Xinanjiang Model Research Articles

Fast hydrological model calibration based on the heterogeneous parallel computing accelerated shuffled complex evolution method

ABSTRACTHydrological model calibration has been a hot issue for decades. The shuffled complex evolution method developed at the University of Arizona (SCE-UA) has been proved to be an effective and robust optimization approach. However, its computational efficiency deteriorates significantly when the amount of hydrometeorological data increases. In recent years, the rise of heterogeneous parallel computing has brought hope for the acceleration of hydrological model calibration. This study proposed a parallel SCE-UA method and applied it to the calibration of a watershed rainfall–runoff model, the Xinanjiang model. The parallel method was implemented on heterogeneous computing systems using OpenMP and CUDA. Performance testing and sensitivity analysis were carried out to verify its correctness and efficiency. Comparison results indicated that heterogeneous parallel computing-accelerated SCE-UA converged much more quickly than the original serial version and possessed satisfactory accuracy and stability for the task of fast hydrological model calibration.

Improvement of hydrological model calibration by selecting multiple parameter ranges

Abstract. The parameters of hydrological models are usually calibrated to achieve good performance, owing to the highly non-linear problem of hydrology process modelling. However, parameter calibration efficiency has a direct relation with parameter range. Furthermore, parameter range selection is affected by probability distribution of parameter values, parameter sensitivity, and correlation. A newly proposed method is employed to determine the optimal combination of multi-parameter ranges for improving the calibration of hydrological models. At first, the probability distribution was specified for each parameter of the model based on genetic algorithm (GA) calibration. Then, several ranges were selected for each parameter according to the corresponding probability distribution, and subsequently the optimal range was determined by comparing the model results calibrated with the different selected ranges. Next, parameter correlation and sensibility were evaluated by quantifying two indexes, RC Y, X and SE, which can be used to coordinate with the negatively correlated parameters to specify the optimal combination of ranges of all parameters for calibrating models. It is shown from the investigation that the probability distribution of calibrated values of any particular parameter in a Xinanjiang model approaches a normal or exponential distribution. The multi-parameter optimal range selection method is superior to the single-parameter one for calibrating hydrological models with multiple parameters. The combination of optimal ranges of all parameters is not the optimum inasmuch as some parameters have negative effects on other parameters. The application of the proposed methodology gives rise to an increase of 0.01 in minimum Nash–Sutcliffe efficiency (ENS) compared with that of the pure GA method. The rising of minimum ENS with little change of the maximum may shrink the range of the possible solutions, which can effectively reduce uncertainty of the model performance.

Application of BP Neural Network Algorithm in Traditional Hydrological Model for Flood Forecasting

Flooding contributes to tremendous hazards every year; more accurate forecasting may significantly mitigate the damages and loss caused by flood disasters. Current hydrological models are either purely knowledge-based or data-driven. A combination of data-driven method (artificial neural networks in this paper) and knowledge-based method (traditional hydrological model) may booster simulation accuracy. In this study, we proposed a new back-propagation (BP) neural network algorithm and applied it in the semi-distributed Xinanjiang (XAJ) model. The improved hydrological model is capable of updating the flow forecasting error without losing the leading time. The proposed method was tested in a real case study for both single period corrections and real-time corrections. The results reveal that the proposed method could significantly increase the accuracy of flood forecasting and indicate that the global correction effect is superior to the second-order autoregressive correction method in real-time correction.

乘坡河流域洪水预报模型比较研究

中小流域山洪资料积累较少，洪水预报方案编制困难。针对该问题，以南方湿润地区海南乘坡河流域为例，选取以蓄满产流为基础的集总式新安江模型和分布式TOPMODEL模型开展洪水模拟研究。结果表明，对于海南湿润地区，新安江模型和TOPMODEL模型产汇流模拟效果都较好，均可用于指导中小流域进行洪水预报预警。但相对而言，前期土壤较为湿润时，新安江模型的模拟效果较好；对于大洪水及时间序列较长的洪水而言，TOPMODEL模型的模拟效果较好。 The accumulation of flash floods is low and the preparation of flood forecasting schemes is difficult. To address this issue, we take the Chengpo river basin in Hainan, one of the humid areas in south part of China. The lumped Xinanjiang model based on the runoff generation and the distributed TOPMODEL model were selected for flood simulation. The results of both models are good for humid regions in Hainan. These two models can be used to guide the small and medium-sized river basin for flood forecasting and early warning. Generally speaking, Xinanjiang model is more accurate when the pre-soil is wetter; the TOPMODEL model performs better for the larger flood events.

Evaluating functions of reservoirs′ storage capacities and locations on daily peak attenuation for Ganjiang River Basin using Xinanjiang model

Flooding is the most prevalent and costly natural disaster in the world and building reservoirs is one of the major structural measures for flood control and management. In this paper, a framework was proposed to evaluate functions of reservoirs′ locations and magnitudes on daily peak flow attenuation for a large basin of China, namely Ganjiang River Basin. In this study, the Xinanjiang model was adopted to simulate inflows of the reservoirs and flood hydrographs of all sub-catchments of the basin, and simple reservoir operation rules were established for calculating outflows of the reservoirs. Four reservoirs scenarios were established to analyze reservoirs′ locations on daily peak flow attenuation. The results showed that: 1) reservoirs attenuated the peak discharges for all simulated floods, when the flood storage capacities increase as new reservoirs were built, the peak discharge attenuation by reservoirs showed an increasing tendency both in absolute and relative measures; 2) reservoirs attenuated more peak discharge relatively for small floods than for large ones; 3) reservoirs reduced the peak discharge more efficiently for the floods with single peak or multi peaks with main peak occurred first; and 4) effect of upstream reservoirs on peak attenuation decreased from upper reaches to lower reaches; upstream and midstream reservoirs played important roles in decreasing peak discharge both at middle and lower reaches, and downstream reservoirs had less effect on large peak discharge attenuation at outlet of the basin. The proposed framework of evaluating functions of multiple reservoirs′ storage capacities and locations on peak attenuation is valuable for flood control planning and management at basin scale.

An improved hybrid data-driven model and its application in daily rainfall-runoff simulation

In previous literatures, a coupled data-driven rainfall-runoff (RR) model, NU-PEK, has been proposed and successfully applied in hourly RR simulation task. However, numerical experiments show that its performance for daily RR simulation is unsatisfactory. It is noticed that the poor performance is due to the inability of the original model to capture the much higher non-linear characteristics contained in the daily data. In order to improve the nonlinearity simulation capability of the original model, an improved model named NU-PKEK and its calibration methodology are developed in this paper. The improved model is constituted by adding a K-means clustering module and utilizing multiple NU-PEK modules instead of using only one NU-PEK model. This study applies the improved model, the Xinanjiang model, and the original model for daily RR simulation in Chengcun catchment for intercomparison and verification. The simulation results prove that the NU-PKEK performs best, and has better simulation and forecasting capability.

Risk Analysis of Reservoir Flood Routing Calculation Based on Inflow Forecast Uncertainty

Possible risks in reservoir flood control and regulation cannot be objectively assessed by deterministic flood forecasts, resulting in the probability of reservoir failure. We demonstrated a risk analysis of reservoir flood routing calculation accounting for inflow forecast uncertainty in a sub-basin of Huaihe River, China. The Xinanjiang model was used to provide deterministic flood forecasts, and was combined with the Hydrologic Uncertainty Processor (HUP) to quantify reservoir inflow uncertainty in the probability density function (PDF) form. Furthermore, the PDFs of reservoir water level (RWL) and the risk rate of RWL exceeding a defined safety control level could be obtained. Results suggested that the median forecast (50th percentiles) of HUP showed better agreement with observed inflows than the Xinanjiang model did in terms of the performance measures of flood process, peak, and volume. In addition, most observations (77.2%) were bracketed by the uncertainty band of 90% confidence interval, with some small exceptions of high flows. Results proved that this framework of risk analysis could provide not only the deterministic forecasts of inflow and RWL, but also the fundamental uncertainty information (e.g., 90% confidence band) for the reservoir flood routing calculation.

A comparison of simple and complex versions of the Xinanjiang hydrological model in predicting runoff in ungauged basins

There are different views on the selection of hydrological model structural complexity for streamflow prediction in ungauged basins. Some studies suggest that complex models are better than simple models due to the former's prediction capability; whereas some studies favor parsimonious model structures to overcome a risk of over-parameterization. The Xinanjiang (XAJ) model, the most widely used hydrological model in China, has two different versions, as follows: (1) the simple version with seven parameters (XAJ7) and (2) the complex version with 14 parameters (XAJ14). In this study, the two versions of the XAJ model were comprehensively evaluated for streamflow prediction in ungauged basins based on their efficiency, parameter identifiability, and independence. The results showed that the complex XAJ14 model was more flexible than the simple XAJ7 in calibration mode; while the two models have similar performance in validation and regionalization modes. Lack of parameter identifiability and the presence of parameter interdependence most likely explain why the complex XAJ14 cannot consistently outperform the XAJ7 in different modes. Therefore, the simple XAJ7 is a better choice than XAJ14 for streamflow prediction in ungauged basins.

Similarity and difference of global reanalysis datasets (WFD and APHRODITE) in driving lumped and distributed hydrological models in a humid region of China

Different conclusions have been drawn in literature as using reanalysis rainfall products to drive hydrological models for the simulation of streamflow, which warrant a need of further investigation before a generalised conclusion can be drawn. This paper assesses the utility of two widely used reanalysis rainfall datasets (WFD (developed by the WATCH project) and APHRODITE (Asian Precipitation–Highly-Resolved Observational Data Integration Towards Evaluation of the water resources)) against the gauged rainfall in terms of flood simulation in lumped (Xinanjiang Model) and distributed (SWAT Model) hydrological models in a tributary basin of Yangtze River with 94,660km2 drainage area in humid region of southern China. Differences in terms of rainfall accumulation, number of rainy days, spatial patterns of the rainfall amount and frequency distribution of the rain rates are evaluated. The APHRODITE product shows high consistency with the gauged rainfall while WFD data gives large errors in various statistical indices in the study region. Simulated discharges from the gauged and reanalysis rainfall data, respectively, are analysed and compared with the observed discharge in basin outlet over the period 1991–2005. The APHRODITE data show relatively high ability in modelling hydrological responses while the WFD data based models give large error in simulating the discharge. For the assessment of the high flows, both datasets exhibit some skills in flood prediction, however, APHRODITE data perform better than the WFD data when forcing into the lumped Xinanjiang Model than into the distributed SWAT Model in terms of flood duration, Probability of Detection, False Alarm Rate, regressions of annual peaks and partial duration series.

Quantifying the effect on flood regime of land-use pattern changes via hydrological simulation in the upper Huaihe River basin, China

Both landscape and river systems over most regions of the world, within China in particular, have experienced unprecedented changes in the past decades, and their impacts from these changes undoubtedly have altered storm runoff generation characteristics and flood regimes in these areas. To assess the effect on flood regime of land-use pattern changes, the upper Huaihe River basin above the Dapoling station was selected as the case study site. On the basis of the topography, land-use/land cover, hydrological and meteorological data in 1990s and 2000s, quantification of the response to the land-use pattern change of flood characteristics was made by using the semi-distributed Xinanjiang model integrated with a distributed dual-source evapotranspiration model. The results revealed that land-use pattern change had significant impact on runoff depth, runoff coefficient, together with flood peak size, duration and timing. Moreover, land-use pattern change also altered the partitioning of runoff components and rainfall–runoff relationship. The output could provide valuable references for the prediction and mitigation of flood hazards, and also for river ecological health, scientific planning, sustainable development and management of the Huaihe River basin in the future.

Modeling the impacts of water transfer on water transport pattern in Lake Chao, China

Water transport pattern plays an important role in determining water and pollution transport, and is worth our investigation. In this study, a hydrological model (Xinanjiang) and a hydrodynamic model (Environmental Fluid Dynamics Code, EFDC) were coupled to investigate the impacts of a water transfer project on the water transport pattern of Lake Chao, China. The Xinanjiang model simulated the inflow discharges for the EFDC model. The validation results were acceptable with a Nash–Sutcliffe coefficient of 0.74 and 0.84 for the Xinanjiang and EFDC models. Based on the calibrated models, four simulations were implemented and compared to evaluate the impacts of the water transfer project on the water transport pattern, described by water age and residence time. The simulation results revealed that the water transfer project had the potential to shorten water age and residence time of the lake, however, should be carried out by considering the influences of wind conditions and water transfer route. This study demonstrated a successful adaption of water age and residence time for describing water transport pattern, and highlighted the importance of model coupling approach in addressing multidisciplinary questions from the complex natural ecosystem.

An optimization based sampling approach for multiple metrics uncertainty analysis using generalized likelihood uncertainty estimation

This paper investigates the use of an epsilon-dominance non-dominated sorted genetic algorithm II (ɛ-NSGAII) as a sampling approach with an aim to improving sampling efficiency for multiple metrics uncertainty analysis using Generalized Likelihood Uncertainty Estimation (GLUE). The effectiveness of ɛ-NSGAII based sampling is demonstrated compared with Latin hypercube sampling (LHS) through analyzing sampling efficiency, multiple metrics performance, parameter uncertainty and flood forecasting uncertainty with a case study of flood forecasting uncertainty evaluation based on Xinanjiang model (XAJ) for Qing River reservoir, China. Results obtained demonstrate the following advantages of the ɛ-NSGAII based sampling approach in comparison to LHS: (1) The former performs more effective and efficient than LHS, for example the simulation time required to generate 1000 behavioral parameter sets is shorter by 9 times; (2) The Pareto tradeoffs between metrics are demonstrated clearly with the solutions from ɛ-NSGAII based sampling, also their Pareto optimal values are better than those of LHS, which means better forecasting accuracy of ɛ-NSGAII parameter sets; (3) The parameter posterior distributions from ɛ-NSGAII based sampling are concentrated in the appropriate ranges rather than uniform, which accords with their physical significance, also parameter uncertainties are reduced significantly; (4) The forecasted floods are close to the observations as evaluated by three measures: the normalized total flow outside the uncertainty intervals (FOUI), average relative band-width (RB) and average deviation amplitude (D). The flood forecasting uncertainty is also reduced a lot with ɛ-NSGAII based sampling. This study provides a new sampling approach to improve multiple metrics uncertainty analysis under the framework of GLUE, and could be used to reveal the underlying mechanisms of parameter sets under multiple conflicting metrics in the uncertainty analysis process.

Real-time correction of antecedent precipitation for the Xinanjiang model using the genetic algorithm

The Xinanjiang model has been successfully and widely applied in humid and semi-humid regions of China for rainfall–runoff simulation and flood forecasting. However, its forecasting precision is seriously affected by antecedent precipitation (Pa). Commonly applied methods relying on the experience of individual modelers are not standardized and difficult to transfer. In particular, the Xinanjiang daily model may result in obvious errors in the determination of Pa. Thus, a practical method for estimating Pa is proposed in this paper, which is based on a genetic algorithm (GA) and is estimated during a rising flood period. In the optimization process of a GA, Pa values form a chromosome, the root-mean-squared error between the observed and simulated streamflow is chosen as the fitness function. Simultaneously, the best individual reserved strategy is adopted between correction periods to avoid complete independence between each optimization process as well as to ensure the stability of the algorithm. Twenty-seven historical floods observed at the gauge station of the Shuangpai reservoir in Hunan Province of China are used to test the presented algorithm for estimation of Pa, and the results demonstrate that the proposed method significantly improves the quality of flood forecasting in the Xinanjiang model.

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.

Misrepresentation and amendment of soil moisture in conceptual hydrological modelling

Although many conceptual models are very effective in simulating river runoff, their soil moisture schemes are generally not realistic in comparison with the reality (i.e., getting the right answers for the wrong reasons). This study reveals two significant misrepresentations in those models through a case study using the Xinanjiang model which is representative of many well-known conceptual hydrological models. The first is the setting of the upper limit of its soil moisture at the field capacity, due to the ‘holding excess runoff’ concept (i.e., runoff begins on repletion of its storage to the field capacity). The second is neglect of capillary rise of water movement. A new scheme is therefore proposed to overcome those two issues. The amended model is as effective as its original form in flow modelling, but represents more logically realistic soil water processes. The purpose of the study is to enable the hydrological model to get the right answers for the right reasons. Therefore, the new model structure has a better capability in potentially assimilating soil moisture observations to enhance its real-time flood forecasting accuracy. The new scheme is evaluated in the Pontiac catchment of the USA through a comparison with satellite observed soil moisture. The correlation between the XAJ and the observed soil moisture is enhanced significantly from 0.64 to 0.70. In addition, a new soil moisture term called SMDS (Soil Moisture Deficit to Saturation) is proposed to complement the conventional SMD (Soil Moisture Deficit).

Tracing Temporal Changes of Model Parameters in Rainfall-Runoff Modeling via a Real-Time Data Assimilation

Watershed characteristics such as patterns of land use and land cover (LULC), soil structure and river systems, have substantially changed due to natural and anthropogenic factors. To adapt hydrological models to the changing characteristics of watersheds, one of the feasible strategies is to explicitly estimate the changed parameters. However, few approaches have been dedicated to these non-stationary conditions. In this study, we employ an ensemble Kalman filter (EnKF) technique with a constrained parameter evolution scheme to trace the parameter changes. This technique is coupled to a rainfall-runoff model, i.e., the Xinanjiang (XAJ) model. In addition to a stationary condition, we designed three typical non-stationary conditions, including sudden, gradual and rotational changes with respect to two behavioral parameters of the XAJ. Synthetic experiments demonstrated that the EnKF-based method can trace the three types of parameter changes in real time. This method shows robust performance even for the scenarios of high-level uncertainties within rainfall input, modeling and observations, and it holds an implication for detecting changes in watershed characteristics. Coupling this method with a rainfall-runoff model is useful to adapt the model to non-stationary conditions, thereby improving flood simulations and predictions.

基于经验单位线的丰乐河流域洪水预报

经验单位线法是洪水模拟与预报中推求地面净雨汇流过程常用的方法之一。以丰乐河流域为研究区域，基于该流域1984年至2013年的降雨、径流和蒸发资料，选取了30场洪水，采用三水源新安江模型进行次洪的产流与分水源计算，以地面径流与壤中流作为直接径流，利用经验单位线法和线性水库法分别模拟直接径流和地下径流的汇流过程，通过合成直接径流与地下径流，构建了丰乐河流域洪水模拟模型。以新安江模型计算的14场率定洪水的径流深与实测径流深平均误差最小化为目标，采用遗传算法确定了新安江模型的参数。采用分析与试错相结合的方法推求了直接径流的汇流经验单位线。采用大洪水推求的单位线(980915场次)对30场洪水进行模拟，得14场率定洪水的径流深平均相对误差为7%、平均洪峰误差为8%、平均峰现误差为2.57h，合格率分别为93%、93%和86%，平均确定性系数为0.91；16场检验洪水的径流深、洪峰和峰现的误差分别为8%、9%和3.38h，合格率分别为88%、94%和88%，平均确定性系数为0.89。表明基于新安江模型与经验单位线所构建的模型可用于该流域的洪水预报，计算精度较高。采用基于大洪水推求的单位线进行中小洪水汇流计算时的精度较低，反之亦然。推求了相应于大、中小洪水的经验单位线对大、中小洪水分别进行汇流计算，可获得较高的模拟精度。建议在该流域根据形成次洪的降水量的大小，采用不同的经验单位线对洪水进行预报，以提高预报精度。 To calculate ground net rainfall runoff, unit hydrograph method is one of the common methods for flood simulation and prediction. Taking Fengle River basin as a study case, 30 floods were selected based on data of its rainfall, flow and evaporation from 1984 to 2013. The Xinanjiang model was used to calculate the runoff and the separate water sources. Considering surface runoff and interflow as the direct runoff, unit hydrograph and the linear reservoir methods were adopted to simulate the confluence process of the direct runoff and underground runoff respectively. Then flood forecasting model for Fengle River basin was established. With the goal of minimizing the average absolute error between the calculated runoff depths with Xinanjiang model and the observed runoff depths of 14 floods as calibrated samples, the genetic algorithm was adopted to calibrate the parameters of Xinanjiang model. The unit hydrograph of flood confluence of direct runoff was quantified by the method based on the combination of analytical and trial and error approaches. Then 30 floods were simulated with the unit hydrograph acquired through flood data of No. 980915. According to the results, the average relative error of the runoff depth, the error of average flood peak and the peak time of 14 calibrated floods are 7%, 8% and 2.57 h respectively, with the qualification rate are 93%, 93% and 86% respectively, and the average deterministic coefficient is 0.91. The error of the runoff depth, flood peak and the peak time of the 16 validated floods are 8%, 9% and 3.38 h respectively, with the qualification rate are 88%, 94% and 88% respectively, and the average deterministic coefficient is 0.89. Practical applications of the flood forecasting for Fengle River water- shed show that, the model based on the Xinanjiang model and the unit hydrograph can be used to forecast the flood occurred within the watershed, and reaches to a high accuracy. However, the accuracy of concentration calculation of small floods with the unit hydrograph acquired through big flood data usually is lower than those big floods, and vice versa. So unit hydrograph should be calibrated with different rainfall intensities, then using different unit hydrograph to forecast flood processes according to their corresponding precipitation amounts.

Hydrological Drought Forecasting and Assessment Based on the Standardized Stream Index in the Southwest China

Southwest China is abundant of rainfall and water resources, however, severe and extremely droughts hits it more frequently in recent years, caused huge loss of human lives and financial damages. To investigate the feasibility of the standardized stream index in Southwest China, the Nanpanjiang River basin above the Xiaolongtan hydrological station was selected as the case study site. Based on long-term daily hydrological and meteorological data series, the generated runoff was simulated by the daily Xinanjiang model, then the standardized stream index was calculated and its feasibility was explored by comparing it with other two hydrological drought index. The result revealed that the standardized stream index performed well in detecting the onset, severity and duration in 2009/2010 extremely drought. The output of the paper could provide valuable references for the regional and national drought monitoring and forecasting systems.

TRMM卫星降雨数据的精度及径流模拟评估

卫星测雨的精度已成为人们关注的热点，并且在水文研究及应用上有很大的发展前景。本文选取湘江流域为研究区，以地面雨量站点数据为参照，评估了不同时空尺度下热带测雨观测计划(Tropical Rainfall Measuring Mission, TRMM)多卫星降雨分析新产品(TRMM 3B42 V7)的精度，同时分析了精度在不同时期、雨强及高程等因素下的影响规律，并驱动新安江模型展现了径流模拟效果。研究表明：TRMM降水数据与站点数据相关性较好，TRMM数据精度在月尺度明显优于日尺度，流域尺度优于栅格尺度，面雨量精度湿润季节高于干旱季节；雨量估计上TRMM相对站点在流域尺度下整体低估，栅格尺度下受高程影响，在高程较大区域TRMM数据倾向低估，反之倾向高估；各雨强级别下TRMM对小强度降雨稍低估而对强降雨和无雨高估；TRMM数据月径流模拟效果比日径流好，均在可接受范围，但对洪峰、洪量及峰现时间不如站点模拟精准。本研究为TRMM数据的精度订正提供相关思路，并分析其在径流模拟应用中的可行性。 The accuracy of satellite precipitation estimates has become a focus of attention, having a promising prospect in the hydrological researches and applications. In this study, the new precipitation product (3B42 V7) of Tropical Rainfall Measuring Mission (TRMM) was evaluated via comparison with the rain gauge precipitation data in Xiangjiang River Basin. The evaluations were conducted at different spatial and temporal scales. Simultaneously, the laws of data accuracy under different periods, rainfall intensity levels, and elevations were analyzed. These two precipitation data were then adopted to drive the Xinan-jiang Model to perform streamflow simulation. The results showed that TRMM satellite precipitation data had a good correlation with rain gauge precipitation data; it had better accuracy under monthly scale than daily scale and showed better performance at watershed scale than it did at grid scale. The area precipitation of TRMM data showed better accuracy in humid season than in arid season. From an overall estimate on the amount of precipitation, areal precipitation from TRMM data on watershed scale was underestimated in contrast with areal precipitation interpolated from rain gauge data. While at grid scale, TRMM data from higher elevations inclined toward underestimation, while the contrary inclined toward overestimation. The frequency statistical result of different rainfall intensity levels indicated that TRMM tended to underestimate the occurrence of light rainfall intensity but overestimate the occurrence of heavy rainfall intensity and no-rain events. The TRMM-driven streamflow simulation performed better at monthly scale relative to the daily scale, but all were within acceptable range. However, the effects of streamflow simulation driven by TRMM data were not as accurate as those driven by rain gauge data displayed in the aspects of flood peak, flood volume and flood duration. This study assesses the accuracy of the latest product TRMM 3B42 V7, revealing the influential factors regarding accuracy for future reference and analyzing the effectiveness and possibility of streamflow simulation.

Uncertainty of Flood Forecasting Based on Radar Rainfall Data Assimilation

Precipitation is the core data input to hydrological forecasting. The uncertainty in precipitation forecast data can lead to poor performance of predictive hydrological models. Radar-based precipitation measurement offers advantages over ground-based measurement in the quantitative estimation of temporal and spatial aspects of precipitation, but errors inherent in this method will still act to reduce the performance. Using data from White Lotus River of Hubei Province, China, five methods were used to assimilate radar rainfall data transformed from the classifiedZ-Rrelationship, and the postassimilation data were compared with precipitation measured by rain gauges. The five sets of assimilated rainfall data were then used as input to the Xinanjiang model. The effect of precipitation data input error on runoff simulation was analyzed quantitatively by disturbing the input data using the Breeding of Growing Modes method. The results of practical application demonstrated that the statistical weight integration and variational assimilation methods were superior. The corresponding performance in flood hydrograph prediction was also better using the statistical weight integration and variational methods compared to the others. It was found that the errors of radar rainfall data disturbed by the Breeding of Growing Modes had a tendency to accumulate through the hydrological model.