This study evaluates the applicability of four metrics for quantifying the spatial concentration of precipitation，namely the coefficient of variation (CV)，Gini coefficient (G)，Theil index (T)，and Herfindahl-Hirschman index (IHH). The performance of these metrics is examined using controlled numerical experiments and ERA5-Land monthly precipitation datasets at global land areas and the Yangtze River Basin. The results indicate that all four metrics are robust to proportional variations in total precipitation，yet they exhibit distinct responses to dry-wet polarization and precipitation footprint contraction. The T shows higher sensitivity to dry-wet polarization，rendering it suitable for identifying variations in the tails of precipitation distribution. By contrast，the IHH is more responsive to footprint contraction and thus effective in identifying changes in precipitation-concentrating regions. Both the CV and G exhibit relatively stable performance across all conditions. Controlled numerical experiments based on ERA5-Land monthly precipitation fields further suggest that dry-wet polarization can be decomposed into two sub-mechanisms: one-sided enhancement of wet-tails and two-sided dry-wet differentiation. The CV is more sensitive to one-sided wet-tail enhancement，while the T is more responsive to two-sided dry-wet differentiation. Results for the Yangtze River Basin are broadly consistent with those obtained for global land areas. These findings provide a methodological basis for selecting appropriate metrics in precipitation-concentration-related studies.
Aquatic community dynamics in semiarid regulated rivers (i.e.，river systems significantly regulated by human activities) are sensitive to environmental changes; however，the precise mechanisms by which external drivers shape these dynamics remain poorly understood. Focusing on the middle and lower reaches of the Daheihe River (a first-order tributary of the Yellow River)，this study utilizes 36 sets of synchronous “three-waters” (water resources，water environment，and water ecology) monitoring data collected across four cross-sections from 2023 to 2024. By integrating a metacommunity dynamics model (MDM)，cumulative dispersal volume (CDV)，and the relative dispersal potential index (RDPI)，we elucidate the niche characteristics，dispersal patterns and underlying mechanisms of phytoplankton，zooplankton，and benthic macroinvertebrate communities. Several key findings emerged. First，the three communities exhibited relatively broad ecological niches (0.80-1.00) across most environmental gradients，with a high degree of overlap in resource utilization (mean competition coefficient of 0.93). Nevertheless，a certain degree of ecological niche differentiation persisted among the communities. Second，spatial dispersal along the river axis exhibited a pattern of net upstream output and net downstream input，with zooplankton showing the highest dispersal potential (RDPI = 0.44). Third，biotic interactions contributed significantly more to community dynamics (51%) than water quality (27%) and hydrological factors (22%)，while exhibiting pronounced spatial heterogeneity along the river axis. These findings provide a scientific basis for the ecohydrological management of semiarid rivers and the coordinated governance of water resources，water environment，and water ecology.
AI-based forecasting models for urban flooding face the challenge of selecting an appropriate sample size: an excessively large sample size may result in unnecessary computational costs，whereas an insufficient sample size may lead to overfitting. In this study，hydrological and hydrodynamic modeling was integrated with AI techniques，and four algorithms，namely ridge regression，k-nearest neighbors （KNN），random forest （RF），and backpropagation neural network （BPNN），were employed. By identifying the performance change points，saturation points，and optimal sample sizes of the four models，the variation in forecasting performance with increasing sample size was investigated. The results showed that moderately sized datasets were sufficient for the models to achieve high and stable forecasting accuracy. Under the prescribed accuracy criteria （Nash-Sutcliffe efficiency coefficient not less than 0.85 and root mean square error not greater than 0.02），the optimal sample sizes for ridge regression，KNN，RF，and BPNN were 100，90，50，and 120，respectively，accounting for 59.5%，53.6%，29.8%，and 71.4% of the full dataset. Compared with the use of the full dataset，the dataset construction time at the optimal sample sizes was reduced by 5.00，5.71，8.61，and 3.39 h，respectively. Among the four models，RF exhibited the greatest potential for sample-size optimization. These results provide a basis for balancing dataset construction costs and forecasting accuracy in AI-based urban flood forecasting.
Modern water network operation requires both the translation of scheduling tasks into executable engineering operations and the linking of engineering effects，task outcomes，and deviation information back to the original task to support subsequent judgments. This study defines this continuous relationship between a scheduling task and engineering execution as a task-level control relationship. Starting from an approved task，the relationship is maintained or adjusted as states，constraints，and authorization change，and is ultimately closed through feedback. To address inconsistent object definitions，incomplete measurement-control links，and discontinuities in responsibility handoffs，four theoretical propositions are developed: object correspondence，relationship establishment，state continuity，and organizational formation. The node-channel structure provides a unified object representation for tasks and engineering objects. The task-level measurement-control coordination criterion determines，across four dimensions-objects，boundaries，authorization，and responsibility-whether the relationship can be established，maintained，and closed. Five state categories and three operational domains are used to identify changes in operating conditions and determine when the original control relationship should be maintained，adjusted，or re-established. Basic control units and their multiscale organization enable multiple engineering units to work in coordination on the same task. On this basis，four methods are proposed: state representation，basic measurement and control，coordinated control，and capability enhancement. Within an overall decision-making-scheduling-control-execution architecture，the study further explains how tasks are transformed，how feedback is returned，how control organization is connected across scales，how three types of boundaries are mapped，and how layered verification is conducted during system construction. The study shows that whether a task-level control relationship can be established，maintained，and closed depends on clear correspondence between tasks and engineering objects; complete measurement，control，and feedback links; the continued validity of states，constraints，authorization，and responsibility during execution; and coordination among multiple engineering units working on the same task.
To clarify the evolution and climatic drivers of flash drought events in the Weihe River Basin，flash droughts were identified using a three-dimensional clustering method based on ERA5-Land soil moisture and meteorological data. Spatiotemporal characteristics were systematically analyzed，and the dominant climatic drivers were examined using principal component analysis. Results showed the following: ① From 2000 to 2024，68 flash drought events were identified，with an average frequency of approximately 2.72 events per year and an average duration of 36.5 d per event. Events in the Southern Jinghe River Basin lasted longer （50- < 90 d） and were more likely to develop into large-scale，high-intensity events. ② Flash droughts mainly migrated toward the west-northwest，east-southeast，and east-northeast （12，10，and 9 events，respectively）. Their migration distance，affected area，and severity were greatest toward the south-southwest. ③ The thermal-atmospheric evaporative demand factor and water supply factor were the first and second principal components，PC1 and PC2，influencing flash droughts，explaining 63.2% and 16.4% of the total variance，respectively. During flash drought periods，PC1 scores increased significantly （+6.95），whereas PC2 scores decreased markedly （− 1.52）. These findings indicate that flash drought events in the Weihe River Basin develop primarily under enhanced thermal-atmospheric evaporative demand，compounded by water supply deficits. Overall，the results provide a scientific basis for flash drought monitoring，early warning，and risk management in the Weihe River Basin.
