Addressing the issue of insufficient flood forecasting accuracy during the Meiyu period (June 18-28) in the Fenshuijiang River Basin, this paper investigates the practical application and improvement method of coupled "Xin'anjiang + Muskingum" hydrological model. Based on the observed rainfall and reservoir operation data, the evaluation of forecasting precision for peak discharge, runoff depth, and time-to-peak was conducted through relative error analysis. The accuracy of the model was verified by comparing the measured values with the predicted values in combination with the deterministic coefficient (D_DC). The forecast effect was optimized by correcting the runoff yield and concentration parameters of the Xin 'anjiang model and the flow routing parameters of the Muskingum model. The results show that after parameter adjustment, the peak discharge relative error is below 15%, runoff depth error is low 40%, time-to-peak deviation is within 2 hours, and D_DC values exceeds 0.85; However, there is still a lag in the forecast during the concentrated period of heavy rainfall. The correction coupled model significantly enhances flood forecasting accuracy during the Meiyu season, offering reliable support for reservoir operation decisions. In the future, it is necessary to further optimize the dynamic coupling mechanism between the input of rainfall spatio-temporal distribution and the parameters of river routing parameters.

In recent years, frequent heavy rains have made Zhengzhou city suffer from the hazards of urban waterlogging. Identifying the spatiotemporal distribution characteristics and driving forces of waterlogging can help improve the current flood-prone situation in Zhengzhou city. Based on the waterlogging data of Zhengzhou from 2020 to 2024, the standard deviation ellipse method, autocorrelation method, kernel density estimation method, and optimal parameter geographic detector were used to systematically analyze the spatiotemporal distribution characteristics of waterlogging points and their driving forces from multiple scales. The results show that the annual distribution of waterlogging points in Zhengzhou is uneven; Spatially, the the focus of waterlogging is located in the old urban area, with significant clustering of waterlogging points at different scales, especially with the strongest spatial clustering within a 3km scale; There is a significant scale effect in the explanatory power of driving factors, with socioeconomic and urban construction driving factors having the greatest explanatory power at different scales, and the combined effect of factors significantly surpasses the independent effects of single factors.

The video image flow monitoring technology has been applied in hydrological monitoring and irrigation area monitoring. Due to the limitation of received light, video image flow monitoring method has limitations in monitoring at conditions of night and low light. The thermal imaging was used to increase the spectral applicability of video flow measurement and compensate for the low applicability of video flow measurement under night conditions. By using a visible light and thermal imaging double-sided camera device mounted on a drone, the flow velocity field inversion results of visible light and thermal imaging camera devices under conditions of visible light and night were compared with the flow velocity monitoring data. The results show that under the condition of sufficient visible light during the day, the inversion effect of visible light video image flow velocity is better, and the accuracy error can reach within -5% except for a few monitoring points; Due to the limitation of its resolution, thermal imaging camera has a certain ability to capture natural tracers on the river surface, and has the ability and potential to invert the flow velocity field on the river surface. If missing particles of cold ice are added to the river section, the monitoring accuracy of thermal imaging at night can be significantly improved. Except for one monitoring point, the relative error of monitoring accuracy is 8.4%, and the accuracy of the other three monitoring points is relatively high, all around -1%. The research results can provide important references for improving the monitoring accuracy of video velocity measurement under conditions of nighttime and low light.

To enhance the application value of precipitation observation data in public meteorological services, this study investigates a quality control method for precipitation data based on multi-source meteorological data and machine learning models. Based on the meteorological station data, multi-source integrated analysis products, meteorological radar, and satellite products, as well as the XGBoost model, a multi-source collaborative quality control method for precipitation data was developd by comprehensive scoring. The accuracy of this method was evaluated using the artificial error implantation technique. The results indicate that when random precipitation anomalies are implanted, the detection rate of erroneous data is 0.85; When errors are implanted at different precipitation levels, the detection rate of erroneous data is 0.79. After collaborative quality control, the average error between the precipitation data to be tested and neighboring meteorological stations is reduced by 15%, and the correlation coefficient is increased to 0.7-1.0. Therefore, this method can effectively identify suspicious precipitation data, which can provide strong data support for meteorological disaster prevention and mitigation.

To analyze the daily-scale spatiotemporal process of the persistent extreme drought in Sichuan and Chongqing in 2022, the daily-scale drought index was calculated based on the ERA5-Land meteorological data from 2018 to 2022. Three-threshold tour theory was used to identify persistent extreme drought events in the Sichuan-Chongqing Region in 2022. The duration, intensity, and frequency of the drought across different terrain types were extracted, and the spatial heterogeneity of drought processes was revealed. The results show that the majority of pixels in the Sichuan-Chongqing region were subjected to two drought events in 2022 and the worst of the drought occurred on 22 August. The average duration of persistent extreme drought in Sichuan-Chongqing region during the summer was 67 days, with the longest duration being 159 days and the shortest being 17 days, showing a increasing trend from west to east. The earliest drought started on 16 June and the latest on 28 August. The correlation between drought severity and drought duration reached 0.98, and their trends were consistent. In different terrain areas, the first type of drought process occurred in spring, with an average of 3.5 drought events, an average total duration of 166 days and an average severity of 193.63. The second type of drought with no drought in spring occurred on average 2.2 times, with an average duration of 120.6 days and an average severity of 136.27, which is only 70.38% of the first type. The research results can provide theoretical basis and technical support for regional daily drought monitoring.

Drought disaster is the combined effects of disaster-causing, socioeconomic and environmental factors. The textual information of historical drought disasters in China was used to classify drought disaster. The method of hotspot analysis was employed to identify prone areas of drought disaster. Based on three elements of hazard, exposure, and vulnerability, the risk assessment of drought disaster was investigated. The results show that three types of key information including water and rainfall conditions, drought severity and social impact extracted from drought disaster texts were well applied to their classification. During the period of 1949-2020 in China, it experienced an average of about 170 drought disasters per year, with mild, moderate, and severe droughts accounting for 52.4%, 27.6%, and 20.0%, respectively. The prone areas of drought disaster mainly occurred in Henan, Shanxi, Shaanxi and Inner Mongolia etc. The levels of the hazard, exposure and vulnerability in Henan, Hubei and Shandong accounted for the higher or above. The levels of drought disaster risk in Henan, Hubei, Sichuan, and Shandong, as the major grain producing areas, were the higher or above. The results of the study can provide a reference for drought disaster mitigation.

Agricultural drought risk assessment is the prerequisite for agricultural drought risk management. Crop growth is highly vulnerable to threats from natural disasters such as droughts, floods, and typhoons, with drought-induced losses accounting for over 50% of the total agricultural natural disaster losses. Based on the regional natural disaster system theory, this study constructs an agricultural drought disaster risk assessment indicator system comprising three components of hazard, vulnerability, and exposure. A total of 22 indicators including annual precipitation were selected. The study adopts county-level administrative regions across China as research units and integrates the fuzzy comprehensive evaluation method to establish the agricultural drought disaster risk assessment framework. The results indicate that the spatial distribution of drought hazard is significantly correlated with annual precipitation, showing a trend of higher hazard in the northwest and lower in the southeast. Provinces with the highest vulnerability include Heilongjiang, Guizhou, and Ningxia, while Guangdong and Zhejiang exhibit the lowest vulnerability due to their strong economic resilience and low disaster impact indices. Henan, Hebei, Heilongjiang, and Ningxia demonstrate higher comprehensive drought risk ratings owing to high agricultural exposure levels. In contrast, regions such as Qinghai, Tibet, Fujian, and Zhejiang exhibit the lowest risk levels due to smaller agricultural scales or abundant water resources. This research can provide a basis for targeted agricultural disaster prevention and mitigation strategies.

To investigate the three-dimensional spatiotemporal propagation characteristics of meteorological drought, based on the CN05.1 gridded data and standardized precipitation index (SPI) during the period 1961-2021 in China, a spatiotemporally continuous event identification method and hierarchical clustering algorithms were applied to classify different propagation patterns of spatiotemporally continuous meteorological droughts and analyze their characteristics. The changing characteristics of events in future climate scenarios were discussed. The results indicate that the spatiotemporal continuous meteorological drought events distribute widely with significant variations in their movement distances, ranging from localized areas to nearly half of China. Three propagation patterns are identified based on movement distance and direction. The Type 1 and Type 2 events are characterized by long duration and large affected areas but occurring less frequently. The Type 3 events occur more frequently, have shorter durations, and affect smaller areas. Under future scenarios, the total number of spatiotemporally continuous drought events is projected to increase, while their duration and intensity are expected to decrease. Both the maximum and total affected areas are likely to expand.

Scientific classification and evaluation of groundwater level funnels is an important basis for the treatment and protection of groundwater overexploitation. Based on the actual situation of groundwater dynamic analysis, evaluation and management in China, this paper put forward the basic provisions, evaluation indexes, classification standards and evaluation methods of influencing factors for the division of deep and shallow funnels. At the national level, the classification and evaluation criteria of groundwater level funnels have been unified. Taking the Beijing-Tianjin-Hebei region as the research object, this paper systematically analyzed the impact of the comprehensive control of overmining in North China since 2019 on the evolution of funnel. The results show that compared with 2018, the shallow funnel presented a stable type in 2023 , while the deep funnel presented a reduced type. The area of the deep funnel was reduced by about 99%, and the water level in the center of each funnel was obviously rising. Based on the XGBoost integrated learning model, the development and change of the shallow funnel in the Beijing-Tianjin-Hebei region were mainly influenced by human factors, followed by meteorological factors. Among them, groundwater exploitation, accumulated precipitation and ecological water recharge of rivers and lakes have the most significant impacts, with impact contributions of 23.1%, 23.0% and 15.2%, respectively.

To address the shortcomings of existing water safety evaluation methods in terms of scientific indicator selection and accurate weight allocation, this study focuses on the optimal selection of indicators and optimization of weight assignment, starting from five aspects of flood control, drinking water, water use, river and lake ecology, and socio-economic security. The key indicators are selected through correlation analysis, principal component analysis, and information sensitivity analysis. The combined weight method, integrating entropy weight method, analytic hierarchy process, projection pursuit model, and cloud model, is used to solve the indicator weights, which are then optimized by simulated annealing. Finally, the cloud model is employed for comprehensive evaluation, with an application case in Chenzhou City, Hunan Province. The results show that the indicator optimization identified 22 key indicators, including industrial recycled water rate and per capita water resource availability; The correlation between the weights obtained by the simulated annealing algorithm and indicator information sensitivity is the highest (R²=0.744 2), indicating a more reasonable weight distribution; The water safety status of Chenzhou City from 2016 to 2022 was generally basically safe, with a continuous improvement trend. The results validate the effectiveness and applicability of the proposed model, which provides new methodological support and practical guidance for urban water safety evaluation.