Comparative Study of Three Updating Procedures for Real-Time Flood Forecasting

Accurate real-time flood forecasting is essential for flood control and warning system, reservoir operation and other relevant water resources management activities. The objective of this study is to investigate and compare the capability of three updating procedures, namely autoregressive (AR) model, recursive least-squares (RLS) model and hydrologic uncertainty processor (HUP) in the real-time flood forecasting. The Baiyunshan reservoir basin located in southern China was selected as a case study. These three procedures were employed to update outputs of the established Xinanjiang flood forecasting model. The Nash-Sutcliffe efficiency (NSE) and Relative Error (RE) are used as model evaluation criteria. It is found that all of these three updating procedures significantly improve the accuracy of Xinanjiang model when operating in real-time forecasting mode. Comparison results also indicated that the HUP performed better than the AR and RLS models, while RLS model was slightly superior to AR model. In addition, the HUP implemented in the probabilistic form can quantify the uncertainty of the actual discharge to be forecasted and provide a posterior distribution as well as interval estimation, which offer more useful information than two other deterministic updating procedures. Thus, the HUP updating procedure is more promising and recommended for real-time flood forecasting in practice.     

基于数据同化校正参数的河流磷迁移估计研究

通过磷迁移数学模型合理估计磷在河流中的时空分布,对防治水体富营养化,抑制水华暴发具有重要的科学和工程意义。数据同化方法可以将模型和观测两种研究手段有机地结合起来,将观测数据融入模型,优化模型状态变量,校正模型参数,进而提高数学模型的模拟预报精度,并依托物联网技术将传统数学模型发展为实时数学模型。本文将粒子滤波数据同化算法引入到水动力-泥沙-磷迁移模型中,以实测的断面磷含量作为观测数据,在观测时刻优化磷含量估计结果,同时校正模型参数磷相平衡分配系数Kd,构建了水动力-泥沙-磷迁移模型同化系统。将其应用于长江上游寸滩至坝前河段的计算结果表明,所构建的同化系统在真实的河流中计算效果良好,可以有效地优化更新状态变量各相磷含量浓度,并反演出模型参数Kd随水沙水环境条件变化的动态变化过程,同化之后模型模拟预报磷输移过程的精度显著提升,为水质实时模型打下基础。     

基于分位点回归模型的洪水概率预报方法

采用分位点回归模型分析洪水预报的不确定性, 提供洪水预报倾向值( 预报概率分布的中位数) 和 90% 置信度的预报区间成果, 实现了洪水概率预报。基于/ 精度2可靠性0 联合评价指标对分位点回归模型计算的预报倾向值 和预报区间成果进行了评估。在信江流域梅港站的应用结果表明: 基于分位点回归模型提供的倾向值定值预报结 果可进一步提升洪水预报的精度; 同时该模型提供的 90% 预报区间结果具有较高的覆盖率( 约 90% ) 且离散度较小 ( 小于 01 20) , 表明预报区间以较窄的宽度包含了绝大多数的实测值, 预报可靠性较强。     

实时洪水预报中基于岭估计的AR修正模型研究

为提高实时洪水预报精度,经常将水文模型与误差修正模型相结合,AR模型因其结构简单广泛应用于实时洪水预报误差修正。然而,实际应用显示,AR模型时常出现修正结果不稳定现象,表现为流量修正幅度过大,甚至出现“震荡”现象,严重影响修正效果。鉴于此,本文从矩阵特征值角度解释了AR模型出现不稳定现象的原因,并引入岭估计方法选择性利用流量信息更新自回归系数,使其更满足真实流量的涨落过程,增强该模型的稳健性。将新方法应用于蔺河口流域,结果显示岭估计方法显著提高了AR模型的稳健性,从而改善了模型修正效果,进一步提高了洪水预报精度。     

2D Finite Element Inundation Modelling in Anabranching Channels with Sparse Data: Examination of Uncertainties

Flood inundation modelling in developing countries is severely limited by the lack of high resolution terrain data and suitable imagery to map flood extents. This study assessed the predictive uncertainty of modelled flood extents generated from TELEMAC2D model using low-cost, sparse input data commonly available in developing countries. We studied a river reach characterised by anabranching channels and river islands in eastern India. In this complex fluvial setting, we analysed computational uncertainty as a function of error in both satellite-derived flood-extent maps using a Generalised Likelihood Uncertainty Estimation (GLUE)-based approach. The model performance was quite sensitive to the uncertainty in the inflow hydrograph, particularly close to the flood peak. Evaluation of the flood inundation probability map, conditioned upon deterministic and probabilistic observed flood extents, reveals that the effect of using probabilistic observed data is only evident for portions of the model domain where the model output is free from consistent bias (over or under prediction) likely created by the imperfect terrain data.     

分布式水文模型结合气象预报方法初步探讨

为了尽量消除因流域空间非均一性引起的水文模拟不确定性,采用基于GBHM分布式水文模型以及具有明确物理意义的模型参数,利用三峡区间2011年5~6月期间的气象预报信息,探讨该区域实时洪水预报方法,以及不同预见期的洪水预报精度。结果表明,分布式水文模型与气象预报数据结合,能够较好地模拟该区间的洪水过程。该方法在一定预见期内能够对实时洪水过程进行预报,预报精度很大程度上取决于降水预报的准确性。     

KALMAN FILTERING CORRECTION IN REAL-TIME FORECASTING WITH HYDRODYNAMIC MODEL

Accurate and reliable flood forecast is crucial for efficient real-time river management, including flood control, flood warning, reservoir operation and river regulation. In order to improve the estimate of the initial state of the forecasting system and to reduce the errors in the forecast period a data assimilation procedure was often need. The Kalman filter was proven to be an efficient method to adjust real-time flood series and improve the initial conditions before the forecast. A new model integrating the hydraulic model with the Kalman filter for real-time correction of flood forecast was developed and applied in the Three Gorges interzone of the Yangtze River. The method was calibrated and verified against the observed flood stage and discharge during Three Gorges Dam construction periods （2004）. The results demonstrate that the new model incorporates an accurate and fast updating technique can improve the reliability of the flood forecast.     

关于水库防洪调度若干问题的探讨

笔者在分析了水库实时防洪调度与设计阶段防洪调度的差别后，对在实时调度可以较大幅度地提高水库效益的两种主要措施──分期防洪调度，预报调度，在实施中应注意的问题作了探讨，分期防洪调度应注意摸清分期洪水的规律；正确划分分期时段，按规范的规定和方法计算分期设计洪水；依分期设计洪水拟定分期汛限水位。预报调度必须细致地编制预报调度方案，尽快建设自动测报系统。挖掘水库潜力关键是对来水尽早，尽准确的预知，由于不要     

基于LSTM实时校正的WRF/WRF-Hydro耦合径流预报

为改进WRF/WRF-Hydro陆气耦合系统的径流预报效果,减小耦合系统在峰现时间、洪峰流量预报上的误差,本文在使用变分数据同化技术充分降低预报降雨误差水平的基础上,采用长短期记忆人工神经网络LSTM对WRF/WRF-Hydro耦合系统的径流预报过程开展了实时校正研究,并与自回归滑动平均模型ARMA实时校正结果进行对比。研究结果表明,通过数据同化技术可有效提升WRF模式降雨预报精度,降低WRF-Hydro模式的输入误差,但径流预报准确性仍有待提升。对比LSTM和ARMA两种实时校正模型对耦合径流预报结果的实时校正：在前3 h预见期,两种模型在中国北方半湿润、半干旱地区山区小流域6场典型洪水预报中的表现基本接近,除场次4外,LSTM和ARMA两种模型在3 h预见期的衰减速率分别为2.04~23.08和9.18~36.47,随着预见期的延长,LSTM径流预报精度的衰减速度在整体上慢于ARMA模型,预报效果优于ARMA模型。     

河道洪水概率预报方法研究

河道洪水预报对防洪减灾和水资源的综合利用有着重要的现实意义。考虑到水文预报不确定性的水文风险,为了更科学地对河道洪水进行预测,提出将贝叶斯方法的思路引入到河道洪水的概率预报中,计算预报流量的概率密度分布和置信区间,获得河道洪水的概率预报结果,为防洪决策提供更丰富的信息。     

数值天气预报的成功经验对洪水预报的启示

鉴于近30年来数值天气预报有了长足的进步,而在洪水预报中试图推行流域水文模型却进展缓慢这一现状,试图寻找洪水预报精度与数值天气预报精度的差距原因。通过对大气模式(型)和流域水文模型的比较,围绕预报方法及预见期、初始场和边值场、数据同化和融合等问题,论述了数值天气预报的成功经验和流域水文模型在洪水预报应用中存在的问题,期望洪水预报能从中得到启示,在不远的将来实现流域水文模型与数值天气预报的无缝衔接,以进一步提高洪水预报精度,增长洪水预报预见期。     

基于流溪河模型的乐昌峡水库入库洪水预报模型研究

采用流溪河模型构建乐昌峡水库入库洪水预报模型,通过"粒子群(PSO)"算法优选模型参数,并对实测洪水过程进行了模拟,对比模型性能。研究发现,采用流溪河模型的乐昌峡水库入库洪水预报性能优良,可满足乐昌峡水库入库洪水预报对精度的要求;模型参数优选可明显提高乐昌峡水库入库洪水预报流溪河模型的洪水模拟精度;"粒子群"算法具有很强的全局优化能力,快速的计算收敛能力,参数优选中种群进化次数在30次以内;乐昌峡水库的建成运行产生了一定的水库洪水效应,10场洪水平均峰现时间提前1.3 h,次洪径流系数增加1.596%,洪峰流量增加0.207%。该模型可用于同类水库入库洪水预报。     

Flood Forecasting and Decision Making in the new Millennium. Where are We?

This paper reviews the development of real time flood forecasting systems from the early 1970 approaches to the recent probabilistic ones. A preliminary discussion on the motivations for developing real time flood forecasting systems is introduced to explain their evolution in the last four to five decades. It will be shown how recent probabilistic flood forecasts are more robust and effective than the traditional deterministic ones. In particular, when combined with Bayesian decision approaches, probabilistic forecasts are the most appropriate tools for rational decision making in flood warning and flood management.Moreover, they allow taking into account the information from several models to be taken into account by combining into a unique predictive density the deterministic predictions of several hydrological or hydraulic models of a different nature, while in the multi-temporal forecasting extensions, they provide to answers questions such as: Which is the probability of overtopping a dyke in the next 24 h? When will this event be more likely to occur during the next 24 h?The work concludes with a discussion on the still unresolved problems, namely how decisions makers can fully take advantage of the probabilistic forecasts and how these forecasts must be communicated to them in order to meet this objective.     

三峡水库实时调度水文气象预报应用风险及控制

为量化水文气象预报误差对水库实时预报调度的影响,以三峡水库为例,基于其历年水文气象预报信息的误差分析成果,假定入库过程,并叠加对应不同概率的预报误差,以此作为水库调洪演算模型的输入。分析了不同预见期、不同保证率水文气象误差条件下,不超过风险控制点的库水位指标,提出了三峡水库实时预报调度风险控制策略,可为水库科学调度提供参考。     

Rainfall-runoff simulation and flood forecasting for Huaihe Basin

The main purpose of this study was to forecast the inflow to Hongze Lake using the Xin＇anjiang rainfall-runoff model. The upper area of Hongze Lake in the Huaihe Basin was divided into 23 sub-basins, including the surface of Hongze Lake. The influence of reservoirs and gates on flood forecasting was considered in a practical and simple way. With a one-day time step, the linear and non-linear Muskingum method was used for channel flood routing, and the least-square regression model was used for real-time correction in flood forecasting. Representative historical data were collected for the model calibration. The hydrological model parameters for each sub-basin were calibrated individually, so the parameters of the Xin＇anjiang model were different for different sub-basins. This flood forecasting system was used in the real-time simulation of the large flood in 2005 and the results are satisfactory when compared with measured data from the flood.     

基于AM-MCMC算法的贝叶斯概率洪水预报模型

本文在贝叶斯预报系统的框架下,利用BP网络能描述非线性映射的特性建立了基于BP网络的先验密度和似然函数的模型,并采用基于自适应采样算法(Adaptive Metropolis algorithm,简称AM)的马尔可夫链蒙特卡罗模拟方法(Markov Chain Monte Carlo,简称MCMC)求解流量的后验密度,最后给出流量的概率预报。实例表明,基于AM-MCMC的BP贝叶斯概率水文预报的精度高,且能给出预报的方差,使得防洪决策可以考虑预报的不确定性。     

水库防洪预报调度的风险分析

本文从水文预报误差的不确定性分析出发，将短期洪水预报精度评定指标转化为入库洪水过程的随机特征值，并引入水库调洪演算随机数学模型，从而实现水文预报风险向预报调度风险的转化，为定量考察预报调度风险率、合理选择动态的汛限水位提供了科学的依据。通过这一方法论证了水文预报精度对水库防洪预报调度风险率的影响，表明提高水文预报精度将有利于降低水库调洪风险率。     

水箱模型在三峡入库洪水预报模型研制中的应用

介绍了应用水箱模型建立三峡入库洪水实时连续预报模型的方法。按照河段流量传播时间将入库站(寸滩)以上干、支流划分为若干子河段,各子河段按照计算时段长分成若干单元河段,各单元河段区间降雨径流预报采用水箱模型,河道流量演算采用连续马斯京根法。对河系洪水预报模型的检验表明,预报精度在85%以上,能够满足实时洪水预报的要求。     

风光水互补系统时间序列变量概率预报框架

风光水互补系统实时调度受风速、太阳辐射强度、径流和电力负荷等时间序列变量的不确定性影响,如何准确预报这些变量并量化预报的不确定性是风光水互补系统面临的关键难题。为此,本研究提出一种基于深度学习模型的时间序列变量概率预报框架。首先,从时间序列数据中挖掘特征输入并采用相关系数对生成的特征进行初选;其次,基于深度学习模型和高斯过程回归构建时间序列变量概率预报模型,同时分别通过0-1规划思想和贝叶斯优化算法实现特征组合优化和超参数优选;进而,从确定性预报、概率预报和可靠性3个方面对预报模型进行全面评价;最后,以雅砻江流域风光水互补先期试点示范基地作为研究对象,分别在径流、风速、光伏和负荷4个数据集上与目前7个不同的时间序列变量预报模型进行全面对比,验证本研究提出预报框架的精度和概率综合性能。     

水库汛限水位实时动态控制模型研究

现行的汛限水位方法一般未能考虑预报信息，因而严重制约了水库汛期兴利效益的发挥。通过对实时调度中的可接受风险的研究，在调度期内控制防洪风险在可接受水平内；讨论实时调度与传统汛限水位之间的关系．由调度末水位耦合汛限水位，由此建立了带有机遇约束的水库汛限水位实时动态控制模型。结合三峡水库围堰发电期的汛期调度，进行了实例研究，结果表明，在不同的预报来水以及预报精度情况下，水库汛限水位实时动态控制模型能在不降低防洪标准的前提下，显著提高水库的兴利效益。