To overcome the difficulty that existing hydrological models cannot accurately simulate hydrological processes with limited information in irrigated paddy areas in southern China, this paper presents a prediction model combining the Ensemble Empirical Mode Decomposition (EEMD) method and the Long Short-Term Memory (LSTM) network. Meteorological factors were set as the multivariate input to the model. Rainfall, regarded as the main variable affecting runoff, was decomposed and reconstructed into a combination of new series with stronger regularity by using the EEMD and K-means algorithm. The LSTM was used to explore the data laws and then to simulate and predict the runoff of the irrigated paddy areas. The Yangshudang (YSD) watershed of the Zhanghe Irrigation System (ZIS) in Hubei Province, China was taken as the study area. Compared with the other models, the results show that the EEMD-LSTM multivariate model had better simulation performance, with an $NSE$ above 0.85. Among them, the $R^2$, $NSE$, $RMSE$ and $RAE$ of the EEMD-LSTM${}^{\left(3\right)}$ model were the best, and they were 0.85, 0.86, 1.106 and 0.35, respectively. The prediction accuracy of peak flows was better than other models, as well as the performance of runoff prediction in rainfall and nonrainfall events, while improving the $NSE$ by 0.05, 0.24 and 0.24, respectively, compared with the EEMD-LSTM${}^{\left(1\right)}$ model. Overall, the EEMD-LSTM multivariations model is suited for simulating and predicting the daily-scale rainfall–runoff process of irrigated paddy areas in southern China. It can provide technical support and help decision making for efficient utilization and management of water resources.

Polydisperse particles are ubiquitous in both the natural and engineered environment, and the precise prediction of the transport and capture of polydisperse particles in a saturated medium is crucial. Several efforts (Yao model, RT model, TE model, MPFJ model, NG model, MHJ model, and MMS model) were developed to obtain accurate correlation equations for the particle capture probability (single-collector removal efficiency), but the applicability of the existing models to the entire porous medium and the retention characteristic of the polydisperse particles are still unclear. In this study, sand column experiments were undertaken to investigate the transport and capture processes of the polydisperse particles in the saturated medium. The mass density was employed to quantize the effects of particle polydispersity and incorporated into the depth-dependent deposition rate. The experimental results showed that the polydisperse particles formed a hyper-exponential retention profile even under favorable conditions (no repulsion). The excellent agreement between the results obtained from the MMS model and the experimentally observed results of the breakthrough curves (BTCs), as well as the retention profiles demonstrated the validation of the MMS model, as the correlation coefficient and the standard average relative error were 0.99 and 0.005, respectively. The hyper-exponential retention profile is caused by the uneven capture of the polydisperse particles by the porous medium. This study highlights the influences of particle polydispersity on particle transport and capture in a saturated porous medium.

Artificial groundwater recharge utilizing stormwater is an effective tool to reduce urban flooding and artificially increase groundwater resources. However, porous media clogging due to the solid polydisperse particles carried in stormwater severely restricts the application of recharge technology. In addition, solid particles play a crucial role in aquifer contamination. If particles are easily transport in groundwater flows, they can act as contaminant carriers to facilitate the movement of contaminants. Conversely, if particles cause porous media clogging, they can form barriers that prevent contaminant migration. While the transport and clogging mechanism of particles has been explored by many macroscopic physical experiments, the intrinsic connection between the deposition behavior of particles in pore scale and the macroscopic presentation of clogging phenomenon remains unclear. In this study, laboratory-scale sand column experiments were combined with scanning electron microscopy (SEM) observations to explore the effect of polydisperse particle size on the mechanism of particle transport and clogging. The median particle sizes (dp50) of the polydisperse particles used in the experiments were 0.66, 4.05, and 11.83 μm, respectively. The interaction energy between particles and porous media grains was calculated using XDLVO theory, indicating that the experiments were conducted under unfavorable condition. The sand column experiments revealed that the influence of large particles on the porous media permeability is limited to shallow layers, and small particles are more likely to transport to deeper layers. Particles with dp50=0.66 μm were tend to form aggregates, reducing particle recovery rate and promoting clogging. The pore-scale observations illustrated that the vast majority of the particles are preferentially deposited in concave regions of the media grain surface. The larger the particle size, the higher the proportion of deposition in concave regions. Due to the different transport mechanisms of particles in pore space, the ratio of dp50 to porous media grain size is not the only basis for identifying the type of clogging. The results proved that the particles with dp50=0.66 μm can form mixed clogging faster than particles with dp50=4.05 μm. The proposed results provide a reference for the theoretical study of the particle transport and deposition mechanism, the prevention and control of aquifer pollution, as well as the development of more effective artificial groundwater recharge scheme.