Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Dawei Zhang has a total of 48 co-authors. Here are the top co-authors by number of shared publications:

  • Liyun Xiang

    State Key Laboratory of Water Cycle and Water Security, China Institute of Water Resources and Hydropower Research, Beijing, China

    3 shared publications
  • Zhenchun Hao

    National Key Laboratory of Water Disaster Prevention Hohai University Nanjing China

    3 shared publications
  • Zhongxiang Wang

    China Institute of Water Resources and Hydropower Research, Research Center on Flood & Drought Disaster Reduction of the Ministry of Water Resources, Beijing, P. R. China

    3 shared publications
  • Chang Chen

    State Key Laboratory of Water Cycle and Water Security, China Institute of Water Resources and Hydropower Research, Beijing, China

    2 shared publications
  • Cheng Zhang

    State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research

    2 shared publications
  • Dingzhi Peng

    College of Water Sciences, Beijing Normal University, Beijing 100875, China

    2 shared publications

Latest Publications

  • Open Access

    Deep Learning Model for Real‐Time Flood Forecasting in Fast‐Flowing Watershed

    by Fan Wang, Jie Mu, Cheng Zhang, Weiqi Wang, Wuxia Bi, Wenqing Lin, Dawei Zhang

    The fast‐flowing watershed is characterized by rapid runoff and confluence, posing challenges for accurate flood prediction. We introduce three flood forecasting model structures, namely GRU‐ED, LSTM‐FED, and LSTM‐DSA to address this issue. Through application research in three representative watersheds, we found that: First, as input information attenuates, the predictive ability of the models may decline with an extended lead time. The incorporation of a feedback mechanism effectively addresses this issue, resulting in an average 5% improvement in Nash efficiency and a significant 26.4% reduction in the interquartile range of relative peak error. Second, the performance of the model is influenced by various factors, including the watershed characteristics, sample size, and temporal resolution. Further investigation is required to determine the extent of their influence. The attention mechanism dynamically assigns weights to input data, significantly improving model performance, especially for larger catchments. This leads to an average increase in Nash efficiency of approximately 7.86% and a reduction in the interquartile range of relative peak error by about 30.7%. Finally, the proposed models demonstrate a high level of accuracy in flood forecasting within a specific lead time, offering an innovative deep learning‐based solution to the problem of fast‐flowing watershed flood forecasting.


  • Open Access

    GKASA–DDPM: a novel flood forecasting model based on Graph Kolmogorov–Arnold Attention and spatio-temporal attention under smoothing DDPM

    by Chang Chen, Dawei Zhang, Xin Qi, Zhongxiang Wang, Liyun Xiang

    Flood forecasting plays an important role in water resources management and flood prevention, leading to enormous research interests. Given the considerable data amount and computational complexity, many scholars have considered this problem from the perspective of time-series forecasting and made substantial contributions in a data-driven manner. Based on this, a novel method is proposed to handle flood forecasting in this study, which is a modified denoising diffusion probabilistic model (DDPM). In addition, a novel attention mechanism, Graph Kolmogorov–Arnold Attention (GKAT) is designed. Since the Kolmogorov–Arnold Network (KAN) utilizes tunable activation functions, it increases the interpretability of complex hydrological models. Meanwhile, spatio-temporal attention (SA) is also adopted, considering the time-variant characteristic of the time-series data. Therefore, this method is termed GKASA–DDPM. Furthermore, Savitzky-Golay smoothing mechanism is deployed in the post-processing procedure to adjust the predicted results according to practical observation. Multiple experiments are executed to exhibit the superior performance of the proposed method, involving seven models and observed hydrological data collected from Xiaoqing River basin above Huangtaiqiao Hydrological Station. Comprehensive results show that GKASA–DDPM achieves the highest prediction accuracy under all experimental conditions, over 0.9 when evaluated by Nash–Sutcliffe Efficiency (NSE).


  • Open Access

    Experimental Study of Dike‐Break Induced Flow Generated by Instantaneous Opening of the Side Gate

    by Dawei Zhang, Zhongxiang Wang, Huiwen Liu, Wuxia Bi

    A large‐scale experimental model of instantaneous dike‐break induced flow was conducted in this work. Water level variations in the river channel and floodplain, breach discharge, and the surface velocity field at the breach were measured during dike failure. The results show that: (i) The water level in the river rapidly decreased to a minimum (15%–22% of the initial water depth), then began to gradually rise, and finally approached stable. The water level in the floodplain gradually increased and ultimately tended towards stability. (ii) The breach discharge initially increased to a peak, then gradually decreased with a decreasing rate. The peak discharge was not only related to the initial river water level before dike‐break, but also to the river velocity. Under the same conditions, the higher the river water level or the higher the river velocity, the greater the flood peak at the breach. And (iii) During the process of dike‐break, the surface velocity of the breach flow gradually decreased. Other things being equal, a higher river water depth or a higher river velocity led to a larger surface velocity of the breach flow. These findings help better understand the hydrodynamic process and provide data support for models.


  • Top Cited Papers
    Open Access

    Quantifying Dynamic Linkages Between Precipitation, Groundwater Recharge, and Streamflow Using Ensemble Rainfall‐Runoff Analysis

    by Huibin Gao, Qin Ju, Dawei Zhang, Zhenlong Wang, Zhenchun Hao, James W. Kirchner

    Understanding streamflow generation at the catchment scale requires quantifying how different components of the system are linked, and how they respond to meteorological forcing. Here we present a proof‐of‐concept study characterizing and quantifying dynamic linkages between precipitation, groundwater recharge, and streamflow using a data‐driven nonlinear deconvolution and demixing approach, Ensemble Rainfall‐Runoff Analysis (ERRA). Streamflow in our mesoscale, intensively farmed test catchment is flashy, but occurs at time lags that are too long to be plausibly attributed to overland flow. Instead, ERRA's estimates of the impulse responses of groundwater recharge to precipitation, and of streamflow to groundwater recharge, imply that this intermittent streamflow is primarily driven by precipitation infiltrating to recharge groundwater, followed by discharge of groundwater to streamflow. ERRA reveals that streamflow increases nonlinearly with increasing precipitation intensity or groundwater recharge, and exhibits almost no response to precipitation or recharge rates of less than 10 mm d −1 . Groundwater recharge is both nonlinear, increasing more‐than‐proportionally with precipitation intensity, and nonstationary, increasing with antecedent wetness. Simulations with the infiltration model Hydrus‐1D can reproduce the observed water table time series reasonably well (NSE = 0.70). However, ERRA shows that the model's impulse response is inconsistent with the real‐world impulse response estimated from measured precipitation and groundwater recharge, illustrating that conventional goodness‐of‐fit statistics can be weak tests of model realism. Thus, our proof‐of‐concept study demonstrates how impulse responses estimated by ERRA can help clarify linkages between precipitation and streamflow at the catchment scale, quantify nonlinearity and nonstationarity in hydrologic processes, and critically evaluate simulation models.


  • Open Access

    An Integrated Model for Dam Break Flood Including Reservoir Area, Breach Evolution, and Downstream Flood Propagation

    by Huiwen Liu, Zhongxiang Wang, Dawei Zhang, Liyun Xiang

    The reasonable and efficient prediction of dam failure events is of great significance to the emergency rescue operations and the reduction in dam failure losses. This work presents a model that is based on the physical mechanism. It is coupled with a multi-architecture (multi-CPU and GPU) open-source two-dimensional flood model, which is based on high-precision terrain and land use data. The aim is to enhance the accuracy of dam break flood process simulations. The model uses DEM data as a computational grid and updates it at each time step to reflect breach evolution. Simultaneously, the breach evolution model incorporates an analysis of stress on sediment particles, establishing the initial erosion state and lateral expansion model while accounting for seepage. The determination of the overflow of the breach is resolved through the application of a two-dimensional hydrodynamic model. This approach achieves a robust connection between the upstream reservoir, the dam structure, and the downstream inundation area. The coupled model is utilized to calculate the failure of earth-rock dams and landslide dams, and a sensitivity analysis is conducted. Taum Sauk Dam and Tangjiashan landslide dam were selected to represent earth dam break and barrier lake break, respectively, which are the main types of dam breaks. The obtained results demonstrate strong concurrence with the measured data, the relative errors of the four important parameters of the application case, the peak discharge of the breach, the top width of the final breach, the depth of the breach and the arrival time of the maximum peak discharge are all within ±10%. Although the relative error of the completion time of the final breach is greater than 10%, it is about 30% less than the relative error of the physical model.


  • Open Access

    Quantifying Dynamic Linkages Between Precipitation, Groundwater Recharge, and Runoff using Ensemble Rainfall-Runoff Analysis

    by Huibin Gao, Qin Ju, Dawei Zhang, Zhenlong Wang, Zhenchun Hao, James W. Kirchner

    Understanding streamflow generation at the catchment scale requires quantifying how different components of the system are linked, and how they respond to meteorological forcing. Here we use a data-driven nonlinear deconvolution and demixing approach, Ensemble Rainfall-Runoff Analysis (ERRA), to characterize and quantify dynamic linkages between precipitation, groundwater recharge, and streamflow in a mesoscale intensively farmed catchment. Streamflow in this catchment is flashy, but occurs at time lags that are too long to be plausibly attributed to overland flow runoff. Instead, the impulse responses of groundwater recharge to precipitation, and of streamflow to groundwater recharge, imply that this intermittent runoff is primarily driven by precipitation infiltrating to recharge groundwater, followed by linear-reservoir discharge of groundwater to streamflow. Streamflow increases nonlinearly with increasing precipitation intensity or groundwater recharge, and exhibits almost no runoff response to precipitation or recharge rates of less than 10 mm d−1. Groundwater recharge is both nonlinear, increasing more-than-proportionally with precipitation intensity, and nonstationary, increasing with antecedent wetness. Simulations with the infiltration model Hydrus-1D reproduce the observed water table time series reasonably well (NSE=0.70). However, the model’s impulse response is inconsistent with the observed impulse response estimated from measured precipitation and groundwater recharge, illustrating that goodness-of-fit statistics can be weak tests of model realism. Our analysis demonstrates how impulse responses estimated by ERRA can help quantify nonlinearity and nonstationarity in hydrologic processes, and can help clarify the mechanistic linkages between precipitation and streamflow at the catchment scale.