Globally, overuse of groundwater resources and their scarcity have emerged as significant obstacles to sustainable development. Groundwater potential maps are critical tools for long-term water resource management, particularly in agriculturally intensive countries such as China. The past decade has seen the adaptive neuro-fuzzy inference system (ANFIS) garner enthusiastic acceptance across diverse research domains. This study aims to evaluate groundwater potential in Mianyang City, China, by leveraging ANFIS within a Geographic Information System (GIS) framework, coupled with two metaheuristic optimizers: the Dandelion optimization algorithm (DOA) and the Nutcracker optimization algorithm (NOA). The groundwater potential was modeled using 14 influencing factors and 326 spring locations. Data for modeling was separated into two main groups: training (70%; 228) and validation (30%; 98). The CAE (correlation attribute evaluation) method was employed to determine the relative contribution of each factor. The results show that rainfall, lithology, and elevation are the primary factors influencing groundwater potential. Three models' performance was tested and compared using the area under the receiver operating characteristic (ROC) curve (AUC), RMSE (root mean square error), and several quantitative indicators. The ensemble ANFIS-DOA model (positive predictive value = 94.4%, negative predictive value = 89.7%, sensitivity = 90.1%, specificity = 94.2%, accuracy = 92.0%, j = 0.859, AUC = 0.939, and RMSE = 0.072) performed the best in mapping groundwater potential. The ANFIS-NOA and single ANFIS models came in second and third, respectively. Hotspot analysis further revealed three statistically significant high-potential zones concentrated in the central and eastern parts of the study area, confirming the spatial robustness of the model outputs. The groundwater potential map produced in this study can inform strategic planning by identifying suitable locations for key infrastructure, such as irrigation systems, industrial facilities, and residential zones. (c) 2025 COSPAR. Published by Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

Landslide susceptibility assessment is essential for maintaining resilient mountain transport corridors where slope failures can disrupt mobility, freight movement, and economic activity. This study evaluated standalone and hybrid machine-learning models for the Kamchik Pass corridor of Uzbekistan, which carries the A-373 Tashkent–Osh highway. Thirteen topographic, hydrological, geological, climatic, vegetation, and land-cover factors were integrated with a landslide inventory. K-nearest neighbours (KNN), Random Forest (RF), XGBoost, and artificial neural network (ANN) models were compared to three RF-based hybrids: RF + KNN, RF + XGBoost, and RF + ANN. Performance was assessed using confusion-matrix metrics and receiver operating characteristic area under the curve (ROC–AUC), together with variable-importance and class-area analyses. All predictors were retained because variance inflation factors remained below 3.3. Slope and elevation were the most consistent predictors across the standalone models. RF + KNN achieved the best performance, with an accuracy of 0.8429, kappa of 0.6857, sensitivity of 0.8857, specificity of 0.8000, and AUC of 0.88. Its map classified 16.94 km2, approximately 10.8% of the study area, as high or very high susceptibility. Compared to standalone RF, RF + KNN increased accuracy by 2.86 percentage points and AUC by 0.03, while sensitivity decreased from 0.9429 to 0.8857 and specificity increased from 0.6857 to 0.8000. Because these differences were obtained from a small point-level hold-out set without spatially independent validation, they are interpreted as descriptive rather than evidence of universal model superiority. The maps provide a first-pass susceptibility screening layer for subsequent field verification and asset-exposure analysis; they do not constitute an implemented infrastructure-risk assessment.

Weathered rock slope damage occurs frequently in mountainous areas, but the effects of weathered layers on the dynamic response characteristics and failure patterns of rock slopes are rarely systematically observed. Taking the Laoshaiba landslide as an example, a time‒frequency joint analysis method based on time-, frequency- and time‒frequency-domain analysis is proposed in this work. Three models, a homogeneous slope (Model 1), a weakly weathered slope (Model 2) and a strongly weathered slope (Model 3), were established via FLAC3D for dynamic analysis. The time‒frequency joint analysis revealed that the weathered layer intensified the propagation of dynamic waves and amplified the dynamic response of the slopes. The amplification effect of the strongly weathered slope is greater than that of the weakly weathered slope. According to field investigations, modal analysis and Fourier spectrum analysis, the presence of a weathered layer affects the dynamic failure pattern of slopes. The weathered layer reduces the dynamic stability of the slopes, whereas it significantly affects the peak Fourier spectrum amplitude and the higher-order natural frequencies of the slopes. Unlike the seismic failure pattern of cracking‒sliding on top of a homogeneous slope, weathered slopes exhibit a surface crushing‒sliding failure pattern. Furthermore, empirical mode decomposition (EMD) coupled with composite multiscale permutation entropy (CMPE) and the wavelet packet denoising (WPD) method is proposed, which reduces the endpoint effect and mode aliasing that appear in the EMD algorithm and provides better processing of seismic signals. This study provides valuable insights into the seismic response of weathered rock slopes.

Existing "seepage-stress-damage" models are mainly used for pre-drilling prediction and post-drilling evaluation of collapse pressure. Such models involve a large number of parameters that are difficult to obtain. In this paper, the "seepage-stress-damage" model is simplified into a strength softening model applicable to real-time collapse analysis while drilling. A method is also proposed to determine the parameters of the strength softening model using logging data. Field engineers of Well X adjusted drilling fluid properties based on the strength softening model. Within 10 days of exposure, only small collapse blocks (0.2 cm × 0.23 cm) were retrieved, and the 1000 m thick clay formation in Well X was drilled efficiently.

To study the influence of the valley topography effect and the geological effect of structural planes on the dynamic response characteristics and failure mechanisms of a steep rock slope, a time-frequency joint analysis method combining time, frequency and time-frequency domains analysis is proposed. Four models were created for dynamic analysis using FLAC3D. According to the time-frequency joint analysis, the dynamic response law of homogeneous high-steep slopes is that the acceleration amplification factor (AAF) varies periodically along the elevation with a certain pattern and is obviously amplified at the slope crest. The existence of structural planes reduces the velocity of wave propagation through rock masses while creating local amplification and attenuation of seismic waves. The valley topography has an amplifying effect on the dynamic response of the slopes but has less influence on the changing laws. Frequency domain analysis shows that the low-frequency and high-frequency components mainly cause overall deformation and local deformation of the slopes, respectively. Moreover, structural planes have a high-frequency filtering effect, and the influence of structural planes on the failure mechanism of slopes is significantly greater than that of valley topography. In addition, the relationship between the local deformation response of anti-dip slopes and the occurrence of landslides is determined. In addition, the dynamic response characteristics of the slope were analysed based on the Hilbert energy, and the ability of the seismic Hilbert energy spectrum to reflect the dynamic deformation characteristics of the slope was identified.

Hydropower development is one way to efficiently harness the energy of rivers. However, damming limits the upstream reproductive behavior of migratory fish, and fish passage facilities are now an important means of restoring connectivity. Flow velocity is an important factor for fish to perceive the environment and maintain swimming behavior. The reservoir in front of the dam has become a near-hydrostatic environment that significantly changes the river properties. Reduced velocity may make it difficult for fish released into the reservoir area by fish passage facilities to find the flow direction, reducing their migratory success rate. This paper takes the Wudongde reservoir as an example and uses numerical simulation to analyze the characteristics of the flow field in the reservoir at different flow levels. It takes the induced velocity of fish as the threshold value to analyze the river sections in the reservoir affected by the reduced velocity. The results show that the flow velocity in the reservoir area decreases significantly from the end of the reservoir to the front of the dam; the flow velocity in the main reservoir is more homogeneous in the direction of water depth, and the tributary estuaries present non-uniform distribution. Under the conditions of low flow (February), medium flow (May), and high flow (August), a certain range of the river section is below the induced velocity of the fish (0.2 m/s), and the suitable locations for migratory fish stocking are located in the upper part of Wudongde Dam, 102, 96, and 38 km, respectively. The results provide technical references for the selection of release sites for fish passage facilities in high dams and large reservoirs, and are of great significance for the protection of water ecology in the dam-built river reaches.

OBJECTIVE:To explore if the nasal profile is influenced by maxillary hypoplasia among patients with congenital cleft lip and palate. METHODS:Young adult patients with cleft and noncleft controls were enrolled. Nasal and maxillary profiles were measured on lateral cephalometric radiographs. The Kruskal-Wallis and Tukey post-hoc tests were employed for intergroup comparison among various the cleft types, and Pearson's product moment correlation coefficients were calculated to detect the correlation between nasal and maxillary cephalometric measurements. RESULTS:A total of 250 participants were enrolled, including 64 with unilateral cleft lip and alveolae; 22 with bilateral cleft lip and alveolae; 49 with unilateral cleft lip, alveolae, and palate; 35 with bilateral cleft lip, alveolae, and palate; and 80 controls. Patients with cleft demonstrated significant difference in nasal and maxillary profiles when compared with the normal controls. The nasal dorsum length had a significant positive correlation with anterior and posterior maxillary height. The upper nasal dorsum length had a significant negative correlation with the maxillary protrusion. Among patients with bilateral cleft, the nasal dorsum length was also significantly negatively correlated with the maxillary protrusion. CONCLUSION:The nasal profile is closely related to the growth of maxillae among patients with cleft and the normal population.

Frequent earthquakes, coupled with discontinuous geological conditions, complicate the earthquake dynamic response characteristics of slopes. To reveal the dynamic accumulative damage effect and instability mechanism of a cross-jointed rock mass slope under continuous earthquakes, two discrete-element models, a homogeneous slope and a cross-jointed slope, are established in this study via particle flow code (PFC2D). The results show that the cross-jointed slope has an obvious dynamic amplifying effect on the elevation and slope surface. Compared with the homogeneous slope, the cross-jointed slope has a more significant slope magnification effect. The seismic wave amplification on cross-jointed slopes is more significant under the influence of S waves than P waves. In addition, on the basis of the characteristics of crack propagation and particle bonding failure, the dynamic cumulative failure effect and evolution process of cross-jointed slopes under cumulative earthquakes are revealed, including the crack initiation stage (<= 0.1g), crack accumulation stage (0.1-0.2g), crack propagation stage (0.2-0.4g), and crack penetration stage (0.4-0.6g). Through the analysis of particle bond rupture characteristics and displacement evolution, the mechanisms of dynamic failure and modes of instability for the cross-jointed slope are identified. Under a small earthquake (<= 0.2g), the slope remains largely uncracked, and slope failure does not occur. Under a strong earthquake (>= 0.4g), bond failure near the slope surface develops rapidly and intensively, and the sliding body gradually experiences instability failure. Moreover, joints control slope dynamic failure behavior. Shear failure and tensile failure mainly occur in cross-jointed slopes and homogeneous slopes, respectively.

This study employs the discrete element method (DEM) to simulate the dynamic processes of landslides induced by rainfall and earthquakes, with the objective of conducting a comprehensive investigation into the influencing factors, failure mechanisms, energy transformation, and movement characteristics of landslides. The methodology employed involved the construction of a two-dimensional landslide model, which was utilized for simulating the instability process of the Temi landslide under conditions of rainfall and seismic activity. The primary conclusions of this study are as follows: (1) While both rainfall and seismic events can lead to landslide instability, earthquakes are the predominant trigger for river-blocking landslides. (2) The Temi landslide event and subsequent damming of the Jinsha River occurred through five stages: seismic-induced rock mass loosening; rear tension cracking and front shear failure; high-speed sliding; front portion crossing the riverbed and climbing the opposite bank; and finally debris accumulation forming the dam. (3) Energy transformation and dissipation are pivotal in landslide dynamics, with frictional heat being the primary energy dissipation source. (4) The evolution of kinetic energy demonstrates initial stability, a gradual increase during sliding, and an eventual decrease to zero upon cessation. The findings of this study indicate that earthquakes are relatively likely to trigger large river-blocking landslides, and the discrete element method has significant advantages in simulating complex geological phenomena, providing important insights for landslide prediction and disaster prevention in earthquake-prone areas.

Under the dual pressures of global climate change and human activities, the carbon sink function of terrestrial ecosystems and ecological vulnerability exhibit a dynamic game relationship. To address the inadequacies in carbon dynamic assessment methodologies and the unclear driving mechanisms in ecologically fragile regions (particularly karst desertification areas), this study developed an integrated framework of "remote sensing monitoring-model coupling-scenario prediction". Focusing on Guangxi's karst desertification region, we analyzed the spatiotemporal evolution of carbon sinks and human-land coupling driving mechanisms from 2000 to 2020, while projecting future trajectories of carbon storage. The research findings indicate that (1) Rocky desertification control measures have significantly reduced the area of severely degraded land, with a 74.5 % decrease in extremely severe rocky desertification. However, latent risks persist, forming an "explicit improvement-implicit risk" coexistence. (2) Vegetation net primary productivity (NPP) exhibited a fluctuating upward trend (2000-2020), spatially characterized by lower values in central regions and higher values at the periphery. The carbon sink-source system has undergone a three-stage evolution of "rapid repair-dynamic balance-game adjustment," spatially manifesting as "marginal carbon fixation-central emission." (3) Partial least squares structural equation modeling (PLS-SEM) revealed that carbon loss stems from the interaction of natural constraints and human interference. Topography indirectly regulates carbon sinks by modulating climate and human activities, while vegetation suppression by human activities emerged as the primary driver of carbon loss. (4) Multi-scenario analyses predict that carbon storage in 2030 will retain the spatial differentiation observed from 2000 to 2020, maintaining a "high-edge, low-center" pattern. The sustainable development pathway (SSP1-2.6) yielded the most pronounced carbon storage gains. These findings provide a scientific foundation for optimizing carbon neutrality strategies in karst regions and insights for global carbon management in analogous geomorphic areas.

Microbial coalescence plays a crucial role in shaping aquatic ecosystems by facilitating the merging of neighboring microbial communities, thereby influencing ecosystem structure. Although this phenomenon is commonly observed in natural environments, comprehensive quantitative comparative studies on different lifestyle bacteria involved in this process are still lacking. The study focuses on 16S rRNA Amplicon Sequence Variants (ASVs) at the Jinsha River hydropower stations (Wudongde [WDD], Baihetan [BHT], Xiluodu [XLD], Xiangjiaba [XJB]), specifically examining free-living (FL) and particle-attached (PA) bacteria. Minimal differences in microbial composition were observed across water layers (surface, middle, and bottom). Analyses of overlapping ASVs, Bray-Curtis dissimilarity, and the SourceTracker algorithm revealed a significant difference in the coalescence ability of FL and PA bacteria, particularly in the surface water of XJB (FL: 31.1% ± 2.0%, PA: 27.6% ± 2.5%, p < 0.05). The coalescence of FL bacteria was primarily influenced by the mixing of adjacent water layers, while PA bacteria exhibited significant geographical variations across water layers (p < 0.05), displaying lower coalescence compared to FL bacteria. Using a cohesion metric, 12 keystone species in PA bacteria were identified and 7 in FL bacteria. Proteobacteria and Bacteroidetes were the most abundant phyla at the keystone species in PA and FL bacteria, respectively. The abundance of keystone ASVs decreased with distance in PA bacteria, whereas FL bacteria showed the opposite trend. At the genus level, Brevundimonas and Chryseobacterium were identified as keystone species in both lifestyles. Moreover, the impact of community coalescence on the stability tends to exhibit differences downstream in cascade stations. This study provides novel insights into the dynamic variations of microbial communities with diverse lifestyles in stratified aquatic environments and assesses the impact of dam construction on microbial coalescence and the alteration of keystone species.

The riparian zone is a crucial interface between the aquatic and terrestrial components of river ecosystems. Organic matter deposited here can serve as an important food source for aquatic organisms, earning it the name of detritus bait. The construction of hydrological facilities has altered the flow rhythms of rivers, reducing the inundated area of banks and the frequency of flow velocity changes, which in turn negatively affects the ecological stability of the river. Given these reasons, it is crucial to understand the distribution characteristics of debris bait on riparian and the recharge pattern to rivers under flow impacts. The physical properties intrinsic to the debris in question were examined in order to ascertain their influence on the dynamics of debris transport through natural sampling methods. The distribution patterns of debris on natural river banks were investigated using density separation. A quantitative experimental model was designed based on the results of natural investigations to analyze the response characteristics of debris in-stream recharge to hydrodynamic structures at different flow velocities. The results indicate that the debris distribution is influenced by the background soil situation and flow effect. The soil layer immediately adjacent to the surface of the bank has a higher recharge effect. The main period for riparian debris recharge occurs at the onset of changes in flow rate. Furthermore, there is a positive correlation between the flow velocity and the quantity of debris entering the river. The findings of this study offer valuable insights into the replenishment of river bank debris bait under the current conditions, with implications for the ecological management of rivers undergoing hydropower development.

This study utilized the discrete element method to simulate the dynamic processes of earthquake-induced landslides, aiming to explore the instability mechanisms, energy transformation, and movement characteristics of such landslides. By constructing a two-dimensional landslide model, typical landslides triggered by the Wenchuan earthquake were simulated, leading to the following key conclusions: (1) The landslide instability process can be divided into three consecutive stages: disintegration of slope materials (bond breakage), initiation and propagation of the slip surface from top to bottom, and coherent downslope movement of slope materials, followed by reorganization into a stable postfailure state. (2) Energy transformation and dissipation are critical factors in landslide dynamics, with frictional heat being the primary source of energy dissipation during the landslide process. (3) The evolution of kinetic energy over time shows initial stability, a gradual increase as the sliding mass moves downwards, and an eventual reduction to zero as the landslide ceases. The increase in heat corresponds to a decrease in kinetic energy, reflecting the mechanisms of energy transformation and dissipation during the landslide process. The results of this study demonstrate that the discrete element method has significant advantages in simulating complex geological phenomena, providing important insights for landslide prediction and disaster mitigation in earthquake-prone areas.

The time-frequency joint analysis method was used to study the seismic response characteristics of high-steep rock slopes with weak structural planes. Two models of homogeneous slopes and anti-dip slopes were established using FLAC3D. The results of the time-frequency joint analysis reveal that the topography, geological conditions and seismic wave propagation directions strongly influence the seismic response characteristics of the slopes. The high slope dynamic response of a homogeneous slope shows that the peak ground acceleration (PGA) periodically changes in a certain pattern with elevation and is obviously amplified at the slope crest. Weak structural planes cause local amplification and attenuation of seismic waves, while the two effects of horizontal seismic waves are greater than those of vertical seismic waves when passing through weak structural planes. Frequency-domain analysis reveals that the direction of seismic wave propagation notably affects the value and variation rule of the peak Fourier spectrum amplitude (PFSA), while weak structural planes strongly affect only this value. The dynamic deformation characteristics of the slopes are clarified according to Fourier spectrum analysis and modal analysis. The low-frequency components and high-frequency components mainly cause overall and local deformations, respectively, of the surface slope. In addition, the dynamic response characteristics of the slopes are further analysed based on Hilbert energy, and the applicability of the seismic Hilbert energy spectrum to reflecting the dynamic deformation characteristics of the slopes is determined. Moreover, the connection between the local deformation of the anti-dip slope and the occurrence of a landslide was discussed.

To investigate the effects of weak structural planes on the seismic response characteristics and failure mechanism of high-steep rock slopes, a time-frequency joint analysis method is proposed on the basis of the time, frequency and time-frequency domains. Four models, a homogeneous slope (Model 1), a horizontal layered slope (Model 2), an anti-dip slope (Model 3) and a bedding slope (Model 4), were established for numerical simulation via FLAC3D. The time-domain analysis shows that the structural planes contribute significantly to the dynamic response of the slopes. Seismic waves passing through a structural plane generate local amplification and attenuation, and the effects of different dip directions follow the order of anti-dip > horizontal layered > bedding. In addition, frequency-domain analysis revealed that the low-frequency components (< 5 Hz) and high-frequency components (> 11 Hz) caused the overall and local dynamic deformation of the slopes, respectively. The distribution and dip direction of structural planes determine the failure modes of slopes. Complete ensemble empirical mode decomposition with adaptive noise and Hilbert transform methods (CEEMDAN-HTs) were proposed as the improved Hilbert-Huang transform (HHT) to process seismic signals for time-frequency domain analysis, and the seismic energy in the marginal spectrum is distributed mainly in the low-frequency section. The time-frequency joint analysis method shows that the acceleration response and the marginal spectrum reflect the local and overall dynamic failure characteristics of the slopes, respectively. The local and overall dynamic instability probabilities of the slopes follow the order of bedding slope > anti-dip slope > horizontal layered slope > homogeneous slope.

The Jinsha River Basin, situated in the arid and warm upper reaches of the Yangtze River in China, presents distinctive challenges for investigating dissolved inorganic carbon (DIC) due to its regional geography and climate. This study explored the spatial distribution and cycling of DIC during the wet season at four cascade power stations: Wudongde (WDD), Baihetan (BHT), Xiluodu (XLD), and Xiangjiaba (XJB). The results indicate that the primary sources of DIC at these stations are the result of carbonate rocks weathering by carbonic acid and sulfuric/nitric acids. Specifically, DIC production from weathering at WDD significantly surpasses that at the other stations (BHT, XLD, XJB) under the sulfuric/nitric acid weathering conditions (p < 0.05). In the cascade reservoirs of the Jinsha River, biochemical processes predominantly facilitate the precipitation of calcium carbonate. Among these reservoirs, XLD exhibits the highest efficiency in converting DIC into calcite, with a transformation rate of Delta DICtrans reaching 2.62 Tg/year. During the wet season, the combined effects of air temperature, wind speed, and partial pressure result in higher average CO2 emissions from the cascade power stations in the Jinsha River and Yellow River basins compared to other cascade power stations, comparable to emissions from the Three Gorges Reservoir. Specifically, XLD exhibits the highest CO2 emission rate (FCO2 = 102.38 +/- 94.71 mmol/m2/d). Nevertheless, these levels are lower than those observed in most natural rivers globally. Buffer factor analysis indicates that the carbonate buffering capacity of the Jinsha River cascade reservoirs is relatively weak, with an average Revelle coefficient of approximately 41.4. These results enhance our understanding of carbon cycling in the Jinsha River's cascade power stations and provide crucial insights. for preserving the long-term stability of river ecosystems.

Hydropower development is one way to efficiently harness the energy of rivers. However, damming limits the upstream reproductive behavior of migratory fish, and fish passage facilities are now an important means of restoring connectivity. Flow velocity is an important factor for fish to perceive the environment and maintain swimming behavior. The reservoir in front of the dam has become a near-hydrostatic environment that significantly changes the river properties. Reduced velocity may make it difficult for fish released into the reservoir area by fish passage facilities to find the flow direction, reducing their migratory success rate. This paper takes the Wudongde reservoir as an example and uses numerical simulation to analyze the characteristics of the flow field in the reservoir at different flow levels. It takes the induced velocity of fish as the threshold value to analyze the river sections in the reservoir affected by the reduced velocity. The results show that the flow velocity in the reservoir area decreases significantly from the end of the reservoir to the front of the dam; the flow velocity in the main reservoir is more homogeneous in the direction of water depth, and the tributary estuaries present non-uniform distribution. Under the conditions of low flow (February), medium flow (May), and high flow (August), a certain range of the river section is below the induced velocity of the fish (0.2 m/s), and the suitable locations for migratory fish stocking are located in the upper part of Wudongde Dam, 102, 96, and 38 km, respectively. The results provide technical references for the selection of release sites for fish passage facilities in high dams and large reservoirs, and are of great significance for the protection of water ecology in the dam-built river reaches.

Cleft lip and palate deformities are usually accompanied by abnormal external nasal morphology which seriously affects facial aesthetics and quality of life. In-depth understanding of the growth pattern of external nasal morphology in patients with cleft lip and palate is essential for deciding the optimal timing of surgical intervention and for furthering improvements in surgical technique. In this review we systematically depict the cross-sectional changes in external nose morphology during growth in the normal population and in patients with clefts, and summarise the iatrogenic impacts on growth of the cleft nose, aiming to provide a reference to further improve the management of the nose in patients with clefts.

The confluence of two rivers gives rise to a distinctive hydrodynamic structure where numerous nutrients can be accumulated, making the confluence area a pivotal biological habitat. The threat of supersaturated dissolved gas in upstream water caused by dam discharging is indispensable for the confluence areas. Therefore, it is imperative to study and assess the risks faced by fish in these areas, particularly when the merging of main and tributary streams leads to varying degrees of supersaturation and non-uniform distribution of dissolved gases. The spatial and temporal distribution of supersaturated dissolved gas at the natural confluence in the Yangtze River utilizing continuous prototype monitoring and unsteady flow numerical simulation was investigated, and a dynamic evaluation method that classifies risk into three levels (high, medium, low) was proposed. The confluence area exhibited evident non-uniform distribution characteristics, wherein the primary factors influencing this distribution were identified as the dissolved gas saturation of inlet flow, and flow ratio. The presence of non-continuous discharge intensity resulted in an increase in medium-risk areas within the river, thereby mitigating the risk to fish from the supersaturation of the discharge. A velocity barrier was formed at the confluence, with the velocity range being 1.8 similar to 3.4 m/s, exceeding the fish's burst speed of 1.7 m/s. Because of the barrier, there was no direct access to the low-saturation areas, and potential avoidance routes are primarily located in the medium-risk zones downstream. The paper offers a useful method to assess risk distribution in critical ecological regions characterized by non-constant characteristics. Additionally, it recommends dam discharging strategies to expand the scope of fish risk avoidance.