Journal of Hydrologic Engineering

Runoff in major rivers in China has been decreasing in recent decades. The attribution of hydrologic variability to human activity or climate change is a challenging problem and an active research area. In this study, a sequential cluster analysis method was used to detect undisturbed parts of the record for the Kuyehe River catchment in China. The variable infiltration capacity (VIC) land-surface model was calibrated and verified using observed hydrometeorological data from a period with relatively little human-induced disturbance. The calibrated VIC model was then used to simulate natural runoff during the human-regulated period. Results indicate that the recorded runoff at Wenjiachuan station had significant decline trend of $-1.34\mathrm{mm}/\mathrm{year}$. Time series of runoff was detected and divided into three epochs at two critical years of 1980 and 1998. The VIC model performed well in simulating monthly discharges in the catchment, both Nash and Sutcliffe efficiency (NSE) critera in calibration (1955 to 1969) and verification (1970 to 1979) were above 70% while relative errors (REs) were less than 5%. Human activity was the main driver behind 68% of the runoff reduction that occurred for the period of 1980 to 2008.

Climate change has become an environmental issue of utmost importance, and one that will challenge existing water resource management practices in many ways. The Huaihe River, one of China’s major rivers, is frequently subject to flooding and drought and, for the purposes of assessing the implications of climate change on water resources in the Huaihe River Basin, the variable infiltration capacity (VIC) model with a resolution of $0.5°\times 0.5°$ was calibrated using data from 11 well-gauged subbasins. The model parameters from the well-gauged stations were then transferred to poorly gauged areas according to similarities in, for example, climate conditions and soil texture. The calibrated VIC model was subsequently used to study the potential impacts of three climate change scenarios on basin runoff, taking projected runoff for 1961–1990 as the baseline. In general, the results showed that although the annual runoff will likely increase across the basin under the different scenarios, regional flooding and regional shortage in water resources will be exacerbated under the impacts of global warming.

In China, design flood computation is generally used by the design flood hydrograph (DFH) method by amplifying the typical flood hydrograph (TFH). TFH selection plays an important role in design flood calculation because the selection result directly affects the final DFH. The traditional method of TFH selection is subjective and varies with designers. So the fuzziness of TFH selection should be considered; the theory of fuzzy pattern recognition is applied in TFH selection to explore a theoretical method. This paper establishes a fuzzy optimization model based on entropy weight to select TFH. The influence indexes reflecting the quality of TFH selection are determined by Pearson correlation test and Spearman’s rho test. A shape parameter, which quantitatively described the shape of the flood hydrograph, is proposed for the first time. The entropy weight method is used to determine the weights of influence indexes and is a better way to avoid subjective influence. The optimal TFH is obtained by the principle of minimum eigenvalue of grades through fuzzy optimization model calculation. The model is applied to Yuecheng Reservoir in the Zhanghe Basin. Results show that No. 520723 flood process is the optimal TFH. Compared with the traditional TFH (1956 TFH), the TFH selected with this model is relatively better to design flood computation.

One-dimensional (1D) hydrodynamic models based on the Saint-Venant equations (the SVN model) have been extensively used for flow simulation in tidal rivers. In rivers with significant tidal effects, the inconsistencies between actual flow characteristics and simplifying assumptions that the Saint-Venant equations are derived based on can introduce unignorable errors into the application of the SVN model. To reduce the errors contributed by the factors the SVN model can not completely account for, such as the tidal effect, an approach of incorporating a function into the calculated velocity by the SVN model was proposed in this study. Fourier series was employed to construct the function. By combining the proposed approach with the flow continuity equation, a 1D hydrodynamic model (the FSV model) was constructed to improve the performance of the SVN model in tidal rivers. The simulation results in the tidal reaches of the Qiantang River and the Changjiang River show that the FSV model performs better than the SVN model alone. It indicates that the proposed approach is reasonable and effective to reduce the errors due to the factors the SVN model can not completely considered in applications.

The Three Gorges Project along the Yangtze River in China, as one of the biggest hydropower-complex projects in the world, plays a significant role in the economic development of the area even of the whole nation. An accurate and reliable flood forecast modeling system is of significant importance for flood control, flood warning, and operation of larger reservoirs. Kalman filter coupling with hydrological models or hydraulic models is one of the efficient methods to adjust real-time flood series for reducing errors from model structure, input data, and calibrated parameters. However, the coupling model is time consuming in computation because the state vectors in this kind of Kalman filter including both water stage and discharge are solved simultaneously. In this study, an alternative coupling method was developed, which separates system state equations and measurement equations allowing the water stage and discharge to be computed alternately. The new method was applied for real-time flood forecasting in the Three Gorges interzone of Yangtze River. The hydraulic model is calibrated and verified against the observed flood stage and discharge before and during Three Gorges Dam construction periods. Study results demonstrate that the new model is efficient in real-time flood forecasting. A comparative study shows that the newly developed approach outperforms the conventional methods in terms of modeling efficiency, root mean square error, as well as the forecasting errors in the maximum water stage and peak flow.

Bayesian model averaging (BMA) is a statistical method that can synthesize the advantages of different models or methods. The objective of this research is to explore the use of BMA to forecast combinations among several hydrological models. BMA is a statistical scheme that infers the posterior distribution of forecasting variables by weighing individual posterior distributions based on their probabilistic likelihood measures, with the better performing predictions receiving higher weights than the worse predictions. The Topographic Kinematic Approximation and Integration and Xin’anjiang models were applied to the Dongwan basin, Yellow River, China, for flood simulation. Observed and simulated discharge time series were transformed into normally distributed variables through the normal quantile transform. The Gaussian mixture model was constructed by weighing the posterior distribution of individual hydrological models in the transformed space. The posterior probability measuring samples belonging to each specific hydrological model were treated as the weights. The parameters of the Gaussian mixture model and the weight of each hydrological model were estimated by the expectation maximization algorithm. Thus, the forecast combination in the catchment was obtained from the two hydrological models. For flood forecasting, the results provided not only the mean discharge values, but also quantitative evaluation of forecasting uncertainties (e.g., standard deviation and confidence interval), because the BMA approach calculated the estimation of the probability distribution of forecasted variables.

In this research, considering the importance and complexity of the water-level forecast in tidal reach, a case study was done in the tidal reach of the Yangtze River. Through the analysis of the characteristics of water-level variations, tidal harmonic analysis was introduced into tidal reach, which was commonly applied in oceanic water-level forecast. An automatic partial tide selection and optimization pattern was used to calibrate harmonic constants, long-term water-level forecast models were developed for Yanglin, Xuliujing, Tianshenggang, and Jiangyin stage stations, and then the long-term forecast models were used to hindcast historical water-level events at the aforementioned stations. Most of the results were highly accurate, but there were also some occasional significant errors, which could be attributed to the effects of upstream runoff and downstream storm surge. Based on these findings, response functions of water-level forecast error to the variation of upstream runoff and the variation of storm surge were developed, which were used in the real-time correction of the simulation results obtained from the long-term forecast model. The forecast accuracy was evidently improved. These models have already been used in water-level forecast in the tidal reach of the Yangtze River for the construction of two large bridges, and they also have the potential to contribute to water-level forecast in similar tidal reaches of other rivers.

This paper investigates the probability of integrating the Xinanjiang model, the most widely used watershed hydrological model in China, with the fully distributed groundwater model MODFLOW for processing a coupled flow simulation. Based on the flow simulation of the Xinanjiang and the SWAT model in the case study area, the correlation between the groundwater recharge and the total flow runoff is analyzed, which is then applied to compute the groundwater recharge in subbasins in the Xinanjiang model. By applying a loosely coupled modeling approach, the spatially distributed groundwater recharge is introduced into MODFLOW with the RCH package via the interface of the subbasins in the Xinanjiang model and the cells in MODFLOW; thus the integration of the two models is successfully realized. The application of the integrated model simulation in Zhangji, a groundwater-dominated area in the Huai River basin, China, demonstrates that the coupled Xinanjiang-MODFLOW simulation produces reasonable groundwater levels comparing to the observations. However the study also shows the need for further research in regarding to the improvement on Xinanjiang model in order to enable its direct integration with MODFLOW. Furthermore, the approach presented in this paper is limited to a unidirectional coupling of Xinanjiang and MODFLOW, which only accounts for groundwater level simulations, while the possible feedbacks from the surface water is neglected. More efforts are required to further facilitate the applicability and capability of the models integration.

The prediction of rainfall is the premise of responding to water resources management and flood defense. The hydrological system is complex, and rainfall time series as a member of it has characteristics of being nonlinear and nonstationary, so accurate rainfall prediction still a much difficult job at present. The paper proposes a conjunction model named EMD-KNN for forecasting annual average rainfall. The model is improved by combining two methods, empirical mode decomposition (EMD) and the K-nearest neighbor bootstrap regressive model (K-NN). It is applied to case studies of forecasting annual average rainfall for Nanjing city and the Dahuofang reservoir basin. Where Nanjing city is water-rich area in East China, Dahuofang reservoir basin is water-deficient area in Northeast China. Three performance evaluation measures results revealed that the EMD-KNN model reduces the prediction mean absolute error (MAE), mean relative error (MRA), and root-mean square error (RMSE) with respect to the single K-NN model by almost 50% each, so the suggested model can effectively improve the forecast accuracy of single K-NN in forecasting the annual average rainfall.

The middle-term and long-term runoff forecasting model of the hydropower station reservoir is established with the Bayesian statistic forecasting theory; uncertainty of the hydrological forecasting is quantitatively described in the form of a probability distribution to explore the statistic forecasting theory and its application value. The uncertainty of the input factor is processed with the forecasting model of grey correlation of meteorological factors, and real-time weather data are effectively combined with the historical hydrological data to break through the restriction of traditional deterministic forecasting methods in the aspects of information utilization and sample study to improve the precision of hydrological forecasting. The established model has been assessed by the example of the reservoir of the Fengman hydropower plant. It is indicated by the analog computation result that this model, compared with the deterministic runoff forecasting method, has advantages not only in quantitatively considering the uncertainty in decision making, but also in improving the precision of runoff forecasting in the expected significance, and has comparatively high application value.

Land-use change has significant impacts on hydrologic processes at the watershed level; thus, quantitative assessment on the impacts of land-use change is vital for basin environment protection and water resources sustainable development. Owing largely to computer and geographical information system (GIS) technology improvements, the distributed hydrological models, which allow describing the temporal variability and spatial distribution of water-balance components, offer an effective approach to quantify the land-use change effects on watershed water quantity. In this study, a soil and water assessment tool model was used to simulate land-use change effects on water quantity in the upper Huaihe River basin in China above the Xixian hydrological controlling station with a catchment area of $\mathrm{10,190}{\mathrm{km}}^2$ by the use of temporal three-phase (1980s, 1990s, 2000s) land-use maps, soil type map (1:200,000 scale), and 1980–2008 daily time series of rainfall from the upper Huaihe River basin. Within the model, potential evapotranspiration was computed using the Penman-Monteith method coupled with a simplified plant growth model. On the basis of the simulated time series of daily runoff, land-use change effects on spatio-temporal change patterns of runoff coefficients and runoff modules, the rainfall-runoff relationship, the sensitivity of rainfall-runoff relationship to rainfall for different types of land use, and impact of land-use patterns on rainfall-runoff relationships were investigated. The results revealed that under the same condition of soil texture and terrain slope, the advantage for runoff generation and the sensitivity of rainfall-runoff relationship to rainfall decreased by farmland, paddy field, and woodland. With the same rainfall, the advantage for runoff generation increased by the 1990s, 2000s, and 1980s land-use patterns. The outputs could provide important references for soil and water conservation and river health protection in the upper stream reach of the Huaihe River.

A prompt availability of precipitation intensity is crucial for accurate and effective flood forecast. When radars are coordinated, the mosaic of weather radar can perform rainfall estimation over large areas. This study presents an experimental method of rainfall estimation based on multiple-Doppler radars over the Huaihe River Basin in China. The method emphasizes the combined use of six radars in the area. The processes reviewed are base-data collection, conversion from radar reflectivity to rain rate, rain gauge calibration, consensus of rainfall, and rainfall mosaics. A case study of an extreme meteorological and hydrological event, which occurred in July 2007 over the Huaihe River basin in China, was examined, and the estimated error of rainfall is lower than 45% after calibration with measurements from rain gauges. The results show that the proposed rainfall estimation method presents a cumulative rainfall distribution that is similar to the observations. It is both practical and reliable for the Huaihe River Basin.

Vegetation plays an important role in water and energy cycles on land surfaces. Nevertheless, the physical effects of vegetation are not explicitly considered in many hydrologic-modeling works. In this study, a coupled land-surface-hydrologic model was used to investigate the vegetation effects on hydrologic processes over the years 1980–1987 in the Huaihe River basin, China. The vegetation coverage of the basin was assessed by the International Satellite Land Surface Climatology Project (ISLSCP) Initiative II historic and potential land cover data set. Farmland was declared the dominant vegetation type of the basin in 1970 by the historic land cover scenario, and the potential vegetation cover is mixed forest. Firstly, the coupled model was calibrated by observed streamflow at Bengbu station. The correlation coefficient and Nash-Sutcliffe coefficient of efficiency of multiannual daily series were 0.987 and 0.968, respectively, which indicate that the capability of the coupled model system is acceptable. Next, typical hydrologic physical processes, including evapotranspiration (ET), soil moisture, surface runoff, and groundwater, were simulated and analyzed for the two vegetation scenarios with same meteorological data, initial conditions, and model parameters. Results indicate that mixed forest has larger ET than farmland for most rainfall events. Compared to the farmland scenario, mixed forest has higher soil moisture in top layer; but mixed forest has a lower soil moisture in deep layer because of more transpiration. Surface runoff with mixed forest decreases significantly compared with the simulated values with farmland, and groundwater is reduced also because there is less drainage and more transpiration in mixed forest. Accordingly, streamflow at Bengbu station decreases by approximately 11% with the potential mixed forest. This study indicates that forestation is effective at reducing peaks of floods.

In the semiarid loess region of northwest China, the use of gravel and sand as a surface-mulching material has been an indigenous farming technique for crop production for over 300 years. A field and modeling study was carried out to quantify the effect of a gravel-sand mulch material on soil water and soil temperature in the semiarid loess region of northwest China. The field experiment (April 4, 2001–July 12, 2001) consisted of two treatments: (1) no surface mulching as a control, and (2) gravel-sand surface-mulching. The numerical model HYDRUS-1D was used in the research reported in this paper. The results showed that compared with the nonmulched condition the gravel-sand mulched field had a more favorable soil-water and temperature environment for plant growth. In the initial stage of watermelon growth, there was larger soil-water storage (3.8 cm versus 1.9 cm) and soil temperature (10.8°C versus 6.2°C) in the 0–20 cm soil layer with the gravel-sand mulch. The mulch provided very beneficial water and temperature conditions for the germination of watermelon. The gravel-sand mulch improved soil-water conditions because it was effective in reducing evaporation and enhancing transpiration. The gravel-sand mulch also improved infiltration of rainwater. For the 0–20 cm soil layers the average temperature in the gravel-sand mulch field was 1.0–5.3°C larger than that in the no-mulch field. In the 20–40 cm soil layer, the average temperature in the gravel-sand mulch was 0.3–3.7°C larger than that in the no-mulch field. In the early stages of plant growth, plant roots primarily centralized in the 0–40 cm soil layer, and the temperatures under the mulch were conducive to plant growth and development. The field and model results consistently showed the value of a gravel-sand mulch to provide warm, moist conditions for watermelon production in the semiarid loess region of northwest China.

The impacts of human-induced land-use and climate changes on hydrologic processes have become a great challenge and attracted widespread attention of many researchers. Dramatic changes in land use and climate have occurred in the mountainous Qingyi River watershed in southwest China in the last three decades. Variable infiltration capacity (VIC), a large-scale hydrologic model, was used in this study to assess the impacts of land-use and climate changes on surface runoff, base flow, streamflow, and evapotranspiration (ET) of this watershed. The analysis for this study includes (1) investigation of change in historical land-use patterns, (2) detection of climate change (precipitation and mean temperature), and (3) simulation and assessment of hydrologic responses to these changes. The Mann-Kendall test was used to identify the long-term monotonic trends in precipitation and temperature for the period of 1980–2005. The results suggest no significant change in annual precipitation and a significant increase in annual temperature, particularly in February, April, July, and September. The analysis of three land-use maps reveals that the conversions between forest and shrubland/grassland were the predominant land-use change over the past three decades. Hydrologic simulations show that the influence of climate change on hydrologic processes was stronger than those of land-use change. Monthly variation of the river flow was mainly attributed to seasonal variation in precipitation. However, ET responded significantly to the land-use change in several subwatersheds where single land cover conversion occurred dominantly. The decrease in surface runoff and base flow caused by climate change was enhanced by changes in land use, whereas the reduction in ET was offset by reforestation over the study period. Furthermore, the impact of deforestation or reforestation on hydrologic processes was more significant in the dry season than other seasons. The results from this study can be a useful reference for decision making in land-use planning and water resource managements in this region.

The Xinanjiang model parameter calibration is an optimization problem that aims to determine the values of model parameters that provide the best fit between observed and estimated flows. Many researchers have used the shuffled complex evolution (SCE-UA) algorithm in the Xinanjiang model parameter calibration and have found some problems in its application. Identifying methods in choosing an objective function, the amount of calibration data to be used, and the best calibration steps in searching for the best parameters are just some of the application problems that must be solved. The goal of the present study is to resolve the aforementioned problems. This work calibrates the Xinanjiang model for the Yuetan basin in China based on the SCE-UA method. In the objective function study, six objective functions are commonly used. Results demonstrate that the water quantity balance error function is the best choice in both daily and hourly Xinanjiang model calibrations. During the experiment, 20 sets of observed data with length differences of 1–20 years are used. The results reveal that data amounting to a period of 16 years are needed to obtain a relatively stable parameter set. Meanwhile, the objective optimization method is introduced in the calibration step study. This method combines the coupling of the objective optimization method with the SCE-UA algorithm. Parameter calibration results reveal that the combined method can increase the calibration speed and reduce the influence of the calibration data period on the results.

Water scarcity in Huaihe River, the sixth largest river in China, is stressed by the changing climate and intense human activities, especially in the headwater area. Xixian, situated in the upper reaches of the Huai River, is an important agricultural county with a population of more than one million people, and nearly one billion kg of crop yield every year. The projected climate changes and increasing population are expected to further complicate the utilization of already stressed water resources, endangering the agricultural activities in this area. There is pressing need for a watershed model to better understand the interaction between land use activities and hydrologic processes and to support sustainable water use. This study evaluated the performance of the Soil and Water Assessment Tool (SWAT) for hydrologic modeling in the Xixian basin; three methods of calibration and uncertainty analysis (sequential uncertainty fitting, generalized likelihood uncertainty estimation, and parameter solution) were compared and used to set up the model. The results showed that SWAT performs well in the Xixian River basin, that the hydrological water balance analysis of the basin indicated that base flow is an important aspect of the total discharge within the study area, and that more than 60% of the annual precipitation is lost through evapotranspiration. The calibrated model can be used to further analyze the effects of climate and land use changes and to investigate the effects of different cultivation styles and management scenarios on local water resources.

Evapotranspiration (ET) is an important, but unmeasurable, component of the hydrological cycle in semiarid regions. Traditionally, actual ET is computed as residual in water balance equations. It is derived from estimates of potential ET or, indirectly, from field measurements at meteorological stations. Recently, researchers have begun using scintillometers, remote sensing data, and hydrological models to estimate areal actual ET. In this study, the surface energy balance algorithm for land (SEBAL) was used to derive ET maps from moderate resolution imaging spectroradiometer (MODIS) images over the Laohahe basin and Shalamulun River basin. The effect of ground parameters on ET of the study area was quantified using the spatial analysis techniques of ArcGIS. At the end, ET estimated from Landsat 7 enhanced thematic mapper plus (ETM+) was compared with that from MODIS data over the Shalamulun River basin. SEBAL is a suitable algorithm for mapping evaporation over semiarid areas, using MODIS and Landsat images with few or no ground measurements. The ET of the study changes from 0 to $6.57\mathrm{mm}/\mathrm{day}$. The land use types, elevation, land surface temperature (LST), and terrain all have a direct effect on the spatial distribution of ET. ET simulated from both MODIS and Landsat data give reasonable values; however, results from Landsat ETM+ are better compared to those of MODIS because Landsat data have higher spatial resolutions than MODIS data.

The Laohahe River basin, located in northeastern China, was selected as a case study to quantify the magnitude of changes in land use and land cover (LULC) during the period from the 1970s to the 2010s and its quantitative effects on surface hydrology, based on hydrologic modeling of a distributed Soil and Water Assessment Tool (SWAT) model and catchment-scale spatial information analyses from remotely sensed data. Land cover maps with 30-m resolution from 1979, 1989, 1999, and 2007, interpreted from Landsat images, were used to analyze LULC changes during the last decades. The observed daily hydro-meteorological data from 1970 to 2006 were divided into four periods: 1970–1979, 1980–1989, 1990–1999, and 2000–2006. The SWAT model was utilized for each period with four LULC scenarios, which were developed by using the four LULC maps. Annual and monthly surface runoff and actual evapotranspiration (AET) were selected as important hydrologic elements to indicate the hydrologic response to LULC changes. The results revealed that distinct land cover changes occurred in the basin; the most important change was the conversion among vegetation cover classes of cropland, forest land, and grassland. Surface runoff always decreased as the LULC scenarios changed from 1976 to 2007 during all periods, but AET did not change regularly following the present LULC changes. Multiple regression equations between quantitative changes of LULC and hydrologic elements were developed. The equations indicated that the changes in three vegetation cover classes of grassland, cropland, and forest areas significantly affected hydrological elements and the increases of vegetation cover class areas all led to decreases in surface runoff and increases in AET. Moreover, given the same quantitative area change, the effects of cropland on hydrologic elements were the strongest, the effects of forest land were the second strongest, and the effects of grassland were the third. The effects of LULC changes on the seasonal distribution of hydrologic elements were also investigated. The results demonstrated that LULC changes have less influence on surface runoff and AET in nonflood seasons, but more influence in flood seasons, especially in July, August, and September, when the crops grow best. The results of this study improved the understanding of hydrologic responses to LULC changes and provided needed knowledge for agricultural decisions and the management of land use and integrated water resources.

Precipitation recycling plays a key role in the hydrological cycle process. The precipitation recycling ratio is a diagnostic measure for interactions between land surface hydrology and regional climate. To demonstrate the influence of the hydrological cycle on the climate of an arid area in northwest China, the authors studied the precipitation recycling ratio and the characteristics of the hydrological cycle in the Tarim River basin, which includes characteristics of water vapor transportation flux divergence and convergence, evaporation and its spatial distribution, and the spatial distribution of precipitation recycling ratio. The results suggest that there is about 14% of annual precipitation coming from the evaporation of the Tarim River basin, and 86% of advected precipitation coming from the surrounding external areas to the region. It is lower than other world famous river basins such as the Amazon, the Mississippi River, the Yellow River, and the Yangtze River basins. The recycling is active over the region during the months from April to July, and extremely inactive during the months from October to next January. The recycling ratio is relatively high in the northwest and north area where human beings live and open up large tracts of wasteland for cultivation and then change the land surface, which acts as a significant source of moisture to the atmosphere during certain seasons. Analysis of the recycling ratios spatial distribution suggests that the basin’s unique topography and land surface use change (through evapotranspiration) have influenced the precipitation and its recycling over the basin.

Groundwater inflow during tunnel excavation is a common problem in practice. How to accurately predict its occurrence during the construction is still a challenging problem for tunnel designers. A numerical method, based on the coupled model involved in artery fractures described by a discrete fractured network model and ramification fractures and rock matrix described by the equivalent continuum medium model, is developed to calculate the groundwater inflow of underground tunnel. The model is calibrated with the observed groundwater levels in the study domain. The results in the model calibration show that calculated and measured groundwater inflows agree well. Sensitivity analysis indicates that groundwater inflow increases with the increase of precipitation rate, hydraulic conductivity of rock matrix and fracture aperture. The effect of fracture aperture on groundwater inflow is predominant, owing to the occurring of many artery fractures around the tunnel, which leads to much groundwater flowing to the tunnel through these fractures.