This study takes the drainage system of Zhengpu Port New District in Ma'anshan as a representative case to explore the application of urban drainage system design based on the combination of storage algorithms and SWMM models. Established in 2012， Zhengpu Port New District is positioned as a modern port logistics area， a modern port industrial area， and an integrated urban area， with an important strategic development function. Firstly， the study collects geographical， climatic， and drainage system data of the new district and applies storage algorithms based on water balance to analyze the drainage system and its scale. Then， the optimized plan is applied to the SWMM model for dynamic simulation and analysis of rainwater runoff processes. The research shows that to achieve the flood control design standard of no water accumulation in a 20-year rainstorm， the study area needs an external drainage pump station with a total design flow rate of 111.9 m³/s. After deducting the total flow rate of the existing three pump stations， an additional 30.5 m³/s of external drainage flow rate is required. The results of this study provide a scientific basis for drainage and disaster reduction and drainage system renovation in the northern part of Laoxia River in Zhengpu Port New District， contributing to the development and people's livelihood of the new district. Meanwhile， this research method can serve as a reference for urban drainage system design application studies based on the combination of storage algorithms and SWMM models， promoting the development of urban drainage system design.

In order to quantitatively evaluate the effectiveness of sponge city construction in the Lihu area of Guangzhou， this study constructed a SWMM model at the scale of drainage zoning through background analysis and data collection. The simulation analysis of the sponge city construction effect in the Lihu area of Guangzhou was carried out from the aspects of annual runoff total control rate， pipe network overflow capacity， and waterlogging risk， in order to verify the disaster prevention and mitigation ability of the area after development and construction according to the sponge city construction implementation plan. At the same time， the results of this study can provide theoretical reference for sponge city construction in high-density areas in the south to alleviate urban waterlogging. The main conclusions drawn from simulation calculations of the Lihu area in Guangzhou are as follows： ① After low impact development and construction， the annual runoff total control rate in the Lihu area of Guangzhou has significantly increased. In 2022， the annual runoff total control rate in the area was 81.46%，which was 15.37% higher than that before the low impact development and construction in 2019； ② The simulation analysis of the drainage network overflow capacity before and after the low impact development and construction in Lihu area of Guangzhou city shows that the area experienced overflow under different return periods of rainfall conditions before the low impact development and construction， and the overflow ratio of the network increased with the increase of the design return period. After the low impact development and construction of the area， the overflow situation of the pipeline network was significantly reduced under different return periods of rainfall conditions； ③ The simulation analysis of waterlogging risk after the low impact development and construction in Lihu District of Guangzhou City shows that the district is basically a risk-free area when the 24-hour rainfall with a 100 year return period （the accumulated rainfall is 322mm） occurs， and very few places are low-risk areas. The sponge city construction in Lihu District can effectively cope with the 100 year return period rainstorm. The construction of sponge cities in the Lihu area of Guangzhou can effectively reduce rainwater discharge and enhance the disaster prevention and mitigation capabilities of the area.

In order to further improve the efficiency of resistance-type turbines and reduce their influence on the liquid conveying capacity of pipelines， the influence of impeller design parameters on their performance was analyzed by single factor test method， and four key parameters and their value ranges were selected. Then， the response surface method was used to design the test scheme， and the function relationship between the key design parameters and the efficiency and head loss was fitted by the least square method， and the regression model is obtained. Finally， MOPSO algorithm was used to optimize the regression model， and the optimal impeller parameters was obtained. The results show that the average efficiency of the optimized resistance-type turbine is increased by 4.053%， and the average head loss is reduced by 0.679%.

Dramatical channel erosion has occurred in the braided reach of the Lower Yellow River （LYR） since the Xiaolangdi Reservoir operation. Study on characteristics and influencing factors of river bed incision has great significance for the river embankment and flood control in the braided reach of the LYR. The accumulated downcutting depths of river bed and channel bankfull depths were calculated based on the observed post-flood profiles at 28 cross-sections in the braided reach from 1999 to 2020. The annual water level-discharge curves of each hydrological station were obtained based on the hydrological data， and the annual change of water level at certain discharges was calculated. The results show that： ① The cumulative scour volume of the LYR was 2.05 billion m³， and the braided reach experiencing the most channel erosion with a cumulative scour volume of approximately 1.47 billion m³ from 1999 to 2020. ② The braided reach of the LYR experienced the most severe river bed downcutting between 2000 and 2007， with an average annual downcutting depth of 0.24 m/a. From 2008 to 2020， the degree of river bed downcutting in the braided reach decreased， with the average annual downcutting depth reducing to 0.05 m/a. ③ There have been significant changes in the water level-discharge relationships at three hydrological stations， with a noticeable decrease in water levels at given discharges at the three hydrological stations. There is a certain time lag in the decline of water levels at hydrological stations along the braided reach， with a lag time of approximately three to four years. The magnitude of water level decline at each hydrological station under a discharge of 500 m³/s is greater than that under a discharge of 3 000 m³/s， with the largest decline observed at Gaocun station. ④ The bankfull depth in the braided reach increased from 1.6 m to 4.0 m from 1999 to 2018， and then decreased to 3.8 m in 2020. ⑤ Influencing factors of river bed cumulative incision were investigated quantitively， including incoming flow and sediment regime and channel boundary conditions. It is found that the previous 6-year average incoming sediment coefficient is the primary factor influencing the accumulated downcutting depth of river bed in the braided reach of the LYR. In addition， the inhibition of continuous river bed coarsening on river bed incision were quantitively analyzed in the braided reach.

Long short-term memory network （LSTM） model can effectively simulate the nonlinear response between rainfall and runoff， and has been widely used in flood simulation and forecast. In order to improve the applicability and simulation accuracy of the model in different application scenarios， this paper， based on the LSTM model and its five variants， carried out a case study on the rainfall and runoff time series of 30 floods from 1986 to 2000 in the Shujia watershed in the southern mountainous area of Anhui Province. The flood simulation effects of LSTM and its variant models under different loss functions， different forecast periods and different training scales are discussed. The ensemble simulation of LSTM model， its variant model and extreme Gradient Rise （XGBoost） model is also studied. The results show that： ① The four loss functions can well simulate the flood process of the outlet section of Shujia Basin， and the simulation accuracy is as follows： relative root mean square error （RSR）> Nash efficiency coefficient （NSE）> mean square error （MSE）> Klin-Gupta efficiency coefficient （KGE）. The Nash efficiency coefficient （NSE） of each RSR test set under LSTM and its variant model can reach more than 0.7. ② With the extension of the prediction period， the model faces problems such as information forgetting or error accumulation when dealing with long time series， and the accuracy of flood simulation using LSTM and its variant models generally shows a downward trend； Under the same forecast scenario， with the increase of training scale， the simulation accuracy of the model first increases to the best and then tends to be stable. ③LSTM model and its variant model are combined with XGBoost model， which reduces the simulation deviation of a single model and makes the overall prediction more accurate and reliable； Moreover， the residual simulation is introduced to make up for the characteristics of compound flood which can not be captured by the single model， and the simulation accuracy of compound flood is further improved.

Long distance tunnels are commonly used to cross complex terrains. The combination of the Drilling and Blasting method （DB）， and the Tunnel Boring Machines （TBM） method for excavation in long tunnels on the plateau can fully leverage the flexibility of the DB method and the efficiency of the TBM method， but it also increases difficulty in schedule management. This paper considers the challenges of long tunnel construction on the plateau， such as long excavation durations， complex geological condition， low air pressure， and low temperature. Based on the studies on excavation simulation models of the DB method and TBM method， the paper examines the duration and impact of switching between the DB and TBM excavation method. It then establishes a simulation model for the excavation progress of long tunnels on the plateau using a combination of the DB method and TBM method， expressing it in the form of a duration function and control flow. Case studies demonstrate that this model can be applied to the situation of simultaneous multiple working-faces in both construction adit and main tunnel， and it can accommodate complex construction boundary conditions of the DB and TBM excavation. The research can provide support for excavation progress management in long-distance tunnels for water diversion， highways， railroads， and other projects in plateau regions.

Correlation and distribution model between compressive strength and permeability coefficient of cutoff wall are the key factors affecting the reliability analysis of cutoff wall strength and permeability failure of earth-rock dam. The testing data of high pressure jet grouting wall of 9 earth-rock dams are collected. AIC criterion is used to identify the best marginal distribution type of compressive strength， permeability coefficient and the best Copula function of correlation joint distribution model. The Bootstrap method was used to simulate the statistical uncertainty of the identification results. The results show that there is a significant negative correlation between compressive strength and permeability coefficient， which mainly obeys Weibull distribution. Frank Copula is identified as the best function used to construct the non-normal joint distribution between the two parameters. The identification difference of the best distribution type and best Copula function is mainly due to the different frequency distribution caused by the construction quality of cutoff wall. The research results can provide a simple and effective distribution mode for the reliability analysis of strength reliability of high-pressure Jet cutoff wall and permeability stability of earth-rock dam.

The combination of hydrological and meteorological forecasts can effectively improve the accuracy and forecast period of flood forecasting. The land-surface hydrological modeling module has become the focus of research by hydrometeorologists. As a new generation of distributed land-surface hydrological model， the WRF-Hydro model has broad application prospects in multi-scale flood forecasting. However， due to the complex parameterization scheme of each physical process and the large amount of calculation of the model， the current parameter sensitivity research of the WRF-Hydro model is insufficient， which also affects the simulation accuracy of the model. This study takes the Tunxi basin in the upper reaches of the Xin'anjiang River in the humid area as the research basin， constructs a set of single-objective and multi-objective functions， and combines Morris method for global sensitivity analysis to explore the parameter sensitivity of the WRF-Hydro model under different objective functions. The result shows that： soil parameters （DKSAT， SMCMAX， BEXP） mainly affect subsurface and surface runoff， and have a significant impact on runoff volume.In particular， DKSAT is the most sensitive and directly affects the movement speed of water in the soil. When increasing， the base flow rate increases significantly while the peak flow rate decreases significantly.； runoff parameters （SLOPE， REFKDT） mainly affect the distribution of surface runoff and baseflow， and have an important effect on the overall shape of the flood hydrograph； the channel routing parameter ManN affects the speed of runoff convergence to channel network and mainly controls the peak time； vegetation parameter MP has a certain influence on annual water volum； overland flow parameter OVROUGHRTFAC and groundwater parameter Zmax are the least sensitive.. There are certain differences in the parameter sensitivity order and optimal parameter values under different objective functions. The single objective function based on relative bias focuses on the simulation of the total annual runoff and the low flow part， while Nash-Sutcliffe and Kling-Gupta coefficients focus more on the simulation of floods and high flow parts. The multi-objective function comprehensively considers the influence of different objective functions， and the results are better than the single-objective function to a certain extent. This work can provide a reference for reasonably determining the WRF-Hydro model parameter optimization strategy.

Taking Jiexi County， Jieyang City， Guangdong Province as the research object， this paper used the time series Sentinel-1 data of the second half of 2021 to anylize the time series backscattering coefficient characteristics and inter-class differences of different covers on cultivated land such as rice， corn， pond water surface， uncultivated， woods and vegetables in the measured samples. The results showed that the optimal polarization mode of crop classification in Jiexi County was VH polarization， on which， a model for the classification of cultivated land crops based on long short-term memory network （LSTM） was constructed. The model accuracy reached 90%.The spatial distribution of rice， corn， pond water， uncultivated land， forests and vegetables in the study area was extracted based on the model， providing a new remote sensing technology for monitoring cultivated land types in cloudy areas.

Rivers carry water resources and are channels for material and energy transfer. The management of river-related construction projects is an important part of ensuring the flood control， drainage， and stable river conditions of rivers. The water resistance ratio is an important control indicator for river-related construction project management， and the study of its calculation method is of great significance for controlling the impact of river-related construction projects. The traditional water resistance ratio， also known as the structural water resistance ratio， refers to the proportion of the cross section area of the river occupied by a river-related construction project at a certain design flood level. The article reveals the inadequacy of the current structural water resistance ratio in reflecting the degree of obstruction of water flow dynamics propagation in river related construction projects. The structural water resistance ratio weakens the flood control impact of river-related construction projects near the main channel of the river， and over-reflects the flood control impact of river-related construction projects near the shallow shoals of the river. On the basis of the structural water resistance ratio， combined with the hydrological properties such as water depth and flow velocity of the river， the article proposes a hydraulic water resistance ratio. The hydraulic resistance ratio refers to the proportion of the flow of the cross section occupied by the river-related construction project to the total flow of the entire river section at a certain design flood level. Compared with the structural water resistance ratio， the hydraulic water resistance ratio more effectively reflects the degree of obstruction of water flow dynamics propagation in river related construction projects， and the calculation method is simple， which can compensate for the shortcomings of the structural water resistance ratio. This article suggests incorporating the hydraulic water resistance ratio into the control indicators of river related construction projects， effectively controlling the degree of impact of river related construction projects on river flood discharge.

In order to study the influence of guide vanes on the flow state and evolution law of the vortex structure of pump as turbine， the evolution process of vortex structure under the optimal working condition of pump as turbine is analyzed based on Omega method. Firstly， the turbine with guide vane number Z ₀=0 is taken as the base， and then guide vanes are added to it. The results show that when the number of guide vanes is Z ₀=0， the vortex structure is relatively disordered， especially the outlet vortex interaction makes its morphological structure disorderly. When the number of guide vanes is Z ₀=9， the outlet vortex range is small， and the vortex shape and evolution are clear. The vortex structure in the guide vanes is simple and the evolution process is very clear. When the number of guide vanes Z ₀=0， the vortex structure and direction are not obvious due to the interaction between the main vortex and small vortex in the tailwater pipe. The spiral main vortex in the tailwater pipe with guide vane number Z ₀=9 has the same rotation direction as the impeller， and the opposite rotation direction of the small vortex. The tangential and axial velocities of the three sections in the tailwater tube of the turbine with Z ₀=9 have been improved compared to Z ₀=0， and the maximum tangential kinetic energy of the P1 section in the turbine with Z ₀=9 is 96.40% higher than that of Z ₀=0. In summary， guide vanes can reduce the generation of vortices， increase the flow velocity of the turbine and reduce the hydraulic loss. The research results can provide a theoretical basis for the design and operation of the guide vane turbine.

Reanalysis datasets are attractive for hydrological modeling and reliable water resource management， especially for areas where meteorological information is scarce. This study takes the Yellow River source as the research area， uses the China Meteorological Assimilation Driving Dataset for the SWAT Model （CMADS） to drive the Soil and Water Assessment Tool （SWAT） model for daily scale runoff simulation， and uses the SWAT-CUP （SWAT Calibration and Uncertainty Program） and SUFI-2 （Sequential Uncertainty Fitting-2） algorithms for calibration and validation to evaluate the accuracy of CMADS and its applicability for hydrological simulation in the Yellow River source area. The results show that： ① The accuracy of CMADS for daily scale temperature in the Yellow River source area is very high， and the correlation coefficients with the measured data from eight meteorological stations in the basin are all above 0.95. The accuracy of daily precipitation during the flood season is satisfactory， with the relative error basically between ±10%， and the accuracy of daily precipitation during the non-flood season is poor， with the relative error basically between -30% and -50%. ②The applicability of SWAT model for hydrological simulation over the Yellow River source area is very strong. The NSE， R2， PBIAS， RSR and KGE evaluation indicators of the calibration period and the validation period obtained by driving the SWAT model using meteorological station observation data are all very good.. ③ Two methods are used to evaluate the applicability of CMADS hydrologic simulation. Method 1 is to use CMADS to drive the SWAT model to rate and validate and perform hydrological simulation； method 2 is to use CMADS to drive the SWAT model， which has already been rate-validated with the best parameters using measured meteorological data， to perform hydrological simulation. It was found that the hydrological simulation of CMADS in the source area of the Yellow River had a high correlation with the measured flow， but it was easy to underestimate the flow. Overall， the applicability of CMADS in the hydrological simulation of the source area of the Yellow River was good， in which Method I was better than Method II in the hydrological simulation. The results of this study proved that CMADS can provide a relatively reliable data source for the alpine mountainous areas where meteorological data are scarce， and provide a possibility to expand the temporal and spatial scales of hydrological simulation.

With the grid-connected operation of large-capacity hydropower units and large-scale new energy sources， the hydro-turbine speed governing system， as the main control equipment of peak regulation and frequency regulation in the power grid， has taken on more and more heavy tasks of regulation and control. Its health status is directly related to the safety and stability of hydropower stations and power generation efficiency. Based on the digital twin five-dimensional model， according to the operating principle and spatial structure of the governing system， a new state evaluation framework for the speed regulation system of hydraulic turbine is proposed in this paper. A digital twin model of hydraulic turbine speed regulation system is established， including a three-dimensional physical model， an Operation Mechanism model and a data-driven model.The digital twin model is used as a support to complete the construction of a state assessment system for the turbine speed regulation system for digital twins. It can provide reference for the digital twin modeling of hydropower unit.

To explore the changing patterns of landslide disasters in Hebei Province under future climate change， and provide scientific basis for the overall planning of disaster prevention and reduction， so that people can better avoid the hazards and risks brought by landslide disasters. This article uses the Ecological Niche Modeling Evaluation Algorithm （ENMeval） to optimize the Maximum Entropy Model （MaxEnt）. Based on 860 landslide disaster points collected from surveys and 11 factors affecting disasters， the article carries out hazard risk assessment on the basis of the main influencing factors， and combines three radiative forcing scenario models of future （recent： 2041-2060， mid-term： 2061-2080， long-term： 2081-2100） climate data （low level forcing scenario： SSP126， medium and high level forcing scenario： SSP370， high level forcing scenario： SSP585）， predict the spatial distribution pattern and change pattern of potential risk areas for landslide disasters. The results show that：①The Area Under the Curve （AUC） of the optimized MaxEnt model is above 0.9， indicating that the model performs well in predicting potential risk areas of landslide disasters.②By optimizing the comprehensive contribution rate analysis calculated by the Maxent model， the main factors affecting landslide disasters are determined to be： precipitation in the wettest quarter， precipitation in the wettest month， and seasonal variation coefficients of elevation and rainfall. This indicates that precipitation related factors are the most important factors affecting landslide occurrence.③A comprehensive evaluation of 14 Coupled Model Intercomparison Project Phase 6（CMIP6） data revealed that BCC-CSM2-MR has the best simulation ability in precipitation compared to other climate models， followed by CMCC-ESM2 and ACCESS-CM2.④In the future climate forcing scenario， the area of low-risk areas for landslide disasters has decreased， while the proportion of areas with medium， medium high， and high risk has increased， which is consistent with the predicted trend of future precipitation changes.

Hydraulic machinery and equipment play an important role in the current national production， and its safe and stable operation is very important. In order to solve the problem that it is difficult for a single depth feature to effectively reflect the fault information of the unit， a fault diagnosis model for hydraulic machinery and equipment based on the fusion of convolutional neural network and graph convolutional network features is proposed. Firstly， the convolutional neural network is used to obtain the convolution depth characteristics of the monitoring signal of hydraulic mechanical equipment， and the fast Fourier transform is used to obtain the spectral value of the monitoring signal， and the monitoring signal graph data is constructed， and the graph convolutional network is constructed to extract the sample association features. Then， the attention mechanism is used to perform weighted summation of different types of features to achieve multimodal feature fusion. Finally， the full connectivity layer is used to realize the fault diagnosis of the equipment. The model is validated by the measured data of the hydroelectric unit and main pump unit faults as well as the bearing fault data. The results show that the proposed model can effectively realize the fault diagnosis of hydromechanical equipment.

In order to improve the sediment removal efficiency of sediment removal tunnels in reservoir， we have developed a new type of swirl sediment removal device. In this paper， the different combinations of impellers in swirl sediment removal device， the velocity distribution of swirl flow field， and the influence of swirl flow device on the scouring and silting characteristics and the efficiency of sediment removal at the mouth of a sand drainage tunnel are studied by numerical simulation and hydraulic model test.The numerical simulation results show that the flow field range of the single impeller in the swirl sand removal device expands with the increase of the rotational speed， but the flow field range does not increase when the rotational speed of the impeller increases to 20 r/s.The range of swirl flow field after the combination of multiple impels in the swirl sand removal device depends on the rotational speed of the impels and their arrangement. When the swirl sand removal device is arranged in a triangular manner， the range of the combined swirl flow field also expands with the increase of the rotational speed of each impeller. However， too small arrangement distance between impeller groups will lead to too high energy dissipation， and too large arrangement distance between impeller groups can not form an effective combined swirl field. The optimal operating state of swirl sand removal device is that the distance between impeller groups is 70 mm and the rotating speed of impeller is 20 r/s.The results of numerical simulation and hydraulic model test show that the swirl sand removal device can effectively increase the sand removal efficiency of the tunnel and the range of strong scouring and silting. When the reservoir water level is deep and the silting form is a scouring funnel， the swirl sand removal device can increase the efficiency of the tunnel by 71.5%；When the reservoir water level is shallow and the form of sediment deposition is a sand flushing corridor， the swirl sediment removal device can increase the sediment discharge capacity of the tunnel by 4.02 kg， indicating that the device has a better removal effect on the sediment deposition in the reservoir with shallow water depth.It is of great significance to solve the problem of sediment deposition in Hexi reservoir with less water and more sediment in Gansu province.

Snowfall， as an integral part of the water cycle in the Tibetan Plateau， plays an irreplaceable role in maintaining the balance of regional water resources and ecological balance. However， with the intensification of global warming， the characteristics of snowfall change in the Tibetan Plateau region not only pose challenges to the stability of the regional climate but also raise new issues for water resource management and ecological environment protection. To deeply understand the spatiotemporal characteristics of snowfall changes in the Tibetan Plateau and their response to climate change， this study selected six representative basins， namely the Lhasa River， Nianchu River， Naqu River， the source of the Yangtze River， the source of the Yellow River， and the source of the Lancang River， as research objects. The aim is to reveal the spatiotemporal variation patterns of snowfall in the Tibetan Plateau during this period through multi-source remote sensing precipitation data fusion technology. This study selected three types of gridded precipitation data： CMFD， CPC， and TRMM， and constructed an evaluation index for precipitation data based on overall deviation and deviation components to assess the accuracy and optimize the integration of these data. The results show that from 2002 to 2018， the snowfall rate in the typical basins of the Tibetan Plateau generally showed a downward trend， especially in areas above an altitude of 5 000 meters， where the reduction in snowfall rate was more significant. The ranking of the decrease in snowfall rate from large to small is： Lhasa River （28.2%）， the source of the Yellow River （18.2%）， the source of the Yangtze River （17.8%）， Nianchu River （15.4%）， the source of the Lancang River （12.1%）， and Naqu River （10.5%）. This downward trend shows a spatial distribution feature that intensifies gradually from south to north. This finding is closely related to global warming and provides important scientific evidence for understanding the response of the Tibetan Plateau to climate change. In addition， this study has important reference value for the formulation of water resource management and climate change adaptation strategies in the Tibetan Plateau region， and also provides data support and research methods for related fields， which helps to promote the in-depth development of climate change research in the Tibetan Plateau.

In the context of the national strategy of carbon emission peak and carbon neutrality， implementing dynamic water level control during the flood season can enhance the utilization of flood resources and fully leverage the comprehensive benefits of water conservancy hubs while ensuring flood safety in the basin. The Goupitan， Silin， and Shatuo hydropower stations serve as the primary cascade power stations in the middle and lower reaches of the Wujiang River. Since the current flood control capacity relies on single-reservoir calculations of flood regulation， neglecting joint scheduling and operation， the operating water level during the flood season is statically controlled， thereby limiting power station output and failing to fully realize comprehensive flood season benefits. Based on forecasted pre-discharge and reservoir capacity compensation， this study coordinates flood control risks across multiple regions in the middle and lower reaches of the Wujiang River， constructed a dynamic control and dispatch model for the operating water levels of cascade reservoirs during the flood season and studied the dynamic strategies and risks of raising the operating water levels of the Silin and Shatuo reservoirs based on the impoundment of the Goupitan reservoir. The study findings indicate that implementing joint flood forecasting and dispatching for cascade reservoirs can increase the operating water levels of Silin and Shatuo reservoirs by 2.10~3.86 meters and 6.09~6.42 meters respectively in the pre-flood season （June-July）， and by 4.17~4.86 meters and 6.21~6.42 meters in the post-flood season （August）， respectively. This approach can also mitigate flood control risks for the cascade hubs， the middle and lower reaches of the Wujiang River （Sinan and Yanhe counties）， and the middle and lower reaches of the Yangtze River. The results of this study can provide technical guidance for the safe management of flood control and the utilization of flood resources during the flood season in the Goupitan， Silin， and Shatuo cascade reservoirs.

In order to solve the problem of information loss that may be brought by a single model in dam deformation prediction， differential evolutionary algorithm （DE） is used for the parameter optimisation of the long- and short-term memory neural network （LSTM） model and combined with the multiple linear regression （MLR） model to establish a tandem combination model of MLR-DE-LSTM. The model is validated based on the horizontal displacement prototype monitoring data of a gravity dam. The results show that the DE algorithm can effectively improve prediction accuracy of the LSTM model， and the LSTM model can effectively mine the information that has not been fully explained by the MLR model. Compared with a single model， the combined model has higher accuracy and stability in predicting displacement data， and the combined model has greater advantages in making full use of the data information. The results provide reference value for improving the accuracy of dam deformation prediction.

The drawdown water levels of cascade reservoirs is one of the important factor affecting hydropower generation. With the construction of the Lianghekou Reservoir with multi-year regulation capacity， determining the drawdown water levels of the Lianghekou Reservoir， and the downstream Jinping Ⅰ and Ertan hydropower stations to increase hydropower generation of cascade hydropower plants in the Yalong River needs further research. This study established a joint optimal operation model for the maximum hydropower generation of the cascade hydropower plants in the middle and lower reaches of the Yalong River. The dynamic programming- progressive optimality algorithm （DP-POA） was implemented to obtain the drawdown water levels， annual runoff and hydropower generation. The K-means clustering algorithm was employed to analyze the relationships between drawdown water levels， and annual runoff and hydropower generation under the medium and long-term optimal operation of cascade hydropower plants. The results showed that： ①Compared with the average power generation of cascade hydropower plants under conventional dispatching， joint optimal operation could improve hydropower generation by 9.108 billion kWh （the corresponding increase rate of 9.36%） with more significant increase in hydropower generation of the Jinping hydropower plant Ⅱ； ② For most inflow scenarios， the drawdown water levels of the Lianghekou， Jinping Ⅰ and Ertan hydropower plants under joint optimal operation are 2 800， 1 825 and 1 175 m， respectively， which could increase the hydropower generation by 4.043 billion kWh with corresponding increase rate of 4.16%， as compared to the conventional operation； ③ There is a positive correlation between the hydropower generation and the annual runoff and a negative correlation between the drawdown water level and the inflow runoff was found in the Lianghekou hydropower plant. While due to the influence of the runoff regulation of the upstream hydropower plants， the correlation between the drawdown water level and the runoff of the downstream Ertan hydropower plant was not strong. The research results can provide references for the joint operation and drawdown water level controlling of the cascade hydropower plants in the middle and lower reaches of the Yalong River.

In earth-rock dam projects， due to the uncertainty of on-site construction compaction and the non-uniformity of material collection in the material field， the mechanical parameters of dam materials have variability in dam body， which leads to the differences between the conventional prediction analysis results of earth-rock dam using deterministic parameters and the actual stress and deformation behaviors of dam body. In this paper， based on the geotechnical test results of dam materials in the built engineering， the spatial random field theory is introduced， and the 3D static analysis of the dam body is carried out by constructing stationary random fields. The influence of the mechanical parameters of Duncan-Chang E-B model on the settlements and horizontal displacements of dam body is obtained. The sensitivity of the mechanical parameters of Duncan-Chang E-B model is studied， and the sensitivity ranking conclusion is given. The results show that the parameters K_b， K and {\phi }_{\mathrm{0}} in Duncan-Chang E-B model have great influence on the deformation of dam body， among which the influence of parameter K_b is the most obvious， followed by the influence of parameters n and R_f， while the parameters Δφ and m are less sensitive. This research conclusion can provide reference for the rational deformation prediction of earth-rock dam and the study of dam material characteristics.

Sudden increase or depletion of dissolved oxygen in river water bodies can cause a series of environmental pollution， species diversity destruction and other problems， and accurate prediction of dissolved oxygen （DO） values in rivers is of great significance to the management of river water environment. In order to improve the interpretability of model input features and model accuracy， and to obtain the optimal prediction model of river DO values， this study utilized the data from water quality monitoring stations in the Yellow River Basin in Shanxi， and used the bidirectional long and short-term memory network （BiLSTM） as the basis， combined with the convolutional neural network model （CNN） and the Attention Mechanism to optimize the feature selection， and conducted feature optimization based on the Random Forest Model （RF）， established RF-CNN-BiLSTM-Attention （RF-CBA） model， and further utilized the bionic optimization algorithms such as Bloodsucking Leech Optimization Algorithm （BSLO）， Black-winged Kite Optimization Algorithm （BKA）， and White Shark Optimization Algorithm （WSO）. A total of five optimization models， BSLO-RF-CBA，， BKA-RF-CBA， and WSO-RF-CBA， were constructed and compared with the CNN-A， LSTM-A， BiLSTM-A， CBA， and RF-CBA models of deep learning， and analyzed to obtain the prediction results of the river dissolved oxygen with the Mean Absolute Error （MAE）， Root Mean Square Error （RMSE）， Mean Square Error （MSE）， Coefficient of Determination （R ²）， Global Performance Indicator （GPI）， and Relative Error （MAPE） to evaluate the different model accuracies， and the results showed that： ① The RF model can eliminate the effect of redundant features on the water quality prediction model and improve prediction accuracies by sorting and filtering the influence of the feature values affecting the DO in rivers. ② Using the bionic algorithm to optimize the number of neurons， learning rate， regularization coefficient and other parameters of the RF-CBA model， the model simulation accuracy was further improved， the overall time series characteristics of DO fluctuations were captured， and the model showed strong stability and generalization ability. ③ The BSLO-RF-CBA model had the highest simulation accuracy， outstanding ability to capture DO changes， and stronger ability to capture global dependence relationships and is recommended for river DO prediction modeling. The model has the ability to be extended to predict the concentration of dissolved oxygen and other pollutants in different rivers， providing technical support for early warning and systematic management of river water pollution.

Due to unscientific groundwater development activities， the groundwater level in some areas of China has declined seriously， triggering a series of geo-environmental problems， and there is an urgent need to formulate the dual control management measures of groundwater level-water quantity to realize the sustainable development and utilization of groundwater. However， a single water quantity control index cannot meet the scientific management demand of water withdrawal control， so this paper takes Xiangyang city plain area as an example， uses groundwater numerical simulation method and combines the simulation prediction of different scenarios to study the response relationship of groundwater level-water quantity， so as to determine the dual-control water withdrawal index. The results show that， the groundwater level in the study area shows a good correlation with the exploitation amount and rainfall. With the increase of exploitation intensity， the groundwater level shows a significant downward trend. The response relationship between groundwater level and exploitation amount can be used as an important basis for determining the control water withdrawal index. By controlling the precipitation conditions and the water level threshold of the over-exploitation area， the water level and water volume thresholds are calculated according to the water level-water volume response relationship. The permissible extraction volume threshold of shallow groundwater under the current exploitation conditions is 1.16×10⁸ m3， and the average water level threshold is 93.11 m. Considering the spatial heterogeneity of the hydrogeological conditions， the spatially differentiated partitioning of the dual-control indexes is carried out， which provides a reference basis for the scientific management of groundwater.

The coordinated development of urbanization development with water security and food security is an inevitable requirement for achieving high-quality development. Metropolitan areas are key drivers for optimizing resource allocation and promoting coordinated regional development. Taking Nanjing metropolitan area as an example， we deeply analyze the evolution pattern of water security， food security， and urban development from 1990 to 2020. We use the coupling coordination degree model to reveal the spatio-temporal characteristics of the coordinated development of water security- food security- urbanization development， and identify the relevant influencing factors through the gray correlation analysis method. The results show that： ① The water security level and urban development index showed a steady increase from 1990 to 2020， while the food security level first declined and then rebounded. Overall， both water security and food security performed well， and by 2020， all cities in the metropolitan area had reached a relatively safe level. Nanjing， as the central city of the metropolitan area， had an urban development index significantly higher than other cities. ② The coupling coordination degree of water security-food security-urbanization development continues to improve， gradually entering the stage of intermediate or good coordination from a state of dissonance in 1990. The development of water security in each city is particularly prominent and prioritized. Among them， Yangzhou has the highest level of coordinated development， while for cities like Xuancheng， Wuhu and Ma'anshan their water safety and food security are far better than the urban development index， and have great potential for development. ③ Expansion of built-up land area， urbanization rate， and increase in population density are the most important factors affecting the degree of coordination of the three couplings. In addition， factors such as green coverage of built-up areas， food prices， the proportion of tertiary industry and the use of fertilizer per unit area also play an important role. The study provides a reference for coordinating the relationship between water， food， and urban development， optimizing the allocation of regional resources， and promoting the high-quality development of the metropolitan area.

In order to solve the problem that the traditional water index easily identifies non-water bodies such as shadows， buildings and ice and snow as water bodies when extracting water bodies， a new multi-band combined water index model （NMCWI） of blue light band （490 nm）， green light band （560 nm）， red light band （665 nm）， vegetation red edge band （783 nm） and near infrared band （842 nm） is proposed by analyzing different spectral characteristics of Sentinel-2 data. In order to verify the water extraction performance of NMCWI， this paper compared several commonly used water extraction algorithms， including the improved normalized difference water index MNDWI， the revised normalized water index （RNDWI）， the normalized multiband water index （NDMBWI）， the vegetation red edge water index （RWI） and the triangle water index （TWI）， and selected six experimental areas for water extraction experiment. The results show that the overall accuracy and Kappa coefficient of NMCWI in each test area are higher than 95%， which is higher in index extraction accuracy compared with other water bodies， which can reduce the interference of most non-water bodies， such as shadows， buildings and ice. Meanwhile， NMCWI is applicable to most areas and can effectively extract water bodies.

Drought is one of the most frequent and destructive natural disasters that occur in arid and semi-arid regions. Its high frequency， long duration， and wide impact have had a serious impact on ecological， economic， and agricultural systems worldwide， and its spatiotemporal characteristics have become increasingly complex under the background of climate change. The research on multi-scale meteorological drought characteristics and potential influencing factors in different climate zones aims to reduce drought losses and promote sustainable development of the watershed. Therefore， this article takes the Yellow River Basin as the research object， fully considering the spatial heterogeneity of meteorological and underlying surface elements in the basin. The Mann-Kendall trend test， run theory， and multi-time scale SPEI index method are used to reveal the temporal and spatial evolution characteristics of multi-scale meteorological drought in different climate zones， and quantitatively evaluate the relative contribution of extreme climate variables to multi-scale meteorological drought in different climate zones. The results show that， in terms of time course， the overall trend of meteorological drought in the watershed at multiple time scales follows a pattern of "humid-arid-humid"， with the highest degree of drought mainly occurring around 90 years； In terms of space， the degree of drought in the middle reaches of the watershed （D and E zones） is the highest， followed by the downstream， and the upstream is the smallest. Drought characteristic variables and extreme climate indicators （precipitation and temperature） exhibit strong spatial heterogeneity in different climate regions. SDII， R10， R20， CDD， R99pToT， R95pTOT， R95p， Rx5day， Rx1day， TX90p， and SU variables explain more than 80% of the evolution characteristics of meteorological drought， and Rx5day variables dominate the contribution of meteorological drought at different time scales， with a contribution rate of about 65%. The research results provide important information for watershed ecological protection， water resource management， and climate change adaptation， and effectively improve the predictability of drought processes.

Tire wear particles （TWPs） generated during vehicle driving can easily enter the bioretention system （BRS） through road stormwater runoff and continuously accumulate， posing a potential threat to the nitrogen removal function of the system. Therefore， the impact of different durations of TWPs stress on the nitrogen removal efficiency of BRS was investigated in this study. By considering the influence of TWPs on soil physicochemical properties and microbial metabolic activity， the primary driving factors of TWPs stress on nitrogen removal performance were identified. The results showed that short-term stress （0~30 d） from TWPs exerted no significant influence on the pollution removal performance of BRS. However， long-term TWPs stress （31~101 d） weakened the removal capability of BRS for NH₄ ⁺-N and TN， resulting in a decrease in removal rates by 16.90% and 6.89%， respectively （P<0.001）. The nitrification rate （NR） of the planting layer was significantly inhibited （P<0.05）， while it significantly enhanced the substrate-induced respiratory rate （SIR） and dehydrogenase activity （DHA） （P<0.01）. Moreover， long-term TWPs stress significantly increased soil organic matter （SOM） and TN content， but decreased soil NH₄ ⁺-N content significantly as well （P<0.01）； however， it had no impact on the soil properties of the submerged layer. Additionally， partial least squares path model analysis revealed a significant positive correlation between soil health chemical indexes （SOM and TN） and biological indexes （DHA and SIR） in planting layer （0.001<P<0.01）. Furthermore， an increase in SIR and DHA played a pivotal role in reducing the removal performance of NH₄ ⁺-N by BRS. The findings confirmed that the long-term accumulation of TWPs in BRS would adversely affect microbial metabolic activity within the planting layer， thereby directly impacting nitrogen removal processes.

One-dimensional consolidation tests， scanning electron microscope tests， and mercury intrusion porosimetry tests were conducted to explore the compression properties and microstructure of compacted clay under dry-wet and freeze-thaw environments with different initial moisture contents. A compression model considering the influence of environmental factors was proposed to describe the relationship between compacted clay void ratio and compression stress. The results indicate that the optimally compacted clay samples with optimal moisture content on the dry side exhibit an aggregate structure with a bimodal pore size distribution curve， while those on the wet side show a dispersed structure with a unimodal pore size distribution curve. Both dry-wet and freeze-thaw environmental effects can promote the transformation of small pores to medium and large pores within the two types of samples， leading to improved compressibility characterized by a reduction in initial consolidation stress and an increase in recompression index， with limited variation in compression index. The influence of dry-wet cycles on the pore system and compressibility of samples is greater than that of freeze-thaw cycles， with a significantly higher enhancement in compressibility observed in wet-side samples compared to dry-side samples under environmental effects. The proposed compression model considering environmental influences effectively describes the compression properties of compacted clay.

The results of multi-objective optimal scheduling of reservoirs in stochastic environment is a set of non-inferior solution sets with uncertainty， and the traditional deterministic decision-making methods are difficult to quantitatively consider the synthetical uncertainties and risks in the decision-making process of reservoir scheduling. If scheduling is based on deterministic decision results， it will create risks for the actual scheduling of the reservoir. In order to cope with the influence of uncertainty factors on the reservoir scheduling decision-making problem and to reduce the risk of reservoir scheduling decision-making in a stochastic environment， this paper proposes a risk-based group decision-making model for reservoir scheduling considering multiple uncertainties. Firstly， we quantify the uncertainties of the decision indicatorsindexes by probability distribution. Then， we propose the resolution of decision-maker group preference conflict and empowerment method based on entropy weight method， fuzzy analytic hierarchy process， game theory and the principle of minimum deviation， and then adopt the feasible weight space to derive the feasible domain of the index weights. Finally， we establish a risk-based group decision-making model based on SMAA-GTDM for reservoir scheduling， and then propose a two-stage process for risk-based group decision-making.The decision risk degree index is defined to assess the reliability of the decision results. Applying the proposed methodology to a flood control system in the Pubugou Reservoir of Dadu River basin in China， the feasibility and superiority of the risk-based group decision-making model based on SMAA-GTDM and its two-phase decision-making process in dealing with decision-making risky problems are demonstrated by a comparative analysis with the deterministic GTDM model. Further onwards， we conduct comparative verification experiments with the original SMAA-2 model. The results show that the probability of obtaining the optimal rank for the scheme with the best comprehensive rank in the two model results is 68.95% and 45.61%， respectively， with confidence factors of 69.1% and 46.22%， respectively， and decision risk degrees of 0.04% and 1.55%， respectively. This indicates that the SMAA-GTDM model can provide more explicit ranking results in the stochastic environments， significantly reduce the decision-making risk degree， and provide more robust decision-making support for reservoir scheduling.

Digital twin technology has been developing rapidly in the water industry for several years. Among it， the digital twin data base plate stores large amounts of information in the whole life cycle of water conservancy projects， and there are problems such as large data volume， complicated management and increased updating cost. Meanwhile， the application layer digital twin platform with WebGL as the core engine rendering load is limited by network bandwidth and computer performance， facing the challenge of light-weight processing. Aiming at the geometric model data in the virtual entity of the digital twin， this paper uses the improved Quadric Error Metrics（QEM） and Low-Poly algorithm to carry out light-weight processing. The hydraulic turbine guide mechanism is selected as a case to study the optimal simplification degree to meet the needs of the digital twin business. Results show that light-weight processing is necessary and the file size is greatly reduced after lightweight processing. The upper limits of the simplification degree of the four precision levels in meeting the digital twin fineness level is the original data， about 40%， about 70%， about 98%.

In order to explore the effect of freezing-microwave method for purifying sewage quality， this study designed an experiment of indoor freezing wastewater， combined with microwave rapid thawing， and the melted water was collected at five times intervals （i.e. T1~T5）. The results show that： ① The microwave thawing of 17 min can completely melt the sewage ice， and the efficiency is much higher than that of the natural thawing process （14.5 h）. ② The microwave treatment of the first 8 min （T1+T2） melts 40% of the total water and removes more than 59.18% of the physical and chemical index content and 64.57% ion content in the sewage. The chromaticity， turbidity and yellow flocculent precipitation of sewage were significantly reduced. ③ In the final aquatic products obtained by microwave thawing for 17 min， the concentration ratio of total alkalinity， HPO₃， PO₄ ^3-， COD_Mn， SiO₂ and H₂SiO₃ is between 0.20 and 0.41， and the maximum concentration ratio of TDS to total hardness is only 0.16. The concentration ratio of eight conventional ions in K⁺， Na⁺， Ca²⁺， Mg²⁺， Cl^-， SO₄ ^2-， HCO₃ ^- and NO₃ ^- wastewater is only between 0.05 and 0.20. The concentration of As is close to 0. ④ The results of principal component analysis （PCA） showed that the purification effect of TDS and total anions was the best when freezing-microwave method was used to purify sewage water， and the principal component comprehensive scores Y were 4.55 and 3.65， respectively. In general， combined with freeze-thaw process and microwave rapid thawing， sewage treatment can be realized quickly and efficiently.

Microcapsule self-repairing technology is an effective method to solve the deterioration of concrete and improve its durability. In recent years， researchers have revealed the self-repairing function of microcapsule by studying the self-repairing microcapsule on millimeter-scale and hundred-micrometer-scale. However， the mechanical properties of matrix concrete are reduced. The microcapsules with particle size of 40 μm and 8 μm were synthesized by controlling the mixing rate and other measures， and the effects of the microcapsules on the mechanical properties of concrete were studied. Results show that： ① The thermal stability of the two particle sizes of microcapsules is good， the large-size microcapsules are broken after adding concrete， and the surface of small-size microcapsules becomes rough； When concrete is damaged and cracks appear， microcapsules will rupture in a timely manner， and the core material will flow out to repair the concrete. ② The compressive strength and flexural strength of concrete decrease with the addition of 40 μm microcapsules， while the compressive strength and flexural strength of concrete increase with the addition of 8 μm microcapsules when the content is less than 4% and begin to decrease after exceeding 4%； ③ Both 40 μm and 8 μm microcapsules can repair the damage caused by compression and bending of concrete.The maximum recovery rates of compressive and flexural strength of 40μm microencapsulated concrete are 110.2% and 113.0% respectively， and those of 8μm microencapsulated concrete are 115.6% and 114.7% ，respectively. However， when the amount of microcapsules is increased to a certain extent， it will reach an inflection point， and then the repair effect begins to decrease.The application of hydraulic concrete can effectively improve the mechanical properties of concrete by controlling the particle size of microcapsules， and timely self repair when the concrete is damaged.

Considering the non-renewable and indispensable nature in crop production of phosphorus resources， the treatment and recovery of phosphorus from wastewater based on struvite crystallization method can achieve resource reuse and ensure longer life span of phosphorus rock. This manuscript systematically studied the response optimization of phosphorus recovery from phosphogypsum landfill leachate under varying temperature （5~30 °C）， providing theoretical and data support for treating leachate wastewater and obtaining high quality of struvite fertilizer. The results demonstrated that under the premise of Mg/N/P molar ratio of 1/1/1， the highest phosphorus recovery rate in the leachate wastewater was achieved at a temperature of 10 °C， being 77.5%， experiencing the largest solution pH reduction from 8.8 to 5.86 after struvite crystallization. As the crystallization temperature continued to increase， the phosphorus recovery rate showed a significant decreasing trend， and the phosphorus recovery rate was only 35.6% at 30 ℃. Meanwhile， the crystal particle size basically presented an increasing trend with the increase of temperature， and the largest size of crystal particle （68.2 μm） were formed at 30 ℃. Through X-ray diffraction and Micro-Fourier transform infrared spectroscopy analysis， the purity of struvite in the crystals formed under different temperature conditions is high， and there is basically a small amount of calcium phosphate. However， the ammonium content in crystals precipitated in the leachate was reduced with an increasing temperature， resulting in a lower purity of struvite. The suggestion is given that phosphorus recovery from leachate should be processed under lower temperature conditions （10~15 ℃）， and the phosphorus resource can be recovered from the wastewater after adding alkaline， achieving an optimal balance among phosphorus recovery rate， struvite particle size and purity.

In order to address the power supply planning optimization issues during power shortages in power systems， considering the economic operation characteristics of the optimal combination of units， the complementary characteristics between multiple power sources， and the fluctuating nature of electricity prices， a multi-factor and multi-objective optimization model for supplementary power sources， which takes into account the characteristics of power systems， is established with the objectives of minimizing the total cost and maximizing the operational efficiency. By treating the costs and benefits of the construction and operation phases as a whole， and using constraint methods to transform the multi-objective problem into a single-objective problem， a solution method based on dynamic programming is provided. The results of the case analysis applied to a regional power system demonstrate that， without considering the constraints of various power resource conditions， the model can more economically and rationally provide an optimized recommendation scheme among a finite set of supplementary power supply options， offering a certain reference for the strategic optimization selection of supplementary power sources in power systems.

In response to the low efficiency and high cost of abnormal personnel and seepage safety detection for small earth and rock dams， an innovative intelligent inspection method is proposed. This method is based on UAV images and deep learning object detection technology， improves the YOLOv8 network structure， and introduces fine-grained convolution and a new MPDIoU loss function. Fine-grained convolution can capture richer image details， while the new MPDIoU loss function helps to improve the positioning accuracy of the model. Experiments were conducted on a self-made dam abnormal personnel and seepage dataset. The experimental results prove that the proposed intelligent recognition algorithm has significant advantages in the comprehensive recognition of personnel and seepage， improves the detection precision of intruders in the reservoir area of the earth and rock dam project and dam seepage， and can adapt to the actual application needs in different environments.

In order to make a network plan better resource leveling for roller compacted concrete（RCC） dams，so as to improve resource utilization and reduce engineering investment， the paper applies the hybrid particle swarm（Hybrid PSO） to the optimization of “Construction Period Fixed and Resource Leveling”of RCC Dam. A resource intensity variance evaluation function with process start time as independent variable is constructed， and simulated annealing（SA） algorithm is used as the convergence criterion of particle swarm optimization（PSO） algorithm to improve its global convergence ability and convergence accuracy. A speed detection method based on dynamic time difference is introduced，the Hybrid PSO is effectively integrated with“Construction Period Fixed and Resource Leveling”of RCC Dam，and an optimization mathematical model under multiple constraints is established.In the application of engineering case study， we calculate and obtain network plan with better resource leveling，and verify that the application of the Hybrid PSO in “construction period fixed and resource leveling” of RCC Dam is reasonable and effective，the convergence speed is faster and the results are better than traditional methods.

A 23 m wide fault fragmentation zone is developed at the foundation of a medium-sized reservoir gravity dam in Southwest China， and the mechanical properties of the rock body in the fragmentation zone are weak， which will have an important impact on the stability of the dam foundation during the construction， impoundment and operation of the dam. Based on the engineering geological conditions and rock mechanical parameters， a micro-arch concrete plug replacement foundation scheme for dam foundation treatment is proposed， and three-dimensional finite element calculations and sensitivity analyses of rock mechanical parameters of the F1 fault at the dam foundation are carried out. It is concluded that in the case of soft bedrock body with wide fracture zone， the difference of deformation modulus and the weak bearing capacity of the fault zone lead to the uneven deformation phenomenon of the dam foundation. After the replacement of the concrete plugs by micro-arch， the load carrying capacity of the dam foundation and the stress deformation of the gravity dam during the impoundment process meet the requirements. As the concrete plugs can work well， the influence of the variation of fault fracture zone parameters on the stress and deformation of the dam and dam foundation is weakened. The study provides an important reference value for the foundation treatment of gravity dams across wide faults.

Water use efficiency of rice and field productivity were affected by irrigation quota. In order to explore the variation law of rice irrigation quota and its correlation with yield， the experiment of water layer location monitoring and supplementary irrigation of paddy field was conducted for eight consecutive years from 2014 to 2021 in Nanning Irrigation Experimental Station， providing fundamental data for quantifying the irrigation quota of rice in the double cropping rice region of southern Guangxi. The upper and lower limits of the water layer in paddy fields were established in regreening stage， tillering stage， jointing and booting stage， heading and flowering stage， milk ripening stage， and yellow ripening stage. The water layer depth was monitored daily by measuring needles. Supplementary irrigation was performed when the depth was below the lower limit， and the cumulative irrigation amount was calculated. At the same time， the rainfall was measured by meteorological facility in Nanning Irrigation Experimental Station. In mature stage of rice， indicators such as panicle length， 1 000-grain weight and yield were measured. The test results show that： ① The amount and time distribution of rainfall were important factors influencing the irrigation quota of paddy fields， and the uneven distribution of rainfall during the growth periods of rice was the main factor causing significant differences of irrigation quota. The irrigation quota of early rice was 210.1~438.4 mm， while that of late rice was 243.9~477.5 mm. ② When the irrigation quota was within the range of 200~320 mm， panicle length， 1 000-grain weight and yield of rice increased with the increase of irrigation amount. While the irrigation quota was more than 320 mm， continued irrigation was not conducive to the increase of rice yield. It could be seen that the appropriate irrigation quota for rice was 320 mm during the experiment. Insufficient or excessive irrigation would affect rice yield and reduce the effective utilization rate of water resources.

In order to address the issue of traditional creep models failing to accurately reflect the characteristics of rock accelerated creep stage， this paper proposes an enhanced unsteady Maxwell creep model based on the conventional Maxwell model. It takes into consideration the unsteady creep parameters of original clay pot elements in series with a nonlinear viscoplastic body （NVPB）， and deduces the corresponding creep constitutive equation. The improved model comprehensively simulates the nonlinear creep behavior throughout all stages of rock creep， thereby enhancing accurate description. Additionally， origin software is utilized for fitting different rock samples. The research findings demonstrate that this new creep constitutive model better fits the experimental data and identifies parameters， effectively simulating rock's deformation under various stress levels while accurately describing the nonlinear creep characteristics of its three stages of creep. This indicates high accuracy and strong applicability of the newly established model. It provides a new creep constitutive model for future geotechnical engineering.

Bacteria are an important part of the aquatic ecological environment， and the bacterial community structure will change with the change of environmental factors， and it also has a counter-effect on the aquatic environment. In order to explore the impact of ecological restoration project on the microbial community structure of drainage ditch water and sediments， water and sediment samples were collected and investigated in August 2021 （before the implementation of the project）， December 2021， May 2022 and July 2022， and the contents of eight physical and chemical factors （pH， conductivity， salinity， total nitrogen， CODMn， CODCr， total phosphorus and F-） in the water were determined. The characteristics of bacterial community structure in the water and sediments of the fifth drainage ditch were analyzed by 16S rDNA high-throughput sequencing， and the response relationship between the bacteria community and the physical and chemical factors of the water body was explored. The results showed that the conductivity， salinity， total nitrogen， total phosphorus， CODMn and CODCr of the water body were significantly reduced and the water quality was improved after the implementation of the ecological restoration project. The species diversity index of bacteria in the water body increased， the community abundance increased， the uniformity increased， the dominant bacteria phyla and bacterial community structure and composition changed significantly， the relative abundance of the dominant bacteria phyla Proteobacteria and Bacteroides increased， and the relative abundance of Cyanobacteria decreased. There was no significant change in the dominant phylum of sediment bacteria before and after the implementation of the ecological restoration project， which were Proteobacteria， Bacteroides and Desulfobacterota， and the relative abundance of Proteobacteria increased after the implementation of the ecological restoration project， while the relative abundance of Bacteroidetes and Desulfobacterota decreased. The diversity， richness and uniformity of bacterial species in sediment were higher than those in water， and there were great differences in spatial and temporal distribution. TP and F- were the main influencing factors of the bacterial community structure in water， and salinity， TP and CODMn were the main influencing factors of the bacterial community structure in sediments.