Evapotranspiration is a key process of energy exchange in the terrestrial-atmospheric system, and the selection of a suitable method for its accurate estimation is the basis and challenges to revealing the energy exchange process and the mechanism of Land-Atmosphere coupling under complex topographic conditions. Based on 60 years of meteorological data from 25 stations in the Red River Basin, China, the FAO56-Penman-Monteith method was used as a benchmark, and the MK statistical tests and Taylor diagrams were used to evaluate 8 potential evapotranspiration (PET) estimation methods from five aspeacts. The results show that the radiative methods Priestley-Taylor and Hargreaves had the best applicability, their results were closest to FAO (R²≈0.9); The temperature methods McCloud and Linacre were followed; The mass transport class methods Rohwer and Penman were worse. This study not only provided a highly adaptive PET estimation method for the Red River Basin, China, but also revealed the spatial and temporal distribution characteristics of the PET.

Affected by the changes of East Asian monsoon and human activities, the hydrological processes in the southeast coastal area have been changed, and the rational development and utilization of regional water resources are facing challenges. Taking Feiyunjiang as an example, the variable parameter Budyko equation based on the moving average treatment was constructed to analyze the runoff change in the basin, and the future runoff change trend was predicted qualitatively combined with the CMIP5 model. The results show that the time-varying Budyko model is much better than the constant parameter Budyko model in the simulation of annual runoff. The hydrometeorological elements in Feiyunjiang Basin showed an upward trend after eliminating the influence of cycle. The contribution rate of climate to runoff variation in the Feiyunjiang Basin reached 85.56%, and showed a downward trend. Compared with before 2000, the contribution rate of the underlying surface to runoff change increased by 6% from 2000 to 2020. At the same time, the influence intensity of climate on runoff change changed sharply. The influence of the underlying surface on runoff reduction was weakened. In addition, under the RCP4.5 scenario, the runoff in the Feiyunjiang Basin will show a decreasing trend in the future. The research results can provide theoretical support for regional water resources planning.

With the continuous development of the city, influenced by climate change and human activities, the runoff of urban water system has changed greatly. In order to further research the runoff trend and its main influencing factors brought by urbanization, the application of the comparison of slope change ratio in hydrological change analysis was further innovated, and the contribution rate of underlying surface to runoff change was separated from human activities. Taking Zhengzhou city as an example, combined with Mann-Kendall test method, the contribution rate of climate change factors, underlying surface factors and other human activities to runoff change was calculated. The results indicate that the measured runoff flow showed an upward trend from 1980 to 2019, and a sudden change occurred in 2008. In the impact analysis of runoff, the change of underlying surface and other human activities are the main influencing factors, and their contribution rates are 69.76% and 56.19%. According to the analysis of the main influencing factors, the main reasons are as follows: The hardening of the underlying surface increases runoff yield and flow concentration, and the population increases and the emissions increase.

Based on the runoff data of Buha River Basin and Shaliu River Basin in Qinghai Lake Basin, the annual base-flow index was obtained by using the filter smoothing minimum method. The temporal and spatial variation of precipitation and the evolution law of base flow in this basin were analyzed by TFPW-MK method, and the influencing factors of base flow were analyzed. The results show that the base-flow index of Buha River Basin first increases, then decreases and increases, reaching the maximum in May and the minimum in August; The base flow index of Shaliu River Basin fluctuates and reaches the minimum in April. The annual average runoff in the Buha River Basin and the Shaliu River Basin showed a significant increasing trend, and both the annual average base flow and base flow index showed an insignificant increasing trend; The base flow inclination rate of the Buha River Basin is twice that of the Shaliu River Basin. The base flow sequence of the Buha River Basin and the Shaliu River Basin is greatly affected by precipitation, flow, drought index and evaporation, and the Shaliu River Basin is more significantly affected by evaporation.

With the development of deep learning and image recognition technology, monitoring the water level of urban rivers and lakes through video has become a hot research topic in recent years. In order to realize the comprehensiveness of urban river and lake water level monitoring, a method of river and lake water level identification based on Mask RCNN was proposed. The water level was obtained by the water level characters and their position relations in the video images, and it was verified by the monitoring video of the real water level station in Dongying City, Shandong Province. The results show that the probability that the comparison error between the water level identification result and the measured data was less than 2 cm was 68.5%, the probability of error less than 3 cm is 76.9%, the probability of error less than 5 cm is 93.5%, the average error is 2.1 cm, and the root mean square error (RMSE) is 3.0 cm, which meets the recognition accuracy requirements of video water gauge level in Technical Outline of Digital Twin Watershed Construction (Trial). Therefore, the model has a good recognition effect and a certain application prospect.

The lakes on Tibetan Plateau are expanding continuously, and the frequency and intensity of natural disasters (debris flow, glacial lake outburst) are increasing, which brings a series of challenges to water resources management, water disaster prevention and sustainable development in Tibetan Plateau and its downstream areas. Taking Yigong Lake as the research object, this paper establishes Tableland NDWI (TNDWI) based on 1 101 remote sensing images. Then the water surface morphology of Yigong Lake from 1986 to 2020 is reconstructed to analyze evalution process of historical water surface morphology in lake. The results show that the user accuracy of the TNDWI for the classification of plateau water bodies is higher than 96.70%, the overall accuracy is higher than 98.75%, and the Kappa coefficient is between 0.972 and 0.992, and the classification accuracy and consistency are better than the tranditional NDWI; The 2000 Yigong landslide has a significant impact on the evolution process of water surface morphology of Yigong Lake; Water surface area after dam-break reduces from 14.44±2.70 km² to 10.99±1.94 km²(P <0.001). Mann-Kendall mutation test results show that the change trend of water surface morphology has a mutation in 2000 (except in dry season); Morlet wavelet periodic analysis results show that the periodic oscillation law of lake area has gradually weakened since 2000. This study overcomes the lack of monitoring data of high altitude barrier lake, providing data basis for the ecological restoration and comprehensive treatment project of Yigong Lake as well as new idea for the quantitative study of the ecological environment in the region.

It has great guiding significance of studying the hydrodynamic distribution strong tide for Shenzhen River estuary management. The physical model of Shenzhen River was established to analyze the hydrodynamic distribution characteristics of beach and trough and the neap tide. The spring tide in dry season and typhoon "Hato" were selected as the boundary conditions in tidal optimal estuary. The results show that the tidal power of the main trough of the Shenzhen Estuary is obviously better than that of the middle beach, the flow velocity tends to increase from the downstream to Shenzhen River for the discharge section is narrower in upstream during rising tidal, the power of spring tide and ebb tide in dry season is basically the same, the power of spring tide in dry season is stronger than that of ebb tide, the estuary flow pattern is mostly affected by riverbed morphology during the low water level with a large area of open beach which made the south trough flow only. During the storm surge, the hydrodynamic force of the estuary is strong, the flow velocity of the middle beach is significantly greater than that of the dry season, the middle beach water is not flowing back to the trough for the high water level and the mangroves in the middle beach has a great influence on the flow pattern. According to the hydrodynamic characteristics and flood discharge requirements, the countermeasures for ecological beach consolidation and main channel widening in Shenzhen Estuary were put forward.

To investigate the spatial and temporal variation of terrestrial water storage in the Yellow River Basin, the inversion of terrestrial water storage in the Yellow River Basin was performed based on the Gravity Recovery and Climate Experiment Follow-On (GRACE-FO) gravity satellite data from 2003 to 2020. Besides, the Variable Infiltration Capacity (VIC) hydrological model was used to simulate the soil water changes. The groundwater storage changes between 2003 and 2014 were calculated through the water balance equation. The results show that the overall terrestrial water storage in the Yellow River Basin is in deficit during the study period, with a significant decreasing trend temporally. Spatially, the water storage is reducing from the river source to the downstream. The inversion results show that the spatial and temporal variation of terrestrial water storage in the Yellow River basin is highly correlated with the topography, climate and human activities. The results can provide a scientific basis for water resources planning and allocation in the Yellow River Basin.

The dynamic changes of the water ecological footprint and carrying capacity of the representative cities along the first Phase Project of South-to-North Water Diversion East Route Project from 2010 to 2020 were analyzed based on the water ecological footprint and carrying capacity model. The SVR model was used to predict the water ecological condition of these cities from 2021 to 2025. The results show that the per capita water ecological footprint of Shandong displays a general uptrend, and the per capita water ecological footprint of the south section of Luoma Lake and Jiaodong Peninsula is at a high level. The per capita water ecological carrying capacity in the north of Luoma Lake and Jiaodong Peninsula shows an upward trend, while in the south of Luoma Lake and across the Yellow River, it shows a downward trend, among which the south of Luoma Lake is at a high level. The water ecological deficit or surplus of the representative cities along the line began to improve in 2013, indicating that the implementation of the project has a positive effect on the improvement of ecological environment. From 2021 to 2025, the water ecological footprint of Jiangsu section will increase and the water ecological carrying capacity will decrease. With the continuous increase of water consumption in various industries, the sustainable development of this region will face a great threat. Although the water ecological situation in Shandong section is developing well, it will continue to be in a state of water ecological deficit for a period of time, and the situation of water ecological pressure is severe.

In order to evaluate the water use efficiency of the central Yunnan urban agglomeration accurately, 14 indicators were selected from five aspects of comprehensive, industrial, agricultural, living and ecological water use, and the evaluation index system was constructed. The subjective weights were obtained by analytic hierarchy process. The entropy weight method was used to determine objective weights. The game theory method was adopted to get the combination weight. The TOPSIS evaluation model was established to calculate the relative closeness degree of water use efficiency and the closeness degree of each criterion layer in central Yunnan urban agglomeration from 2011 to 2020, and the national average data was added to the evaluation sample as the criterion for judging the results. The results show that in terms of spatial distribution, Kunming has the highest water use efficiency, followed by Qujing and Yuxi, both of which are at a good level, and Honghe and Chuxiong have the lowest water use efficiency, both of which are lower than the national average water use efficiency, at a poor level; In terms of time series, except for Qujing City, the overall water use efficiency shows an upward trend. Qujing City shows a downward trend before 2016 and slowly recovers in 2016. By analyzing the causes of the changes, this paper put forward the shortcomings of water use efficiency in each city, which can provide reference for improving water use structure and water use efficiency in central Yunnan urban agglomeration.