的潜水临界水位作为上警戒线值,以潜水开采极限深度作为下警戒线值。在对系统模型概化的基础上,进行了潜水位的数值模拟和预报。利用GIS空间分析方法对图像数据进行栅格叠加和属性判断,通过对预报水位与警戒水位对比,进行水环境预警。结果表明,研究区2015年的警情较1999年增加,且以灌溉后潜水位上升引起的土壤次生盐渍化的警情为主。     

线性扩散波方程解析解及其在水位预报中的应用

提出一个可考虑回水顶托及旁侧入流影响的线性扩散波方程水位解析解模型,并对其响应函数特性进行了研究。该模型是一个集总参数的水位预报模型。模型只包括:谢才系数C和反映河道断面不规则影响的扩散系数DK,并可由试错或优化法求出。对于给定的上、下边界条件和旁侧入流,模型可快速给出各内插断面的水位过程。选用长江荆江河段1981年～1990年34次洪水资料进行参数率定,1991年～1993年7次洪水资料进行模型检验,结果令人满意。     

冲淤河道水位预报模型及其应用

根据河流动力学基本方程,采用水文学与水-沙关系相结合的方法,建立了适用于不受回水顶托影响,以悬移质为主且水-沙关系良好的冲淤河道的水位预报模型,应用于渭河咸阳-华县河段,精度提高。     

三峡水库动态汛限水位洪水调度风险指标及综合评价模型研究

汛限水位是综合利用水库运行和调度的重要参数之一,也是协调防洪和兴利矛盾的焦点所在。现行的汛限水位过多地考虑了小概率洪水事件,不能充分挖掘水库汛期的兴利效益,因此,采用动态汛限水位进行调度,对综合利用水库的运行具有重要的理论意义和实用价值。根据三峡水库围堰发电期的调度规程,建立预报预泄调度模型,采用宜昌站1882-2001年汛期实测日流量资料,实现了考虑预报信息的动态汛限水位洪水调度模拟;提出了多目标风险指标体系;计算了9种动态汛限水位方案下的风险指标值,通过综合评价模型对各方案进行比较和优选,得到了相对合理的动态汛限水位方案。     

用径向基函数神经网络模型预报感潮河段洪水位

径向基函数神经网络方法是一类比较优越的前向式多层神经网络,将其应用于感潮河段的洪水位预报。利用K 均值算法和最小二乘法来确定径向基函数神经网络的参数,并给出了具体计算方法。由于该方法比传统的BP算法有较快的收敛速度,使其具有较大的应用价值。基于感潮河段的具体特点,构建了具有若干个时段预见期的径向基函数神经网络模型。该模型应用于沂河的水位预报,结果表明,该模型运算快速、简便,预报精度较高。     

Tidal variability in the water quality of an urbanized estuary

Tide and water quality data were collected concurrently in the Hackensack River estuary (HRE), a tidal tributary of the New York-new Jersey Harbor estuary. Harmonic analyses of tidal elevation data indicate that HRE tides are predominantly semidiurnal, though modulated by diurnal and fortnightly components. Nearly uniform tidal ranges (averaging approximately 1.6 m) were observed at three stations within the HRE. Periodogram estimates reveal significant tidal variability for the water quality parameter NH₄−N under dry-weather conditions. Lag correlation analyses associate NH₄−N concentration variations with water level fluctuations. Longitudinal profile plots for NH₄−N reveal a consistent pattern of tidal translations, with peak concentrations oscillating about a major wastewater discharger. These analyses suggest that the distribution of NH₄−N concentrations in the HRE is controlled primarily by major point source loadings and horizontal advection. A simplified, one-dimensional model is used to describe this distribution. Effects of tidal variability in masking water quality status and waak trends are also analyzed. These analyses highlight the potential importance of short-term water quality variability in tidal estuaries where concentration gradients are large.     

碳酸盐岩旋回沉积的形成机制

为了更好地理解碳酸盐岩旋回沉积的形成机制,计算机模拟可谓是一种既直观又定量的好方法。我们重点选择了与沉积相有关的五个重要参数：（１）与水深相关的沉积速率;（２）时滞（Ｌａｇｔｉｍｅ）;（３）盆地线性沉降速率;（４）潮差;（５）海平面振荡幅度与周期等。由于碳酸盐岩沉积物对海平面变化非常敏感,因此研究中特别注意了海平面振荡幅度与周期的变化对碳酸盐岩旋回沉积的控制和影响。一个几米厚的潮坪旋回层序通常是由海平面的低幅振荡形成的,振荡幅度也就是几米,时滞亦较短。当海平面振荡幅度较大时,在海平面下降期间,滨线向海迁移的速率会超过潮坪的侧向加积速率,因此在内陆棚上缺乏潮坪沉积物。海平面振荡幅度的增加也能造成海平面下降快于潮坪的沉降,滨线向海迁移速率比潮坪进积的速率要快得多,从而在旋回层序内发育较厚的渗流带以及不整合界面。某些高幅振荡（１００ｍ或更大）还可造成碳酸盐岩台地的早期淹没以及沉积旋回内部的深水相与浅水相并置     

珠江三角洲网河区低水位时空变化规律

珠江三角洲近30年的大规模无序采挖沙改变了网河区水多沙少、河势稳定的演变趋势,促使河床由缓慢淤积转变为大幅下切,低水位出现明显下降。这种人为作用已涵盖并远超同期河流自然演变的程度,对取水供水、生态保护、航运保障等产生了较大影响。在总结河口三角洲地区水位变化研究的基础上,以珠江三角洲网河区为例,按低水位变化划分为常年潮流段、季节性潮流段和常年径流段。其中,常年潮流段基本不受采砂影响;季节性潮流段依据径流、潮流比例,越靠近枯水潮流界,低水位变化越大;常年径流段还存在溯源冲刷引起的低水位下降。结合水沙变异过程,试图探讨水动力对河床急剧演变的响应机制,以期为通航低水位设计提供依据。     

钱塘江河口杭州湾风暴潮溢流计算方法研究

建立钱塘江河口杭州湾风暴潮模型,探讨风暴潮出现溢流的计算方法。将可能出现溢流的沿海堤防以及海水侵入的陆地均依照高程概化为计算区域,采用糙率控制潮水的溢流流量,以模型的堤顶单宽流量和根据计算潮位采用宽顶堰公式换算流量的一致性来率定糙率值。在此基础上模拟了风暴潮漫溢堤防的过程,结果表明风暴潮出现溢流后,钱塘江河口杭州湾之间两岸大片的陆地存在淹没风险,沿程潮位由于溢流出现不同程度的降低响应。     

Impact of River-Tide Dynamics on the Temporal-Spatial Distribution of Residual Water Level in the Pearl River Channel Networks

The behavior of the residual water level in estuarine environment is complex due to the highly nonlinear interaction between river flow and tide and the contributions made by these two external forcing to the dynamics of the residual water level are not yet fully understood. In this study, we investigate the effect of river-tide dynamics on the temporal-spatial changes of flow in terms of residual water level in the Pearl River channel networks, which is one of the complex channel networks in the world. Making use of a nonstationary tidal harmonic analysis, the continuous time series observations of water level covering a spring-neap cycle in 1999 (representing flood season) and 2001 (representing dry season) collected from around 60 stations in the Pearl River channel networks have been used to extract the temporal-spatial changes in stage and tidal properties (including amplitudes and phases) as a function of variable freshwater discharge and ocean tide. It was shown that the averaged residual water level during the flood season (ranging 0–5 m) is one order magnitude than that during the dry season (ranging 0–0.35 m). The distribution of the residual water level clearly indicates that the Pearl River channel networks feature two sub-systems, i.e., the central part of the channel networks being river-dominated with high value of residual water level and the eastern and western sides being tide-dominated with low value of residual water level. To understand the relative importance of river flow and tide on the temporal-spatial distribution of the residual water level, an idealized model is subsequently applied to the Modaomen estuary, which debouches the largest portion of river discharge into the South China Sea. Analytical results showed that the residual water level is mainly determined by the variation of the freshwater discharge for the flood season, while it is primarily controlled by the tidal forcing for the dry season and features a typical spring-neap cycle.     

钱塘江河口段盐水入侵的时空变化及预测模型

杭州城市供水85%取自钱塘江河口段,取水水质在枯水大潮期都不同程度地受到盐水入侵的威胁,分析钱塘江河口盐水入侵时空变化及研制二维数值预测模型对保障城市供水安全十分必要。根据钱塘江河口段实测水文氯度资料,分析了强潮作用下盐水入侵的时空变化特征;据此构建考虑斜压作用的二维水流、盐度输移的耦合数学模型,计算格式采用守恒性较好的有限体积法;在模型验证的基础上,数值分析了径流和潮汐对钱塘江河口段盐水入侵的影响,结果表明河口段的盐水入侵明显地受径流和潮汐的影响,据此可通过增大上游新安江水库的下泄流量抑制盐水入侵上溯以减小取水口氯度及超标时间,确保用水安全。     

Rainfall-Runoff Modeling using Clustering and Regression Analysis for the River Brahmaputra Basin

In this research, k-means, agglomerative hierarchical clustering and regression analysis have been applied in hydrological real time series in the form of patterns and models, which gives the fruitful results of data analysis, pattern discovery and forecasting of hydrological runoff of the catchment. The present study compares with the actual field data, predicted value and validation of statistical yields obtained from cluster analysis, regression analysis with ARIMA model. The seasonal autoregressive integrated moving average (SARIMA) and autoregressive integrated moving average (ARIMA) models is investigated for monthly runoff forecasting. The different parameters have been analyzed for the validation of results with casual effects. The comparison of model results obtained by K-means & AHC have very close similarities. Result of models is compared with casual effects in the same scenario and it is found that the developed model is more suitable for the runoff forecasting. The average value of R² determined is 0.92 for eight ARIMA models. This shows more accuracy of developed ARIMA model under these processes. The developed rainfall runoff models are highly useful for water resources planning and development.     

Modeling the impact of land reclamation on storm surges in Bohai Sea,China

A nested model for the simulation of tides and storm surges in the Bohai Sea, China, has been developed based on the three-dimensional finite-volume coastal ocean model. The larger domain covers the entire Yellow Sea and Bohai Sea with a horizontal resolution of ~10 km, and the smaller domain focuses on the Bohai Sea with a fine resolution up to ~300 m. For the four representative storm surges caused by extratropical storms and typhoons, the simulated surge heights are in good agreement with observations at coastal tide gauges. A series of sensitivity experiments are carried out to assess the influence of coastline change due to land reclamation in recent decades on water levels during storm surges. Simulation results suggest that changes in coastline cause changes in the amplitude and phase of the tidal elevation, and fluctuations of surge height after the peak stage of the storm surges. Hence, for the assessment of the influence of coastline changes on the total water level during storm surges, the amplitudes and phases of both the tidal and surge heights need to be taken into account. For the three major ports in the Bohai Bay, model results suggest that land reclamation has created a coastline structure that favors increasing the maximum water level by 0.1–0.2 m. Considering that during the storm surges the total water level is close to or even exceeds the warning level for these ports, further increasing the maximum water level by 0.1–0.2 m has the potential to cause severe damages and losses in these ports.