Hydrologic Modelling Calibration for Operational Flood Forecasting

Runoff prediction in flood forecasting depends on the use of hydrological simulation models and on the input of accurate precipitation forecasts. Reliability of predictions thus obtained hinges on proper calibration of the model. Moreover, when the model is intended to be used systematically in operational forecasting of streamflows, the calibration process must take into account the variation of the model parameters over time, namely in response to changing weather and hydrological conditions in the basin. The goal of the study was to build a process to adjust, on a daily basis, the simulation model parameters to the current hydrological conditions of the river basin, in order for the model to be run operationally for prediction of the streamflow for the next 10-days period, and, thereby, to forecast the occurrence of flood events. Towards this end, hydrological simulations using the HEC-HMS model were performed, using a 3 h period time step. The present communication focuses on the hydrological model calibration and verification processes and on the evaluation of forecasts’ accuracy. The procedure was applied to a part of the largest (full) Portuguese river basin, the Mondego river basin, corresponding to the Aguieira dam section watershed, which comprises an area of 3070 km². Four wet periods, associated with the occurrence of flooding, were selected for the calibration and verification of the model, by adjustment of the model parameters. The results of the study aim to define the optimal calibration parameters values to model the observed streamflow for various hydro-meteorological states, thus enabling adequate prediction of flow in flooding situations and proper application of the model in operational flood forecasting.     

系统微分响应参数率定方法在建阳流域SWAT模型中的应用

利用基于系统微分响应理论的参数率定方法率定闽江建阳流域SWAT模型,通过流域1992—2000年日资料验证了该方法的实际应用效果,采用纳什效率系数(NSE)、偏差百分比(PBIAS)和相关系数r作为评价指标对率定后的SWAT模型模拟精度进行评价,并与传统利用SUFI-2方法率定得到的SWAT模型模拟结果进行对比分析。结果表明：系统微分响应方法率定的参数在建阳流域的日径流模拟中表现较好,率定期和检验期NSE均为0.65以上,径流量PBIAS在5%以内,r在0.65以上;利用系统微分响应率定参数方法率定SWAT模型参数在实际应用中可行,且效率与精度均高于传统的SUFI-2方法。     

Evaluation of Using Remote Sensing Evapotranspiration Data in SWAT

This study applied a time series evapotranspiration (ET) data derived from the remote sensing to evaluate Soil and Water Assessment Tool (SWAT) model calibration, which is a unique method. The SWAT hydrologic model utilized monthly stream flow data from two US Geological Survey (USGS) stations within the Big Sunflower River Watershed (BSRW) in Northwestern, Mississippi. Surface energy balance algorithm for land (SEBAL), which utilized MODerate Resolution Imaging Spectro-radiometer (MODIS) to generate monthly ET time series data images were evaluated with the SWAT model. The SWAT hydrological model was calibrated and validated using monthly stream flow data with the default, flow only, ET only, and flow-ET modeling scenarios. The flow only and ET only modeling scenarios showed equally good model performances with the coefficient of determination (R²) and Nash Sutcliffe Efficiency (NSE) from 0.71 to 0.86 followed by flow-ET only scenario with the R² and NSE from 0.66 to 0.83, and default scenario with R² and NSE from 0.39 to 0.78 during model calibration and validation at Merigold and Sunflower gage stations within the watershed. The SWAT model over-predicted ET when compared with the Modis-based ET. The ET-based ET had the closest ET prediction (~8% over-prediction) as followed by flow-ET-based ET (~16%), default-based ET (~27%) and flow-based ET (~47%). The ET-based modeling scenario demonstrated consistently good model performance on streamflow and ET simulation in this study. The results of this study demonstrated use of Modis-based remote sensing data to evaluate the SWAT model streamflow and ET calibration and validation, which can be applied in watersheds with the lack of meteorological data.     

妫水河流域水文模拟及参数不确定性分析

以北京市延庆区妫水河为例,使用SWAT(Soil and Water Assessment Tool)模型对妫水河流域进行月尺度水文模拟,使用SUFI-2(Sequential Uncertainty Fitting)算法分析参数的敏感性,依据SWAT-CUP自动率定得到P因子和R因子分析模型的不确定性,从而完成本流域分布式水文模型的构建。率定结果显示,率定期确定系数R 2为0.65,效率系数NSE为0.61;验证期确定系数R 2为0.89,效率系数NSE为0.88;不确定性分析结果中P-factor均大于0.5,R-factor均小于0.3。通过以上分析可得该模型对妫水河流域的水文模拟有良好的效果。     

目标函数对滦河流域SWAT模型径流模拟不确定性的影响

针对水文模型参数和径流模拟结果不确定性问题,选取2Nash-Sutcliffe效率系数(NSE)、改进的决定系数(R_m²)、相对误差(PBIAS)、Kling-Gupta效率系数(KGE)4种目标函数,对构建的滦河流域潘家口水库上游SWAT模型进行参数率定及验证,分析了不同目标函数下模型参数的敏感性差异及径流模拟的不确定性。结果表明：参数敏感性会随迭代次数增加和抽样范围变化发生改变,不同目标函数下率定的参数范围和最优值显著不同;NSE和KGE作为目标函数在各站点径流模拟中更稳健,分别表现出较高的模拟精度和较低的模拟不确定性。     

大伙房模型在石佛寺流域洪水预报方案中的应用

为了提高辽河流域洪水预报精度,缓解流域下游的防洪压力,为水库防洪调度提供科学依据,现以水文自动测报系统为技术支持,以大伙房模型为例研究了石佛寺水库的洪水预报方案,并使用粒子群算法对大伙房模型进行了参数优选。计算结果表明,模拟预报结果都属于甲级方案,预报方案合理,并且符合石佛寺流域的结构特点,充分反映了洪水的时空变化,可用于石佛寺水库的预报调度。     

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。     

基于CMADS驱动下SWAT模型的敖江流域径流模拟

目前针对气象、水文资料不足的小尺度流域进行径流模拟的研究相对较少。选取敖江流域为研究区,利用中国大气同化驱动数据集(CMADS)驱动SWAT模型对该流域2008—2016年进行逐月径流模拟,并使用水文比拟法结合邻近的2个传统气象站计算出模拟时段的观测数据以完成模型参数的率定与验证。结果表明:模拟结果与观测值较吻合,率定期(2010—2013年)评价指标决定系数R²和Nash-Suttcliffe系数NS分别为0.84和0.76,验证期(2014—2016年)两者分别为0.85和0.74,均达到了模型的评价要求。研究成果表明CMADS驱动下的SWAT模型适用敖江流域的径流模拟,采用水文比拟法适合在水文资料不足的地区进行径流计算,可为缺少气象与水文资料的小尺度流域进行径流模拟提供参考。     

HYPE模型在成屏一级水库的适用性研究

水文模型是模拟水库入库流量的重要工具之一。以浙江省遂昌县成屏一级水库为研究对象,基于LHOAT方法、DE方法和HYPE模型,通过参数敏感性分析、参数率定和误差分析,评估了HYPE模型对成屏一级水库入库流量模拟的适用性。结果表明:HYPE模型中极敏感的参数为cevp,较敏感的参数为pcluse、cevpcorr、rrcscorr、rrcs1; HYPE模型在率定期的纳什效率系数(NSE)为0. 87,在验证期的纳什效率系数(NSE)为0. 83,这表明HYPE模型具备在长序列日尺度水文模拟的能力; HYPE模型在洪峰模拟上模拟值低于实测值,原因可能是输入数据的精确性不足以及率定的模型参数并不能较好地反映出暴雨过程中的流域特征;对暴雨情况采用新的参数进行流域特征刻画,模型模拟效果更好。     

淤地坝对流域水沙影响模拟研究

使用SWAT模型,结合淤地坝特点对模型自带的水库模型进行修正来设置淤地坝模块,研究淤地坝对流域径流和输沙的影响,并在内蒙古十大孔兑之一的西柳沟流域进行模拟和验证。以1980-1990年为率定期、2006-2015年为验证期,龙头拐水文站模拟径流量和输沙量与实测值拟合较好,线性拟合系数R²和纳什效率系数E_NS均超过0.6,说明SWAT模型中的水库模型可用于淤地坝模拟。结果表明：其他参数不变的情况下,淤地坝对流域的径流量有一定影响,能够拦截一部分径流量,对流域输沙量的影响巨大,减沙效果明显;淤地坝在一定程度上能影响流域的汇流过程,使得汛期后的月份中出现模拟径流量大于实测径流量的现象;相较于经验公式法,SWAT模型能够提高淤地坝对流域水沙影响的模拟精度;为更好地利用SWAT模型模拟淤地坝对流域径流和输沙的影响,淤地坝模块仍有待改进。     

一种高效的SWAT模型参数自动率定方法

本文分析了SWAT模型和PSO算法的原理,将PSO算法引入SWAT模型中,构建了新的SWAT模型参数自动率定模块,通过在天津蓟运河流域实例研究,发现该方法率定精度较高,收敛速度快,运行结果稳定,整体率定效果优于模型自带的参数率定模块。如果用改进后的模块在Linux平台开展自动率定,可以使模型自动率定效率提高到当前水平的7倍,适用于大型流域或长时间系列模拟。而PSO算法作为一种通用的优化算法,可广泛用于各种水文模型的参数率定。     

SWAT和WetSpa Extension在罗玉沟流域的应用比较

近年来,运用水文模型来研究流域水文状况越来越普遍,本文以黄土高原丘陵沟壑区第三副区的典型小流域为例,基于SWAT和WetSpa Extension两个模型,从模型的输入数据、模型的参数和模型模拟结果等方面进行比较,以期为该区水沙模拟选择合适水文模型并为将来模型性能的改进提供参考。连续15年的校准验证表明:两个模型都能得到较好的结果,而SWAT模型能应用于水文及泥沙等模拟过程,相对来说功能更全面;WetSpa Extension模型目前仅能提供水文模拟的模块,优点在于模拟时间步长较任意,可进行洪水模拟。     

基于PDM的飞来峡水库实时水文预报模型

水文预报模型是实时预报预警系统的核心。基于新型的概率水文预报PDM(The Probability Distributed Model),构建了北江流域飞来峡水库概率水文预报模型。将整个流域划分为18个子流域,对每个子流域分别建立PDM,并采用历史降雨流量数据序列对各子流域和流域PDM进行参数率定;以横石站为参考站,采用52场次实测历史洪水进行模型验证和精度评定,洪峰预报合格率为99%,精度等级为甲等,模拟精度高,说明了PDM结构和参数取值的合理性,以及模型在该流域的适用性。     

Evaluating the SWAT Model for Hydrological Modeling in the Xixian Watershed and a Comparison with the XAJ Model

Already declining water availability in Huaihe River, the 6th largest river in China, is further stressed by climate change and intense human activities. There is a pressing need for a watershed model to better understand the interaction between land use activities and hydrologic processes and to support sustainable water use planning. In this study, we evaluated the performance of SWAT for hydrologic modeling in the Xixian River Basin, located at the headwaters of the Huaihe River, and compared its performance with the Xinanjiang (XAJ) model that has been widely used in China. Due to the lack of publicly available data, emphasis has been put on geospatial data collection and processing, especially on developing land use-land cover maps for the study area based on ground-truth information sampling. Ten-year daily runoff data (1987?C1996) from four stream stations were used to calibrate SWAT and XAJ. Daily runoff data from the same four stations were applied to validate model performance from 1997 to 2005. The results show that both SWAT and XAJ perform well in the Xixian River Basin, with percentage of bias (PBIAS) less than 15%, Nash-Sutcliffe efficiency (NSE) larger than 0.69 and coefficient of determination (R²) larger than 0.72 for both calibration and validation periods at the four stream stations. Both SWAT and XAJ can reasonably simulate surface runoff and baseflow contributions. Comparison between SWAT and XAJ shows that model performances are comparable for hydrologic modeling. For the purposes of flood forecasting and runoff simulation, XAJ requires minimum input data preparation and is preferred to SWAT. The complex, processes-based SWAT can simultaneously simulate water quantity and quality and evaluate the effects of land use change and human activities, which makes it preferable for sustainable water resource management in the Xixian watershed where agricultural activities are intensive.     

Utilization of the Bayesian Method to Improve Hydrological Drought Prediction Accuracy

This study established a hydrological drought forecasting system based on the Bayesian method and evaluated its utilization for South Korea. The regression result between Historical Runoff (HR) and Ensemble Streamflow Prediction Runoff (ESP_R) was used as prior information in the Bayesian method. Additionally Global seasonal forecast System 5 Runoff (GS5_R) produced using a dynamic prediction method was used in a likelihood function. Bayesian Runoff (BAY_R), as posterior information, was generated and compared with the ESP_R and GS5_R results for predictive ability evaluation. The Standardized Runoff Index (SRI) was selected for the drought prediction, and the BAY_SRI, GS5_SRI and ESP_SRI were computed using BAY_R, GS5_R and ESP_R, respectively. The Correlation Coefficient (CC), Nash-Sutcliffe Efficiency (NSE) and Receiver Operating Characteristic (ROC) score of BAY_SRI were the highest, and the Root Mean Square Error (RMSE) of BAY_SRI was the lowest among the methods. The Bayesian method improved the behavioral and quantitative error of drought prediction and the predictive ability of the occurrence of drought. In particular, the simulation accuracy was significantly improved during the flood season. Additionally, BAY_SRI represented past drought scenarios better than did the other two methods. Overall, we found that the Bayesian method could be applied for hydrological drought predictions for based on 1- and 2-month lead times.     

基于智能聚类识别的 SWAT 模型参数优化方法 −以石头口门水库流域为例

为探究 SWAT 模型参数优化过程与方法,降低参数估计不确定性,采用敏感性分析方法遴选关键参数,针对 关键参数采用拉丁超立方抽样构建参数样本集,进而结合各组关键参数组合下的模拟精度指标构建聚类指标集, 采用 SOM 聚类算法进行聚类,并基于模拟精度较高且波动较小类别识别各关键参数取值范围,形成一种 SWAT 模型关键参数优化系统方法。以石头口门水库流域为例,选取 1980—2016 年（1980—1986 年为预热期,1987— 2009 年为率定期,2010—2016 年为验证期）的月径流实测资料,建立流域 SWAT 模型,引入 SOM 聚类算法进行参 数优化,不断缩小模型关键参数合理取值区间,并应用 SUFI-2 算法进行模拟结果对比。结果表明：SWAT 模型适 用于石头口门水库流域,且参数优化前验证期的决定系数 R2为 0.79,纳什效率系数 ENS为 0.74,P-factor 为 0.65,R-factor 为 0.56;参数优化后验证期 R 2为 0.88,ENS为 0.83,P-factor 为 0.70,R-factor 为 0.50,模拟效果较好。故 应用 SOM 算法进行 SWAT 模型参数优化可以降低模型不确定性,提高径流模拟精度,为水文模型参数优化算法 的选择提供思路,对水资源管理政策制定与水库优化调度具有重要意义。     

洪水预报系统在左江洪水预报中的应用研究

分析了左江流域的水文气象条件和崇左水文站的洪水特征及成因。介绍了利用＂洪水预报系统＂的应用平台建立左江崇左水文站产汇流预报模型,并进行模型参数率定,该方案在精度评定及实际应用中取得较满意的效果,可用于发布过程及洪峰预报,为崇左市及下游的防汛抢险提供决策依据。     

Application of hydrometeorological coupled European flood forecasting operational real time system in Yellow River Basin

This study evaluated the application of the European flood forecasting operational real time system (EFFORTS) to the Yellow River. An automatic data pre-processing program was developed to provide real-time hydrometeorological data. Various GIS layers were collected and developed to meet the demands of the distributed hydrological model in the EFFORTS. The model parameters were calibrated and validated based on more than ten years of historical hydrometeorological data from the study area. The San-Hua Basin (from the Sanmenxia Reservoir to the Huayuankou Hydrological Station), the most geographically important area of the Yellow River, was chosen as the study area. The analysis indicates that the EFFORTS enhances the work efficiency, extends the flood forecasting lead time, and attains an acceptable level of forecasting accuracy in the San-Hua Basin, with a mean deterministic coefficient at Huayuankou Station, the basin outlet, of 0.90 in calibration and 0.96 in validation. The analysis also shows that the simulation accuracy is better for the southern part than for the northern part of the San-Hua Basin. This implies that, along with the characteristics of the basin and the mechanisms of runoff generation of the hydrological model, the hydrometeorological data play an important role in simulation of hydrological behavior.     

南四湖洪水预报调度模型研究

南四湖流域水文特性错综复杂,至今没有比较实用的洪水预报调度系统,因此,开展南四湖洪水预报调度模型研究具有重要的理论意义和实际应用价值。在采用新安江模型率定有实测资料的11个子流域后,对于湖区周围无实测资料区间,进行邻近子流域参数移植处理,并通过分析模拟结果在一定程度上掌握了流域水文特性。最后根据实际需要确立2种防洪调度方式,其中人工控制方式充分考虑了调度人员的经验和调度中出现的意外情况,保证了实际洪水调度的灵活性。     

SWAT模型在金沙江上游小流域径流模拟中的应用

为探讨SWAT分布式水文模型在金沙江上游小流域水文过程模拟中的适用性,选择金沙江一级支流冲江河流域为研究区,利用该小流域控制站来远桥水文站1961~2012年月径流量数据进行模型参数率定与模型验证研究。结果表明：SWAT模型对金沙江上游山区小流域径流模拟精度较高,冲江河流域率定期和验证期月径流模拟值与实测值的相关系数分别达0.97和0.81,平均相对误差分别为16%和20%,纳什系数Ens分别为0.84和0.82。