The Simulation of daily water and nitrogen output processes during the snowmelt runoff generation period in seasonal freeze-thaw zones remains challenging due to influences from processes such as snowmelt and soil freeze-thaw. In this study， we took the Heidingzi River basin in Changchun City， Jilin Province， a typical seasonal freeze-thaw zone， as an example. Calibration and validation of the Soil and Water Assessment Tool （SWAT） model were conducted using daily monitoring data of water and nitrogen at the basin outlet during the freeze-thaw period from 2014 to 2016. The applicability of the SWAT model in simulating daily runoff and daily nitrogen loads during the thawing period was explored. The results indicated that within the SWAT model， the parameters CN₂， CNFROZ， SNOCOVMX （for daily runoff） and CN₂， SDNCO， CNFROZ （for daily nitrate nitrogen output） were the three most influential parameters. The SWAT model performed well in simulating daily runoff， with Nash-Sutcliffe Efficiency （NSE）， coefficient of determination （R2） and relative error （Re） values of 0.75、0.78 and -12.76% during the calibration period， and 0.54、0.51 and 5.65% during the validation period， respectively. The primary reasons for the decline in accuracy were that the SWAT model did not consider the lag effect of snowpack on runoff generation and the snowmelt-refreezing process. Additionally， parameters adjusted for accurate simulation of the snowmelt runoff process during the soil thawing period often led to overestimation of runoff and underestimation of baseflow during non-freeze-thaw periods. Due to biases in runoff simulation and inadequate representation of the impact of freeze-thaw processes on nitrogen transformation within the model， the SWAT model exhibited relatively low accuracy in simulating daily nitrate nitrogen loads， with NSE， R2 and Re values of -0.19、0.44 and 2.7% during the calibration period， and 0.60、0.26 and -13.79% during the validation period， respectively. This study validated the applicability of the SWAT model in seasonal freeze-thaw zones and revealed its limitations， providing a scientific basis for model optimization and application.

For a long time， due to excessive exploitation of water resources and pollution emissions， the ecological environment of many rivers and lakes has been severely damaged. Therefore， it is imperative to promote ecological protection and governance of important rivers and lakes. Taking the Maota River Basin as an example， this paper elaborates on the engineering measures for the comprehensive governance of rivers and lakes， including actively promoting water system connectivity， dredging rivers， ecological treatment of river banks， and restoration of water space. These measures have solved the problems of river water body fragmentation and siltation， restored important ecological corridors and connection channels of rivers and lakes， improved water environmental quality， improved the habitat conditions of organisms， effectively protected regional ecological diversity， and at the same time enhanced regional flood control capacity and ensured regional water supply security.

The Lingnan Waterfront Landscape， rich in regional characteristics， is facing the challenges of losing its unique features and the disconnection between people and the landscape under the rapid urbanization process. This study applies the Spatial Gene theory， using Langtou Village in Huadu， Guangzhou as a case study. By combining online social data with field investigation， the spatial genes of the Lingnan water town landscape are extracted from three levels： terrestrial landscape， settlement landscape， and architectural landscape. The study further analyzes and evaluates the interaction between these spatial genes' traditional mechanisms and the current interactive relationship between people and landscape. The findings reveal that the spatial genes of the Lingnan water town can be summarized as “Rivers meandering in front， Fish ponds and paddy fields surrounded with dykes，” “Comb-like layout， Narrow alleys，” and “Black bricks and grey tiles， Closely packed Huoer roof ridges.” These spatial genes show varying disruptions in the modern context. Additionally， in light of the educational tourism development trend of the case study， the study explores the design paths for the landscape renewal of the Lingnan waterfront， providing a strategy to address the unique features losing problem in water waterfront landscapes. This promotes the continuity and innovation of the Lingnan waterfront landscape’s value， characteristics， and form.This study carries out the discussion on the renewal design path of Lingnan Waterfront Landscape， promots it’s inheritance and innovation in value connotation， element characteristics， expression form and other aspects， and enriches the solutions to the common problems of Waterfront landscape area.

The arrangement of cylindrical obstacles in the pool chamber is a method to change the flow structure of vertical slot fishway. In order to investigate the influence of cylindrical obstacles arrangement on the hydraulic characteristics of vertical slot fishway， the RNG k-ε turbulence model was used to numerically simulate the vertical slot fishway. This paper mainly studies the hydraulic characteristics of cylindrical obstacles in five different locations and three different sizes， focusing on the flow pattern， velocity， turbulence kinetic energy and other hydraulic characteristics of three cylindrical obstacles of different sizes in different water depth sections. The results show that： the installation of the cylindrical structure in the reflux zone of conventional vertical slot fishways will make the flow velocity larger at the vertical slot. The hydraulic characteristics of the cylindrical arrangement in the position related to the main flow of a conventional vertical slot fishway are better than those in the reflux zone. Three different sizes of columns at position 3（defined as the position with a 30° clockwise angle to the vertical bisector of the vertical slot and 2b from the center of the vertical slot， b being the width of the vertical slot） significantly reduced the maximum flow velocities and turbulence kinetic energies， and the placement of a 0.5b-diameter column at position 3 nearly doubled the maximum total hydraulic strain when compared to a conventional vertical slot fishway， which may adversely affect the upstream movement of fish； The 1.0b diameter column at position 3 was the most effective in optimizing the flow； the 1.5b diameter column at position 3 was not as effective as the 1.0b diameter column， and the maximum Total Hydraulic Strain of the two combinations is similar to that of conventional vertical slot fishways. The size and location of the columns have an effect on the hydraulic characteristics， and it is extremely important to choose a reasonable size and location of the columns for the hydraulic characteristics of the vertical slot fishway.

River-crossing bridges can alter the flow structure of river channels， and excessive flow-blocking effects may compromise both flood control safety and the structural safety of the bridges themselves. The blockage ratio， a key parameter quantifying the water resistance of the bridge， deserves thorough investigation. This study evaluates the effects of bridge types， bridge span arrangements， and skewed piers on blockage ratios through data analysis and generalized calculations. Results show that arch bridges and suspension bridges exhibit relatively smaller blockage ratios， followed by cable-stayed bridges， while continuous girder bridges demonstrate comparatively higher blockage ratios. Under identical pier diameters， larger span lengths between piers correlate with reduced blockage ratio. Notably， an increase in the oblique angle between piers and water flow significantly amplifies the blockage ratio， suggesting optimal control of oblique angles within 5° or 10°. The 1D and 2D mathematical model are employed to calculate the backwater height under the design flood condition in Wuhan river-reach of the middle Yangtze River. Results show that when the blockage ratios range between 2% and 20%， 1D modeling indicates a backwater height of 0.02~0.22 m at the cross-section upstream of bridge， 2D modeling indicates local backwater height of 0.12~0.30 m in front of piers and an average backwater height of 0.01~0.14 m at the same cross-section. The discrepancy in computational results between the two models stems from differences in modeling principles and simplification approaches. Based on comprehensive backwater calculations and relevant technical standards， it is recommended to maintain the blockage ratio of river-crossing bridges within 5%， and the blockage ratio requirements for river sections in mountainous canyon reaches without critical flood-protection objects may be appropriately relaxed upon technical confirmation of minimal impact on flood control safety. Research proposes the following measures to reduce the blockage ratio： selecting bridge types with lower blockage ratio， maximizing span lengths， adopting streamlined or circular pier shapes， aligning pier axis parallel to the flow direction， and so on. These recommendations provide a scientific foundation for bridge engineering design， flood impact assessments， and regulatory management of river-related projects.

Heavy metal pollution in rivers is a prominent issue faced by mining activities， which seriously threatens the ecological and environmental health of the basin. In order to clarify the content distribution and pollution sources of heavy metals in the river close to the metal mining area， the contents of Fe， Mn， Cu， Cd， Pb， Zn， Cr and Ni in sediments of the Le′an River during the wet and dry seasons were determined. The enrichment factor （EF）， consensus-based sediment quality guidelines （CBSQGs）， principal component analysis （PCA）， positive matrix factorization （PMF） and metal stable isotope （δ⁶⁵Cu） were used to identify the contamination level， ecological risk and source apportionment of heavy metals in Le′an River. The results show that the enrichment of Cu and Cd is more prominent in the sediments， and the increases of Fe， Cr and Ni are not significant comparing with their background values. In terms of spatio-temporal distribution， Cu content is high in the Dawu River sediment and Cd， Pb and Zn concentrations are high in the Jishui River sediment， but the variations are not obvious between the wet season and the dry season. The enrichment factors are in the order of Cd > Cu > Zn > Pb > Cr > Ni. Heavy metals （mainly Cu and Cd） in the sediment of the Dawu River， the Jishui River and the downstream of their confluence with the Le’an River have toxic effects on benthic organisms. Productive activities of copper and lead-zinc ores are the main sources of Cu， Cd， Pb and Zn， while Fe， Cr and Ni are mainly from natural processes. Cu isotopes further indicate that Cu may be derived from the peripheral waters of the copper mining area， and the influence of tailings-derived mine drainage water and other sources on the downstream Cu distribution of the Le′an River still needs further research.

The release of total nitrogen （TN） from river sediment is one of the key factors influencing water quality. This study focuses on the Guanlan River in Longhua District， Shenzhen， using intact sediment core samples to simulate the TN release process at the sediment-water interface through laboratory experiments. This study analyzes the release characteristics of TN under various environmental conditions and its impact on water quality. The results indicate that the release of TN in sediment exhibits seasonal variation， with the highest release rate in summer and the most significant fluctuations in winter. The release rate follows the trend of "static < weak disturbance < strong disturbance" as disturbance intensity increases. Under static conditions， TN release from sediment samples at each point is primarily driven by adsorption. Under disturbed conditions， there are significant differences in release rates at various points along the mainstream. Multi-year average runoff calculations suggest that the release of TN in sediment has a limited effect on water quality improvement； however， under flood seasons or extreme flooding conditions， localized releases may affect water quality. The study indicates that sediment release characteristics are influenced by seasonal runoff variations and human activities， and water environmental management should prioritize high-risk river sections to develop targeted pollution control and ecological restoration strategies. This research provides a scientific basis for the management of river water environments and sediment in China.

Exceeding the standard of total phosphorus concentration is one of the reasons for varying degrees of eutrophication in different areas of Chaohu Lake. Accurately predicting the changes in total phosphorus concentration is of great significance for pollution control of the watershed. In order to precisely predict the total phosphorus concentration in Chaohu Lake， a VMD-TCN-BiLSTM model based on XGboost feature screening is proposed. This model integrates the advantages of Variational Mode Decomposition （VMD）， Temporal Convolutional Network （TCN）， and Bidirectional Long Short-Term Memory （BiLSTM）. VMD is used for preprocessing the raw data to extract components of different frequencies； TCN captures local dependencies in time series data； and BiLSTM， which is skilled at handling long-term dependencies， effectively addresses the vanishing gradient problem often encountered in traditional LSTM models. Finally， the SHAP algorithm is utilized to interpret the constructed model. The VMD-TCN-BiLSTM model has an excellent predictive performance in predicting the total phosphorus concentration at the Zhongmiao section of Chaohu Lake， with R2 of 0.986 6， MAE of 0.002 2， and RMSE of 0.003 2， and has solved the problem of low prediction accuracy and predicted values being less than 0 in other comparative models. In the sections of the center of the western half of Chaohu Lake， Huanglu， and Xinhe inflow area of Chaohu Lake， the model still has high predictive accuracy （R2＞0.98）， indicating that the model has strong generalization ability. The SHAP algorithm is used to interpret the neural network model， revealing the degree of influence of characteristic variables on the predicted value of total phosphorus. The ranking of influence at Zhongmiao section is：Turbidity > Permanganate Index > Total Nitrogen > Dissolved Oxygen > Water Temperature > Conductivity.The ranking of influence at Huanglu section is： Turbidity > Permanganate Index > Dissolved Oxygen > Conductivity > Water Temperature > pH.The ranking of influence at the West Half Lake Center section is： Turbidity > Permanganate Index > Water Temperature > Dissolved Oxygen > Total Nitrogen > Conductivity.The ranking of influence at the Xinhe Lake Inflow section is： Turbidity > Permanganate Index >Water Temperature > Total Nitrogen > Dissolved Oxygen > pH.

Precipitation is one of the most important driving factors in hydrological processes， and its data source and accuracy are key factors affecting the accuracy of flood forecasting. Affected by many factors such as the imperfect water and rain condition monitoring system and the small range of the basin， the accuracy of the measurement of precipitation data in the small and medium -sized basin is high， but the distribution is uneven， which will affect the simulation results of hydrological models. In order to improve the accuracy of flood forecasting in the watershed， this study combines the measurement data with satellite product precipitation data by constructing a CNN-LSTM neural network model， and combines the HEC-HMS model to carry out the flood simulation Taking the catchment area above the Jinpen Reservoir in the Heihe River Basin in Shaanxi as an example， the application effect and applicability of satellite-ground fusion precipitation in the flood simulation are discussed. The results show that ① The HEC-HMS model with the site precipitation data as the input has good applicability in the research area， and can achieve Class B accuracy in both the rate period and verification period. ② The correlation coefficients between the two types of IMERG satellite products and measured precipitation are low， which are generally higher than the actual value， and the error is large. After CNN-LSTM data fusion， it is close to the measured precipitation data， and the fusion effect of the IMERG-Early product is better. ③ The HEC-HMS model after fusion of satellite products has a simulation pass rate of 80% for the 10 floods， with an average definitive coefficient of 0.856 9， and the average absolute value of the absolute value of the flood peak is 0.8h， reaching Class A accuracy. ④ The certainty coefficient of the flood simulation results of satellite-ground fusion precipitation increases， and the absolute value of the average flood peak time difference is reduced； the simulation effect of the flood peak flow decreases and is significantly smaller. The results show that integrating the actual measurement precipitation data and appropriate satellite precipitation product data in small and medium -sized basin can improve the accuracy of flood forecast models to a certain extent.

Exploring the driving effects of extreme floods and human activities on riverbed evolution not only helps improve river flood safety， but also provides scientific guidance for river management. Based on the multi-year measured river terrain data and contemporaneous water-sediment data of the Nandu River and the Wanquan River in Hainan Island， the interannual variation of water depth was analyzed， revealing different spatiotemporal variation modes and driving mechanisms of water depth. The results showed that the average water depth of the Nandu River and Wanquan River increased significantly by more than 1 m， and the overall riverbed showed a trend of erosion and downward cutting. However， there were differences in spatiotemporal variation characteristics and driving factors， mainly including extreme floods， water-sediment conditions， and direct impacts of human activities. The extreme flood has driven the overall erosion of the riverbed of the Nandu River， with a relative contribution of 59%； In addition， the riverbed has experienced shallow erosion and main channel siltation due to water-sediment conditions， accounting for 30% of the total water depth variation. In the Wanquan River， the water-sediment conditions dominated the occurrence of shallow siltation and main channel erosion， with a relative contribution of 77%； Meanwhile， the riverbed of the Wanquan River is approximately eroded due to extreme floods， accounting for 10% of the total water depth variation. Under the direct influence of human activities， water depth variation of both the Nandu River and the Wanquan River exhibited spatiotemporal randomness， with relatively small contributions （3.6% for the Nandu River and 3.9% for the Wanquan River）.

Based on the comprehensive regionalization results of human geography， territory spatial， agriculture cultivates and water resources in Yunnan province， it was divided into 7 geographic-landforms regions include mountainous basin （abbr. MB）， plateau lake watershed （abbr. PLW）， dry-hot valley （abbr. DHV）， low hills region of southwest and southern Yunnan （abbr. LHR）， Wumeng mountain area （abbr. WMA）， Qinghai-Tibet plateau southeast edge （abbr. QTPSE）， Rocky desertification of southeast Yunnan （abbr. RD）， etc. Lorenz Gini coefffcient， asymmetry coefffcient， matching distance， and imbalance index are used to assess the spatial matching and balance of sunshine thermal， water resources， farmland and social- economic development. The results show： ① yearly and monsoon seasonal precipitation of the whole province decreased from 1950s， while the dry season precipitation increased to some extent except WMA， the time serial of precipitation′s Gini coefffcient and Concentration degree decreased gradually while asymmetry coefffcient was increased in most regions. The relationship between precipitation′s Gini coefffcient and Concentration degree was highly linear related， Gini coefffcient and asymmetry coefffcient also correlated. ② The mismatches among various combinations of farmland with water resources， regional Gross domestic production （abbr. GDP）， and population were higher than other factors in all geographic -landforms regions. PLW was the highest mismatches distance region of human-land-water-heat system， and the order from highly down to lowly after PLW in turn was QTPSE， LHR， MB， DHV， RD and WMA. The main reason was highly centralized population， city， secondary and tertiary industries， it caused multi-type water shortage， forced those lakes supplied more and more water demands， and result in many longer distance trans-watershed water supply engineering planned and constructed. ③ It was the most important measure to solve the problems， that includes water resources saving and protected， built modern water net system on Yunnan plateau based on the central Yunnan water diversion project， green and harmonious development between water and human being. Furthermore， it also provide continuous water safety guarantee for rural revitalization， vitalize border areas and enrich the people living there， and modernization construction.

In order to deeply analyze the transformation relationship between desert groundwater and lakes in the contact zones of basin in the western region of China， this study takes Tengger Lake in Ningxia as an example， and uses the in-situ monitoring profiles constructed in the lakeshore zone to analyze the water level dynamics， hydrochemistry and isotopes of lake water and surrounding groundwater， so as to determine the conversion relationship between lake water and groundwater. The coupling model of lake water and groundwater was established by using moldflow and LAK3 modules in GMS to simulate the conversion amount of lake water and groundwater. The results show that the Tengger Lake is a typical groundwater-draining-type lake， which means the upper part of the groundwater recharges the lake， the part passes through the bottom of the lake， and the lake also leaks and recharges the lower part of the groundwater.The results of the coupled model of lake water and groundwater show that the water level change simulated by the model is basically consistent with the actual water level change trend of the observation value. The Nash index is 0.709 between the observed value of lake water level and the simulated value， the root-mean-square error is 0.026 8 m， the decisive coefficient R ² is 0.799， and the Nash index between the observed value and the simulated value is 0.718. The root-mean-square error is 0.052 m and the decisive coefficient R ² is 0.789， which can accurately simulate the fluctuation of lake level and describe the interaction between groundwater and lake water. The simulated groundwater recharge amount of lake water is 516.11×10⁴ m³/a， and the amount of water leaking from lake to groundwater is 117.3×10⁴ m³/a.

The accuracy of runoff simulation is very important for hydrological forecasting， water resource management and basin flood control. At present， runoff simulation at home and abroad mainly adopts hydrological model method， statistical method and machine learning， among which ARMAX-GARCH time series model belongs to mathematical statistics method， which has the advantages of less required data， easy operation and being able to describe nonlinear processes， and has been gradually applied to hydrological process simulation. In the middle and lower reaches of the Yangtze River， Dongting Lake， the Hanjiang River and Poyang Lake converge into the main stream of the middle and lower reaches of the Yangtze River successively， and the runoff processes of the main stream and tributaries interact with each other， and the intricate relationship between the exchange of dry and tributary water also increases the difficulty of runoff simulation in this region. Using the time series model ARMAX-GARCH， we first conduct a cross-correlation analysis on different runoff sequences. And then， the runoff sequences with strong correlation with the simulated target station are added into the model as exogenous variables to simulate the runoff process of the target station. In this paper， the runoff from Yichang， Chenglingji， Hankou， Xiantao， Hukou， and Datong hydrological stations in the middle and lower reaches of the Yangtze River is simulated， and the runoff sequences at Yichang， Chenglingji， Xiantao， and Hukou stations are modeled by an ARMA-GARCH model， and the runoff sequences at Hankou and Datong stations are modeled by an ARMAX-GARCH model. The results show that the model simulates runoff sequences well and can effectively take into account the influence of runoff sequences with strong correlation with the target hydrological stations； and analyzes the influence of the construction of the Three Gorges Reservoir on the downstream runoff mean， which is mainly manifested in the weakening of extreme runoff events in the downstream mainstem and the gradual decrease of this effect in the direction of the water flow. The study confirms the feasibility of using a time series model to simulate runoff and provides an important reference for hydrological simulation and forecasting in areas with complex river-lake relationships.

Studying the future precipitation and aridity characteristics of the Weihe River Basin under the background of climate change provides a theoretical basis for policy formulation and decision-making in water resources management and flood-drought disaster prevention in the basin. Based on the measured precipitation， average temperature， sunshine duration， average relative humidity and average wind speed from 18 meteorological stations in the Weihe River Basin， the potential evapotranspiration of these stations were calculated by the Penman-Monteith method， the four scenarios of SSP1-1.9、SSP2-4.5、SSP3-7.0 and SSP5-8.5 from the CanESM5 model in CMIP6 were selected， and the SDSM was used to predict the precipitation， potential evapotranspiration and aridity index in the Weihe River Basin from 2031 to 2090. The results show that SDSM can simulate potential evapotranspiration and precipitation in the Weihe River Basin well， with better simulation effects on potential evapotranspiration compared to precipitation. Climate change has a significant impact on precipitation and aridity index in the Weihe River Basin， with varying degrees of influence under different climate scenarios and time periods. Temporally， the future precipitation in the Wei River Basin will increase， with the interdecadal trend rates of precipitation under four scenarios are 2.34、13.3、32.5 and 25.1 mm/10 a respectively； In particular， the precipitation in the SSP3-7.0 scenario from 2071 to 2090 will increase by 35.2% relative to the baseline period； Except for the SSP1-1.9 scenario， potential evapotranspiration of the other scenarios shows an increasing trend； The aridity index decreases by an average of 0.145. Spatially， precipitation shows an increasing trend from northwest to southeast in the basin， with significant increases in precipitation above the Beiluohe Zhuangtou station as the forcing scenarios enhance； Potential evapotranspiration increases from west to east， and higher forcing scenarios result in greater increases in the same region； The aridity index decreases from northwest to southeast， and the climate of the basin becomes humid overall. In the future， the Weihe River Basin should enhance its ability to prevent and control flood and drought disasters by strengthening risk monitoring， allocating water resources scientifically and establishing a water and drought disaster early warning mechanism， so as to effectively respond to risks and challenges brought by climate change.

The network of rainfall measuring stations monitors the temporal and spatial characteristics of rainfall in the water basin， which is very important for regional water resources management and flood forecasting and early warning. Based on the information entropy theory， in this paper， a multi-objective rainfall station network optimization model with the maximum total rainfall information and the minimum mutual information of rainfall stations in the basin as the objective function is established， and the multi-objective optimization model is solved. At the same time， according to the rainfall and mutual information data of the water basin， the rainfall information-distance transmission characteristics of the water basin is analyzed，and the correlation coefficients of rainfall between rainfall stations in the water basin is calculated. The regional characteristics of rainfall in the water basin are analyzed and studied. Rainfall station network topology in watersheds is optimized by combining information-distance transfer characteristics and rainfall spatial and temporal characteristics. Taking the Yitong River Basin of Liuhe County as an example， the rainfall station network of the Yitong River Basin is optimized and evaluated. The total information entropy of the rainfall stations in the Yitong River basin is calculated to be 57.305 bits， and the total mutual information is 33.478 bits. Solving the multi-objective rainfall station network optimization model， the ranking of the non-inferior solution set of the basin rainfall stations is：Y19， Y18， Y5， Y21， Y4， Y15， Y1， Y14， Y20， Y11， Y2， Y16， Y6， Y10， Y22， Y12， Y3， Y8， Y7， Y9， Y17， Y13. Combining information-distance transfer modeling with spatial correlation of rainfall in watersheds， it is suggested to adjust the location of three rainfall stations and delete one rainfall station. The optimized network of rainfall stations can improve the efficiency of regional rainfall monitoring and reduce the cost of monitoring rainfall in the watershed.

Detailed and accurate monitoring of water depth and flow velocity is essential for effective urban flood rescue operations and risk assessment. Due to the increasing frequency of extreme weather events and the advancement of urbanization， traditional urban flood monitoring methods are no longer sufficient for regional and real-time monitoring. To address the lack of integrated regional monitoring methods for water depth and velocity in urban floods， this study developed a binocular vision-based monitoring system for urban flood areas. Two monitoring cameras with overlapping fields of view were employed to capture flood area images， enabling high-precision measurements of water depth and flow velocity. A flood area water depth reconstruction method based on feature points extraction and matching was proposed. Using the parallax principle of binocular vision， matching feature points in flood areas were extracted to calculate three-dimensional coordinates， thereby reconstructing the water depth distribution. A surface flow field measurement method based on the identification of floating objects coupled with the water depth distribution was also proposed. By identifying floating objects （such as leaves， debris， etc.） on the flood surface， the surface velocity is calculated and integrated with water depth distribution， enabling comprehensive monitoring of both water depth and velocity. To validate the feasibility of the method， a laboratory water tank was constructed to simulate flood conditions， and monitoring cameras were used to capture images for calculating water depth and velocity. The experimental results show that the water depth error is approximately 2.7 mm compared to wave probes， and the flow velocity error remains within 5% relative to particle velocity measurements. Compared to traditional monitoring techniques， this approach offers non-contact operation， reduced risk， regional monitoring capabilities， and eliminates the need for additional equipment.

Phase modulation and runner drainage control is one of the common operating conditions of hydropower stations， but failures often occur in actual phase modulation experiments. Among numerous influencing factors， the valve opening of the gas system is a key parameter. Due to the high cost of water pressurization experiments and the inability to obtain detailed information about the internal flow field， this paper adopts numerical simulation methods to conduct computational fluid dynamics research on air-water two-phase flow under different valve openings in the entire air supply system. The reasons and laws of the influence of valve opening on the water pressurization process and air escaping from the draft tube are obtained， and a combination of liquid level gauges and turbine power consumption is proposed as a criterion for determining the success or failure of phase modulation. The research results are helpful in solving the problem of phase modulation and runner drainage failure in hydropower stations.

The stability of excavation support for the top arch layer of deep buried caverns is significantly affected by local rock changes， structural surface distribution， and other factors， especially the overlying soft rock structure. Its deformation analysis and prediction are crucial for the construction safety of the top arch layer of underground caverns. Based on the excavation project of the underground powerhouse of Gongyi Pumped Storage Power Station， numerical simulation methods are first used to analyze the stress of the overlying soft rock structure and reveal its influencing laws； The characteristics of excavation unloading deformation and surrounding rock stability of the overlying soft rock top arch layer under the design of excavation support scheme are explored， and the stability effects of different excavation procedures on the overlying soft rock top arch layer are compared. The results show that the overlying soft rock has a significant impact on the displacement of the surrounding rock of the top arch layer. As the distance between the overlying soft rock and the top arch layer decreases， the displacement， stress concentration， and plastic zone of the top arch surrounding rock rapidly increase， with deformation having the most significant impact； Due to the influence of the overlying soft rock structure， the displacement of the main factory building in the top arch layer can be reduced by about 8.6% under the pilot tunnel scheme， while the displacement of the main transformer room can be reduced by about 15.9% under the pilot tunnel scheme on both sides. The distribution of stress and plastic zone is not significantly different. Therefore， the use of a group of caverns and a middle guide tunnel leading scheme is more applicable in the main power house. When the deformation control requirements of the main transformer room are high， the excavation procedure of “pilot tunnels on both sides first and middle pier follow-up” can be considered.

Establishing a water-wind-solar complementary power generation system is an effective way to fully utilize large-scale wind and photovoltaic power. Although there have been considerable studies on capacity allocation and optimization of operation modes for large-scale "wind-solar-cascade hydropower" complementary power systems， the impact of multi-energy complementarity on the operation mode of cascade reservoir systems， especially on water level control， has not been fully revealed. This paper takes a large-scale "wind-solar-cascade hydropower" complementary power generation system as an example and establishes a multi-objective optimal dispatch model that considers both the maximization of power generation and the maximization of minimum output. Finally， three typical years （abundant water year， normal water year， and dry year） are selected for comparative analysis of output processes and water level control processes， comparing the differences between the complementary operation of cascade hydropower stations and independent optimized operation in terms of output indices， output processes， and water level control processes. The results show that： ① For the average output during the non-flood period， the complementary effect has a significant impact on the minimum output of the cascade hydropower station， while the impact on the average output is minimal. For example， in a typical abundant water year， the minimum output decreases by 76.02% and 60.71%， and the average output decreases by 0.23% and 0.60%， when changing from non-complementary operation to extreme complementary operation； ② Different minimum output values determine the regulation capacity used to drive the complementary output of water and wind， which reflects the degree of complementary operation of water， wind， and solar power. For example， in an abundant water year， the minimum output during the non-flood season varies from 6 844.3 MW to 10 227.6 MW， resulting in the average output for different operating conditions gradually decreasing from 10 389.76 MW to 10 244.06 MW. The higher the degree of complementary operation， the lower the capacity used for optimization to enhance total power generation.

The electromagnetic vibration state reflects the operational health of hydropower units. Developing a monitoring and evaluation model for the electromagnetic vibration state of hydropower units based on the startup transition process is crucial for ensuring the safe and stable operation of the units. This paper proposes a method for extracting electromagnetic vibration features and evaluating the state of hydropower units during the startup transition process， based on principal component analysis （PCA） and cluster analysis theory. Firstly， it constructs two types of electromagnetic vibration state samples by extracting two categories of features： electromagnetic vibration frequency characteristics and the correlation between vibration amplitude and excitation current， which are typical frequency amplitude characteristics and correlation characteristics. Secondly， PCA is introduced for dimensionality reduction， and the principal component statistics are calculated as indicators to reflect the electromagnetic vibration state， thereby constructing the evaluation model for the electromagnetic vibration state during the startup transition process of hydropower units. This model assesses the unit's electromagnetic vibration state by analyzing whether the principal component statistics of the monitoring samples exceed the health threshold and whether there are significantly distinct clusters in the principal component subspace through k-means clustering. The method is validated using measured data from a hydropower unit after an overhaul， demonstrating its effectiveness in tracking and evaluating the trend of changes in the unit's electromagnetic vibration state， identifying potential early safety hazards， and ensuring the safe and stable operation of the unit.

The optimal scheduling of intra-day compensation of cascade hydropower station in the integrated operation of regional power grid is an important issue that needs to be solved urgently. Guangxi power grid has a high installed proportion of new energy such as wind and solar， which puts forward higher requirements for the optimal operation of cascade hydropower stations in the Hongshui River basin that bear the function of compensation regulation. Based on the work process of power grid day-ahead planning and real-time scheduling， this paper constructs three optimal scheduling models， day-ahead， 4 h and 15min， for the purpose of compensating optimal scheduling of cascade hydropower station. According to the wind and solar forecast output process， wind-solar-water daily sequential compensation optimal scheduling research is carried out. It is found that when the wind-solar output is in the 85% confidence interval， there is no risk of load loss， but there is a certain risk of power abandonment in the study area. In the dry season， the potential power abandonment is 441 MWh， and the potential power abandonment rate is 0.92%； in flood season， the potential power abandonment is 11.2 MWh， and the potential power abandonment rate is 0.01%. The research in this paper provides an optimal scheduling method of intra-day and real-time compensation for cascade hydropower station dispatching in new energy grid with high proportion， which plays a positive role in promoting the consumption of new energy and maintaining the safe and stable economic operation of power grid.

Developing reasonable cascade reservoir operation schemas is crucial for enhancing hydropower energy utilization. However， the multi-stage decision-making process increases the difficulty of solving the optimization operation model for cascade reservoirs， and optimization algorithms exhibit a certain degree of operation result instability， low accuracy， or even failure to find the optimal solution when solving. The moth-flame optimization algorithm （MFO） is widely used in many complex problems due to its high efficiency， but it exits defects such as the easy tendency to prematurity and population convergence. To enhance the operation effect of the MFO algorithm， an enhanced moth-flame optimization algorithm （EMFO） is proposed， which adopts the varying number of flames to improve the ergodicity of the optimization space， integrates a light communication strategy that dynamically shares the positions of the top three flames to avoid falling into local extremes and introduces the worst individual opposition-based Learning to improve convergence. The effectiveness of the proposed algorithm is verified by the CEC2022 test function and a case study of the medium- and long-term power generation optimization operation of cascade reservoirs in the upper reaches of the Yellow River. The results show that EMFO is significantly superior to the existing algorithms in terms of optimization accuracy and convergence speed， and in terms of engineering cases， the EMFO can generate operation results with higher convergence accuracy， less water abandonment， and more stable performance under different runoff and initial water level conditions. Its optimized scheme has a higher power generation guarantee rate and the minimum damage depth. Taking the wet year as an example， the power generation increases by 2.50% and 1.56%， the standard deviation decreases by 16.48% and 42.86%， and the abandoned water decreases by 11.375 m3 and 6.839 m3， respectively. The rationality of the operation results is analyzed through the process of water level output and the discharge flow of Liujiaxia during the ice flood season. The optimized operation scheme through the EMFO algorithm can ensure an increase in overall power generation while meeting the flood control needs during the flood season and the ice flood prevention needs of Liujiaxia during the ice flood season， effectively improving the comprehensive utilization rate of water resources， which provides a new feasible method for solving the cascade reservoir operation optimization.

Accurate identification of oil-film binding force of guide bearing is very important for the vibration characteristics of shafting of hydropower units. In this paper， numerical simulations are used to study the variation of the equivalent stiffness coefficient and equivalent damping coefficient of lubricating oil film of the guide bearing of hydropower units with eccentricity， eccentricity velocity and oil film thickness. The results show that there is a great difference between the oil film force calculated by the dynamic computational method and the static computational method. With the increase of eccentricity， direct stiffness coefficient in the X direction first decreases and then increases， while cross stiffness coefficient in this direction continues to decrease. Direct damping coefficient in the X direction first decreases and then increases， and cross-damping coefficient in this direction continues to increase. When the eccentricity velocity increases， the direct damping coefficient in the X direction remains constant， and the cross-damping coefficient decreases in this direction. With the increase of oil film thickness， the direct stiffness coefficient， cross stiffness coefficient， direct damping coefficient and cross damping coefficient in the X direction all gradually decrease. The study further confirms that during the operation of the unit， when the main shaft is eccentric in a certain direction， the damping coefficient in this direction is about 100 times of the stiffness coefficient， which has the effect of inhibiting the movement in this direction， and the stiffness coefficient in the orthogonal direction is about 10⁴ times of the damping coefficient， and the large stiffness coefficient can prevent continuious movement in the orthogonal direction.

In this paper， the study collected and calculated the water level difference between upstream and downstream of the annual peak of more than 40 typical ship locks in China， and obtained the uncertainty parameters and coefficients of variation of hydrostatic pressure. Based on the statistical parameters of resistance under the current code， the modified-JC method was used to calibrate and calculate the reliability index for different components of all levels of the mitre gate， and a method for determining the five partial factors in the ultimate limit state design formula was proposed. The research methodology and theoretical framework can be used as a reference for the research on the structural reliability of similar hydraulic steel gates， provide methodological guidance for the revision of specifications， and help to promote the updating and improvement of specifications.

This paper systematically analyzes the multiple factors influencing the willingness of immigrants in water conservancy projects to resettle and their internal mechanisms using structural equation modeling （SEM） and bootstrap mediation effect analysis， providing a scientific basis for relevant policy formulation. The results show that both natural environment and infrastructure， as well as information access and cognition， have significant positive impacts on economic development， employment opportunities， and resettlement willingness. Economic compensation and policy guarantees， social relations， and cultural adaptation all have significant positive effects on economic development and employment opportunities. Furthermore， economic compensation and policy guarantees， natural environment and infrastructure， information access and cognition， social relations， and cultural adaptation all have significant positive effects on immigrants' resettlement willingness through economic development and employment opportunities. Therefore， enhancing the resettlement willingness of immigrants in water conservancy projects requires joint efforts of the government， society， and individual immigrants. Through improving resettlement conditions， promoting economic development， perfecting policy guarantees， strengthening cultural integration， and other measures， we can jointly promote the smooth progress of resettlement work and achieve a win-win situation for the construction of water conservancy projects and the harmonious development of immigrants.

The excavation under the high side wall of underground plants under medium-high geostress faces significant unloading deformation， and the accurate analysis of the deformation mechanism is crucial for the stability control of the complex rock structure， especially the soft and hard rock bodies. This paper firstly analyzes the unloading deformation difference between soft rock and hard rock based on the theory of unloading model of homogeneous rock body with joint surface， and uses numerical simulation to study the loosening deformation law of soft and hard rock body during transient unloading； relying on the excavation of high side wall layer of Gongyi pumped storage power station underground plant， it extends and analyzes the influence of typical construction parameters on the unloading deformation effect， and elucidates the unloading deformation effect of the dynamic excavation of high side wall. The study shows that： the transient unloading deformation of deep rock body with soft and hard layers differs significantly， and the damage of soft rock is further aggravated compared with that of hard rock by the influence of misalignment friction between layers. Setting reasonable single excavation layering height can better control the transient deformation of side wall. The lower the layering height， the lower the dynamic unloading deformation， which can provide reference for the control of high side wall deformation in the cavern.

For the reuse of failed inclinometer tubes， a flexible inclinometer method with multi-core ultra-weak fiber grating arrays is proposed. Based on the wavelength-deformation inversion algorithm， a theory of fiber optic flexible inclination measurement is constructed， an ultra-weak fiber grating deformation sensing fiber optic cable with a diameter of 6 mm， embedded with 3 cores and a resolution of 0.5 m is designed and directly grouted and buried into the failed inclination measurement tube. With solar power supply， real-time acquisition by low-power demodulation module， and 4G remote transmission， the automated remeasurement of the failed inclinometer holes was realized in the Bazhimen landslide in the Three Gorges Reservoir area. The results show that the multi-core ultra-weak fiber grating flexible inclination measurement system can monitor the obvious displacement of the Bazhimen landslide. The maximum displacement change at a depth of 6 m from the ground during the rainy season in July is 24.4 mm. The measurements of the 3-core optical fiber can back up and corroborate with each other， which can satisfy the requirements for the remeasurement of the failed inclination measurement holes better.

Large deformation or landslide of dam slope seriously threatens the long-term operation safety of the reservoir area. The mainstream traditional slope deformation prediction models fails to fully considered the temporal and spatial characteristics of deformation. This article introduces four representative deep learning methods， namely Transformer， Spatio-Temporal Graph Convolutional Neural Network （STGCN）， Temporal Convolutional Network （TCN）， and Graph Convolutional Neural Network （GCN）， and proposes a spatiotemporal prediction method for slope inclinometer deformation based on deep learning models. The deformation monitoring data of a certain hydropower slope inclinometer hole are used， and the monitoring data are systematically analyzed. The prediction results indicate that all four models， GCN， TCN， STGCN， and Transformer， are suitable for slope spatiotemporal prediction. Among them， the TCN model shows higher prediction accuracy and reliability compared to the other three spatiotemporal prediction models， with evaluation indicators MAE、MSE、RMSE、MAPE and R ² of 1.007、2.208 2、1.486、102.40% and 0.988 4， respectively. In addition， the prediction results of the four models on different dates have an error distribution of 0-4mm compared to the actual measurement， which verifies the accuracy and effectiveness of the four models in predicting the deformation of slope inclinometers in time and space. The research results provide new ideas for short-term spatiotemporal prediction of slope deformation in the reservoir area.

High-speed long-distance landslide is a geological disaster phenomenon with strong destructiveness and wide influence. In order to avoid casualties and economic losses， it is necessary to set up protective structures in landslide-prone areas. At present， researches on debris flow protective structures mostly focus on the blocking ability and improving structural strength， and rarely consider the guiding function of the retaining wall. In this paper， taking Zhaojiagou landslide as an example， the PFC method is used to numerically simulate the guiding control process of the oriented-retaining wall during the movement of debris flow， and the protective effect of the oriented-retaining wall on residential areas is studied. The regulatory effect is also analyzed. The results show that the oriented-retaining wall can effectively prevent the landslide from flowing through the living area. In addition， setting up diverting retaining walls to divide the debris flow into multiple strands and assist in guiding the particle flow can better control the speed of the debris flow， reduce the impact force on the main oriented-retaining wall， and further reduce the possibility of debris flow particles climbing over the retaining wall， thus achieving a better guiding and blocking effect.

The integration of relocated reservoir migrants in their new localities directly affects their sustainable development and social harmony and stability. Exploring the development patterns and influencing mechanisms of migrant integration， and facilitating their swift transition from the special status of “migrants” to full members of the local community， are urgent challenges that need to be addressed to ensure a smooth conclusion of post-resettlement support policies. This paper， based on field tracking survey data from 2008 to 2021 in six relocated migrant villages of the JDX Reservoir in Gansu Province， constructs a comprehensive evaluation index system for the integration of relocated migrants. It uses methods such as Vertical and horizontal differentiation method -Cross entropy combined weighting method， improved rank sum ratio evaluation， Moran's I， and variance decomposition to measure the annual integration levels and analyze their influencing mechanisms. The results show： ① By the 14th year after relocation， the integration levels of all relocated migrant villages have exceeded 80%， indicating great integration； ② The development patterns of integration among resettlement villages share commonalities and exhibit a phased characteristic of “rapid in the early stage， slowing down in the middle stage， and stabilizing in the later stage”； ③ Villages with high integration levels show spatial clustering， and during the early stages of integration， a focus on infrastructure development and promoting livelihood integration significantly enhanced integration. In the later stages， emphasis on guiding the mindset of migrants and promoting psychological integration has the greatest impact on improving integration.

To explore the impact of different bucket foundation structure sizes on the horizontal bearing characteristics of suction pile dams， finite element numerical simulation analysis was conducted in this study. The displacement loading control method was utilized to systematically investigate the influence mechanism of different bucket foundation structure sizes （length L， breadth B， height H） and inflow angle α on the horizontal bearing performance， failure mode， and earth pressure distribution of suction pile dams， while keeping the relevant parameters of the dams consistent. The results indicate that the variation in bucket foundation size has a positive growth trend on horizontal bearing performance. When the length， breadth， and height are increased to 14、7 and 8 m respectively， the ultimate horizontal bearing capacity is maximally enhanced by 14%， 7.3%， and 16.4% compared to the standard condition， primarily due to the increased contact area between the bucket and the soil and the enhanced anti-overturning friction resistance. The loading angle significantly affects the distribution of the plastic deformation zone and the horizontal bearing performance. When the inflow angle is between 45° and 60°， the plastic zone expands the most， and the horizontal bearing capacity reaches its peak. In the 90° direction， the horizontal bearing capacity of the suction pile dam decreases due to the reduced contact area between the cylinder and the soil. The earth pressure curve exhibits an asymmetric characteristic distribution. The soil pressure on the inner wall in front of the dam and the outer wall behind the dam initially increases and then decreases with the burial depth. The soil pressure on the outer wall in front of the dam and the inner wall behind the dam is approximately linearly distributed above the rotation point along the burial depth and increases rapidly below the rotation point. Increasing the breadth B and height H of the bucket foundation significantly increases the soil pressure value in the passive zone， while increasing the length L has a minimal effect on the soil pressure. The correlation between the structural parameters of the bucket foundation and the horizontal bearing performance was quantified in this study， providing a theoretical basis for the optimization design and anti - overturning stability assessment of suction pile dams.