Runoff Generation Research Articles

Application of a hybrid algorithm of LSTM and Transformer based on random search optimization for improving rainfall-runoff simulation

Flood forecasting using traditional physical hydrology models requires consideration of multiple complex physical processes including the spatio-temporal distribution of rainfall, the spatial heterogeneity of watershed sub-surface characteristics, and runoff generation and routing behaviours. Data-driven models offer novel solutions to these challenges, though they are hindered by difficulties in hyperparameter selection and a decline in prediction stability as the lead time extends. This study introduces a hybrid model, the RS-LSTM-Transformer, which combines Random Search (RS), Long Short-Term Memory networks (LSTM), and the Transformer architecture. Applied to the typical Jingle watershed in the middle reaches of the Yellow River, this model utilises rainfall and runoff data from basin sites to simulate flood processes, and its outcomes are compared against those from RS-LSTM, RS-Transformer, RS-BP, and RS-MLP models. It was evaluated against RS-LSTM, RS-Transformer, RS-BP, and RS-MLP models using the Nash–Sutcliffe Efficiency Coefficient (NSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Bias percentage as metrics. At a 1-h lead time during calibration and validation, the RS-LSTM-Transformer model achieved NSE, RMSE, MAE, and Bias values of 0.970, 14.001m3/s, 5.304m3/s, 0.501% and 0.953, 14.124m3/s, 6.365m3/s, 0.523%, respectively. These results demonstrate the model's superior simulation capabilities and robustness, providing more accurate peak flow forecasts as the lead time increases. The study highlights the RS-LSTM-Transformer model's potential in flood forecasting and the advantages of integrating various data-driven approaches for innovative modelling.

Impacts of climate change and land cover factor on runoff in the Coastal Chinese Mainland region

The increasingly frequent storms pose significant threats to the sustainable development of coastal regions, particularly in densely populated and economically vibrant areas. Comprehending the dynamics and intricate mechanisms underlying runoff generation is crucial in the context of climate change and anthropogenic interference. Based on hydro-meteorological and land-use data from 1980 to 2018, this study investigates the runoff variation and its driving factors in the Coastal Chinese Mainland (CCM). The aims of this study are to reveal the temporal and spatial trends of runoff yield, to clarify the sensitivity of runoff in coastal cities to the integrated and individual parameters of climate change and anthropogenic interference, including precipitation (P), potential evapotranspiration (E0), and land cover factor (n), and to support the establishment of spatially tailored adaptation strategies. The results show that: (1) runoff has generally increased over the study period, particularly in regions such as the Yangtze River Delta, Shandong, and Guangxi, while it has decreased in western Liaoning and eastern Guangdong; (2) in the northern CCM with larger aridity index, the land cover factor plays a dominant role in runoff production, while in the wetter southern CCM, precipitation is more influential, and potential evapotranspiration mainly hinders runoff generation all over CCM; (3) urban expansion tends to negatively impact n, while the loss of grasslands and shrinkage of croplands tend to undermine the value of n. To facilitate the achievement of sustainable development goals in the CCM, it is imperative to introduce a more comprehensive and theoretical framework that encompasses the natural, technical, and social dimensions of human-water systems into traditional flood regulation and water resource management. This framework should promote interdisciplinary collaboration from an integrated perspective, to bridge the administrative and watershed boundaries, to effectively address the complex challenges posed by climate change and anthropogenic activities on runoff and water resources in coastal regions, and to enhance the realization of local sustainable development goals (UN SDGs).

Runoff generation in a semiarid environment: The role of rainstorm intra-event temporal variability and antecedent soil moisture

Rainfall intensity and antecedent soil moisture are key variables affecting the initiation and accumulation of runoff, and therefore, impact on the occurence flood events. We used time series of rainfall intensities characterizing rainstorms with different mean, standard deviation, and skewness to simulate runoff response of a sealed loamy soil profile under various initial soil moisture conditions. We found that shifting from wet to dry initial soil moisture leads to a 300 % increase in ponding time and a 350 % decrease in total runoff. Intense rainfall events catalyze runoff development and runoff occurs earlier with increasing mean, maximum intensity, and variance of rainfall. Higher variance in intra-event storm-intensity temporal distribution yields more runoff. Earlier runoff formation aligned with negatively skewed storm structures, but largest cumulative runoff resulted from storms displaying near-zero skewness values. The combination between rainfall intensification and extended dry periods between storms might reduce potential runoff. These insights improve our understanding of anticipated impacts of climate variations on floods and may also contribute to ecosystem health evaluations, and bolster predictive capabilities of surface processes.

Comparison of two hydrodynamic models for their rain-on-grid technique to simulate flash floods in steep catchment

In this study, two hydrodynamic models, TELEMAC-2D and HEC-RAS 2D, were compared for their Rain-on-Grid (RoG) technique with a particular focus on runoff generation processes in a small and steep catchment. Curve number (CN) method was applied in both the models to simulate two single storm events up to 20 h of duration, whereas the Green-Ampt Redistribution (GAR) method was additionally applied in HEC-RAS 2D for a multi-peak flood event with sustained flow between the peaks. CN and GAR methods were compared for this flood event, and a sensitivity analysis of the GAR parameters was also done. Moreover, the two models were compared for their calibration process, computational time, mesh size and shape, and model availability, in general, as well as the results including inundated areas, water depth, and velocity. The results indicate that both the models are capable of reproducing short duration single storm floods. NSE and R2 for both models ranged from 0.70 to 0.90 and from 0.93 to 0.95. However, the models struggled to reproduce the long- duration multi-peak flood event. The sensitivity analysis showed that the results are not very sensitive to the two GAR parameters which are responsible to influence the flow of the second peak in the flood event. Neither the CN nor the GAR infiltration method successfully replicated such events because the hydraulic models permanently lose infiltrated water from the domain. The returned sub-surface flow significantly contributes to river flow during these flood events; however, none of the model incorporates a return flow algorithm.

Sensitivity of Runoff to Climatic Factors and the Attribution of Runoff Variation in the Upper Shule River, North-West China

Climate change and human activities exert significant impact on the mechanism of runoff generation and confluence. Comprehending the reasons of runoff change is crucial for the sustainable development of water resources. Taking the Upper Shule River as the research area, the M-K test and the moving t test were used to diagnose the runoff mutation time. Furthermore, the slope changing ratio of cumulative quantity method (SCRCQ), climate elasticity method, and Budyko equation were utilized to quantitatively evaluate the impacts and contribution rates of climate change and human activities. The following results were obtained: (1) The Upper Shule River experienced a significant increase in runoff from 1972 to 2021, with 1998 marking the year of abrupt change. (2) The runoff sensitivity showed a downward trend from 1972 to 2021. The main factor affecting the decrease in runoff sensitivity was the characteristic parameters of underlying surface (n), followed by precipitation (P), while the influence of potential evapotranspiration (ET0) was the weakest. (3) The response of runoff changes to runoff sensitivity and influencing factors were 90.32% and 9.68%, respectively. (4) The results of three attribution methods indicated that climate change was the primary factor causing the alteration of runoff in the Upper Shule River. The research results supplement the hydrological change mechanisms of the Upper Shule River and provide a scientific basis for future water resources management and flood control measures.

Simulating the Influence of Crop Cultivation Practices and Slope Gradients on Curve Number and Runoff Generation in Semi-arid Flat Catchments

Simulating the Influence of Crop Cultivation Practices and Slope Gradients on Curve Number and Runoff Generation in Semi-arid Flat Catchments

GIS and cellular automata based slope rainwater movement process model and its application

Rainfall serves as a significant factor contributing to slope stability challenges in mountainous areas, and simulating the process of slope rainwater movement is a crucial approach for analyzing the stability of slopes triggered by rainfall. By combining computer numerical simulation technology with traditional hydraulic and hydrological calculation theories, it is possible to create an efficient and precise rainwater movement model that can simulate and analyze the process of rainwater movement on slopes. Utilizing natural slopes as the focal point of our research, the cellular automaton method was applied to simulate rainfall runoff on slopes, and a Cellular Automata (CA) based model for rainwater movement process was developed. This model modified the Green-Ampt (G-A) infiltration model by adopting an elliptical water content curve and introducing a coefficient that quantifies the ratio of saturated to unsaturated depth. Additionally, we refined the rules governing runoff generation and convergence within the slope and on its surface, enabling a comprehensive simulation of the entire rainwater movement process on the slope. Furthermore, the effectiveness of the model was validated through analytical solutions derived from simplified assumptions, laboratory experiments on infiltration and runoff in the flume, and a case study of a natural slope. The results show that the infiltration calculation results of the rainwater movement model are closer to the experimental values, and their overall values are slightly higher than the measured values, which are basically consistent with the model test results; The runoff calculation results show a phenomenon of initially increasing and gradually approaching the measured values compared to the measured values. When applying the model to an actual slope, it was found that the model comprehensively accounts for the influence of slope seepage, infiltration and runoff process, has better performance compared to G-A modified model. The model can be used to describe the spatial distribution and temporal variation of infiltration and runoff processes.

Soil moisture dynamics and associated rainfall-runoff processes under different land uses and land covers in a humid mountainous watershed

Identifying soil moisture dynamics is critical for understanding watershed hydrological processes. Soil moisture responses to rainfall vary with land use and land covers (LULCs) and thus influence rainfall-runoff processes, however, the knowledge about how different LULCs affect such processes was less revealed, especially in humid areas. In this study, we investigated the characteristics of soil moisture content (SMC) and rainfall-related soil moisture responses under four typical LULCs, i.e., waxberry forest, farmland, sloping farmland, and bamboo forest, based on in-situ observations in a humid mountainous watershed. We then used the HYDRUS-1D model to analyze the patterns of the rainfall-infiltration processes and quantified corresponding runoff generation to reveal the partitioning between infiltration and runoff at the event scale. We found significant differences of the average SMC at multiple depths for four LULCs. The deep soil (80 cm) had the largest SMC for the waxberry forest, farmland, and sloping farmland, while the surface soil (10 cm) was the wettest soil layer for the bamboo forest. During the representative rainstorms, although soil moisture responses exhibited depth gradients, the earlier responses at some deep layers indicated the occurrence of preferential flow. Generally, the cumulative infiltration increased while the infiltration rate decreased with the rainfall grade. Farmland and bamboo forest showed the most and the least infiltration, respectively. These results suggested that both the rainfall properties and LULCs had a strong impact on the soil moisture responses. Meanwhile, the estimated average runoff coefficients of four typical LULCs increased with rainfall grade, and the sloping farmland generated the most runoff during the representative rainfall events. Our findings provide new insights into soil moisture dynamics and infiltration regimes under different LULCs in humid areas, which can contribute to hydrological modeling at the field-scale and regional water resource management.

Clustering of small watersheds in hilly areas based on complex network theory and similarity analysis

ABSTRACT Clustering analysis of small watersheds is an effective tool for identifying the similarity of runoff generation and concentration. In this paper, 545 small watersheds in the hilly areas of Shandong Province were investigated, and 12 indicators representing their climate and subsurface characteristics were selected to identify communities based on hydrological similarity. We further analyzed the hydrological connections among the small watersheds within each community using three indicators (network mean, centrality, and k-core). Finally, the clustering results were evaluated on the basis of the small watershed flood peak modulus. The results of this complex network method indicate that the study area contained six large communities and nine small communities. The community-clustering results were reasonable and showed the interconnectedness of the watersheds within each community. The three network indicators adequately described the degree of similarity, the representativeness of the watersheds, and the spatial scales of similar hydrological features. This method should be helpful for addressing the issue of parameter transplantation in ungauged watersheds and implementation of a flood risk management strategy.

Application of multi-criteria decision-making on low-impact development practice selection for the Kinyerezi River sub-catchments in Dar es Salaam, Tanzania.

The impermeable areas in catchments are proportional to peak flows that result in floods in river reaches where the flow-carrying capacity is inadequate. The high rate of urbanization witnessed in the Kinyerezi River catchment in Dar es Salaam city has been noted to contribute to floods and siltation in the Msimbazi River. The Low-Impact Development (LID) practices that includes bio-retention (BR) ponds, rain barrels (RBs), green roofs (GRs), etc. can be utilized to mitigate portion of the surface runoff. This study aims to propose suitable LID practices and their sizes for mitigating runoff floods in the Kinyerezi River catchment using the Multi-Criteria Decision-Making (MCDM) approach. The results indicated that the BR and RBs were ranked high in capturing the surface runoff while the sediment control fences were observed to be the best in reducing sediments flowing into the BR. The proposed BR ponds were greater than 800 m2 with 1.2 m depth while RB sizes for Kinyerezi and Kisungu secondary schools and Kinyerezi and Kifuru primary schools were 2,730; 2,748; 1,385; and 1,020 m3, respectively. The BR ponds and RBs are capable of promoting water-demanding economic activities such as horticulture, gardening, car washing while reducing the school expenses and runoff generation.

Evaluation of hydraulic adequacy of runoff disposal in urban watersheds of rapidly expanding towns

ABSTRACT The growth of cities significantly alters natural catchments by increasing impervious surfaces and necessitating the installation of an appropriate drainage system. Arba Minch is rapidly expanding and facing street flooding. The objective of this study is to assess the hydraulic adequacy of runoff disposal in urban watersheds located in rapidly expanding towns. Stormwater Management Model (SWMM) was used to perform rainfall-runoff simulation. Personal Computer SWMM (PCSWMM) was used to carry out calibration and validation in the watershed. The primary and secondary data were used. Five Land Use/Land Cover (LULC) categories identified were asphalt, cobble, vegetation, bare soil, and roof. The result shows 33.49% area is covered with highly impervious land cover. The overall calibration and validation are a very good fit with the observed flow. The total runoff volume of 9494.15 km3 was generated from the town area which is 13.7 km2. The peak runoff determined from 373 sub-catchments varied from 0.01 to 4.48 m3s−1. According to the simulated result, 7.46% of existing drainage channels were flooded. Anticipating future runoff generation, this study evaluates the inadequacy of existing drainage channels in urban areas. Hydraulic analysis is recommended before constructing drainage structures to protect from flooding.

Hydrological responses to permafrost degradation on Tibetan Plateau under changing climate

The Tibetan Plateau (TP) has undergone significant warming and humidification in recent years, resulting in rapid permafrost degradation and spatiotemporal variations in hydrological processes, such as subsurface water transport, hydrothermal conversion, and runoff generation. Understanding the mechanisms of hydrological processes in permafrost areas under changing climate is crucial for accurately evaluating hydrological responses on the TP. This study comprehensively discusses the permafrost hydrological processes of the TP under changing climate. Topics include climate conditions and permafrost states, subsurface water transport under freeze–thaw conditions, development of thermokarst lakes and hydrothermal processes, and runoff response during permafrost degradation. This study offers a comprehensive understanding of permafrost changes and their hydrological responses, contributing significantly to water security and sustainable development on the TP.

Assessment of land use and land cover dynamics and its impact in direct runoff generation estimation using SCS CN method

Assessment of land use and land cover dynamics and its impact in direct runoff generation estimation using SCS CN method

Responses of streamflow to forest expansion in a typical subhumid watershed under future climate conditions

Water availability in the subhumid region is highly vulnerable to frequent droughts. Water scarcity in this region has become a limiting factor for ecosystem health, human livelihood, and regional economic development. A notable pattern of land cover change in the subhumid region of the United States is the increasing forest area due to afforestation/reforestation and woody plant encroachment (WPE). Given the distinct hydrological processes and runoff generation between forests and grasslands, it is important to evaluate the impacts of forest expansion on water resources, especially under future climate conditions. In this study, we focused on a typical subhumid watershed in the United States – the Little River Watershed (LRW). Utilizing SWAT + simulations, we projected streamflow dynamics at the end of the 21st century in two climate scenarios (RCP45 and RCP85) and eleven forest expansion scenarios. In comparison to the period of 2000–2019, future climate change during 2080–2099 will increase streamflow in the Little River by 5.1% in the RCP45 but reduce streamflow significantly by 30.1% in the RCP85. Additionally, our simulations revealed a linear decline in streamflow with increasing forest coverage. If all grasslands in LRW were converted into forests, it would lead to an additional 41% reduction in streamflow. Of significant concern is Lake Thunderbird, the primary reservoir supplying drinking water to the Oklahoma City metropolitan area. Our simulation showed that if all grasslands were replaced by forests, Lake Thunderbird during 2080–2099 would experience an average of 8.6 years in the RCP45 and 9.4 years in the RCP85 with water inflow amount lower than that during the extreme drought event in 2011/2012. These findings hold crucial implications for the formulation of policies related to afforestation/reforestation and WPE management in subhumid regions, which is essential to ensuring the sustainability of water resources.

Analysis of Water Conservation Priority Locations in Upper Sub Watershed

Water is useful for supplying household, industrial, and agricultural needs. The availability of water supports food security and decent life. The process of sustaining water resources requires conservation, which is in line with sustainable development goals. The Keyang sub watershed is one of the catchment areas of the Upstream Madiun River, located in East Java. The development of agriculture land in Keyang sub watershed increases the number of disasters. The Central Statistics Office of Ponorogo Regency recorded 24 floods, 11 drought disasters, and 42 landslides in Keyang sub watershed during 2020. Floods and drought disasters in the same location show strong fluctuations in water resources, so Keyang sub watershed requires water conservation actions. The aim of this study is to analyze priority locations for water conservation. Indicateurs de Ruissellement Intense Pluvial (IRIP) is a method for determining water conservation priority locations. IRIP divides surface runoff vulnerability area into three types: generation area, transfer area, and accumulation area. Each area requires different adaptations to conservation techniques. The result shows that Keyang sub watershed has 4.96% strong potential surface runoff generation area. This priority area is suitable for wetland development, agroforestry, and sustainable agriculture. Then, Keyang sub watershed has 0.03% strong potential area of surface runoff transfer, this area is compatible with prevention of erosion and landslides. Keyang sub watershed has 0.63% very strong potential and 18.42% strong potential for surface runoff accumulation area. Accumulation area appropriate for developing flood mitigation strategies.

Technical note: Testing the connection between hillslope-scale runoff fluctuations and streamflow hydrographs at the outlet of large river basins

Abstract. A series of numerical experiments were conducted to test the connection between streamflow hydrographs at the outlet of large watersheds and the time series of hillslope-scale runoff yield. We used a distributed hydrological routing model that discretizes a large watershed (∼ 17 000 km2) into small hillslope units (∼ 0.1 km2) and applied distinct surface runoff time series to each unit that deliver the same volume of water into the river network. The numerical simulations show that distinct runoff delivery time series at the hillslope scale result in indistinguishable streamflow hydrographs at large scales. This limitation is imposed by space-time averaging of input flows into the river network that are draining the landscape. The results of the simulations presented in this paper show that, under very general conditions of streamflow routing (i.e., nonlinear variable velocities in space and time), the streamflow hydrographs at the outlet of basins with Horton–Strahler (H–S) order 5 or above (larger than 100 km2 in our setup) contain very little information about the temporal variability of runoff production at the hillslope scale and therefore the processes from which they originate. In addition, our results indicate that the rate of convergence to a common hydrograph shape at larger scales (above H–S order 5) is directly proportional to how different the input signals are to each other at the hillslope scale. We conclude that the ability of a hydrological model to replicate outlet hydrographs does not imply that a correct and meaningful description of small-scale rainfall–runoff processes has been provided. Furthermore, our results provide context for other studies that demonstrate how the physics of runoff generation cannot be inferred from output signals in commonly used hydrological models.

Soil-Forming Factors of High-Elevation Mountains along the East African Rift Valley: The Case of the Mount Guna Volcano, Ethiopia

The soils of the high-elevation mountains along the East African Rift Valley are poorly understood. Assessing the potential of soils for agriculture, climate change mitigation, and environmental functioning requires insight into how they relate to the factors influencing soil formation. Between 3000 and 4120 m a.s.l., 85 soil profiles of Mount Guna were described and sampled. Standard physicochemical analyses were done on all pedons. Additionally, X-ray diffraction, Alox and Feox content, and P fixation were performed on six selected profiles. Soils on Mount Guna included Andosols, Phaeozems, Leptosols, Regosols, Cambisols, Luvisols, and Vertisols. With increasing elevation, clay content, bulk density, and pH decreased while the C:N ratio remained constant. In contrast, sand, silt, silt-to-clay ratio, SOC, Ntotal, and SOCS increased. With a factor analysis, the soil-forming factors’ elevation/climate could be disentangled from the factor’s parent material as these affect topsoil and subsoil differently. In the ordination based on climate/elevation and parent material, Andosols and Vertisols stood out while other Reference Soil Groups (RSG) showed indistinct patterns. Soil erosion appeared as an additional soil-forming factor not accounted for by the factor analysis. The distribution of the RSG was significantly associated with elevation belts (p < 0.001), lithology (p < 0.001), and landcover (p < 0.003). On the summital ridge, the Andosols were crucial for groundwater storage due to high precipitation. Shallow and stony soils in the mid-elevation belt contributed to runoff generation. Average soil carbon stock ranged from 8.1 to 11 kg C m−2 in the topsoil and from 29.2 to 31.9 kg C m−2 in the upper meter, emphasizing the global importance of high-elevation areas for carbon sequestration.

Hydrological effects of the conversion of tropical montane forest to agricultural land in the central Andes of Peru

AbstractAgricultural expansion is one of the main causes of deforestation of tropical montane forests in the central Andes of Peru. The hydrological impacts of converting forests to cropland are well known; however, an issue that is not entirely clear is related to the hydrological effects that can result from the conversion of montane forests to agroforestry systems and the recovery of the hydrological functions of montane forests after a disturbance. In this study, we compare the hydrological processes of different land covers previously modified by agricultural expansion, in order to determine the impact of the conversion of tropical montane forest to agricultural land on the ecosystem service of water provision and regulation. To achieve this, we establish study plots in four land cover types located in the central Andes of Peru (mature montane forest (BMC), natural regenerating secondary forest (BMR), coffee agroforestry systems (AF), and cropland (C)), for the purpose of measuring the vegetation structure and soil properties in them, and subsequently carry out a soil water balance in each plot to calculate the actual evapotranspiration, surface runoff, and groundwater recharge. The results revealed the following percentages (based on precipitation) for the hydrological components in the four land cover types: annual actual evapotranspiration—BMR (41.2%), AF (40.4%), BMC (40.0%), and C (38.0%); annual surface runoff—C (16.1%), BMC (8.3%), BMR (5.2%), and AF (4.6%); and annual groundwater recharge—AF (43.0%), BMR (41.6%), C (34.0%), and BMC (31.7%). Furthermore, the study also examined the relationship between vegetation structure and the hydrological components across the four land cover types. The findings indicated that the reduction of thicker and taller trees could increase surface runoff generation, whereas the presence of thinner and shorter trees could facilitate groundwater recharge. These results shed light on the complex interactions between land cover types, vegetation structure, and hydrological processes, emphasizing the importance of considering these factors in water resource management and land use planning.

Stability analysis of unsaturated loess slopes subjected to extreme rainfall incorporating creep effects

Rainfall-induced landslides are common natural hazards, particularly in cases involving unsaturated soils. Slope instability issues may arise due to extreme climate events, while, creep effects can also contribute to landslide problems. In this study, a modified ViscoElastic-ViscoPlastic constitutive model is developed to more accurately characterise the mechanical properties of unsaturated loess, considering stress-level dependency of stiffness and creep properties. In addition, to better characterise the collapsibility of loess, the relationship between strengths and matric suction has been introduced. The infiltration process of an unsaturated porous medium is illustrated by the Van Genuchten model and the generalized Darcy’s law. Additionally, the atmospheric boundary condition is incorporated in the numerical simulations, considering factors such as runoff generation and various climate events.In the context of slope stability analysis, a numerical approach has been developed to determine the safety factor of the slopes, by using displacement changes as the key assessment criterion. A good agreement between numerical and analytical results has been observed, verifying the proposed approach to determining safety factor. Finally, in parametric analyses, the effects of creep and matric suction on safety factor have been identified. Slope stability of unsaturated loess slopes subjected to different rainfall intensities and rainfall durations has been studied. The maximum depth of the sliding part increases obviously when considering the contribution of creep deformations, which is crucial in analysing slope stability.

Sensitivities of hydrological processes under climate warming and Landuse/landcover change in the Lhasa basin, Tibetan plateau

Study Region: The Lhasa River Basin (LRB) on the Tibetan Plateau.Study Focus: This study assessed the influence of climate change and landuse/landcover (LULC) change on hydrological processes using the water balancing simulation model (WaSiM). We used multiple data sources in a multi-objective calibration and validation approach, such as streamflow, snow cover, and glacier inventory, to minimize the uncertainties in the simulation results. Heat transfer in the soil was implemented in this study, which is crucial for permafrost regions.New Hydrological Insights for the Region: The results revealed spatiotemporal variation of precipitation change induced by climate warming in the last few decades. Runoff, actual evapotranspiration, and glacier melt are experiencing an increased trend. Meanwhile, decreased trends are signaled for snowmelt and water storage changes. Snowmelt showed an earlier peak that changed from June to May, which, combined with increased rainfall, resulted in an earlier peak of runoff in the basin. Rainfall is the main factor affecting runoff generation in LRB, while snowmelt and storage change are driven by the amount of snowfall, and actual evapotranspiration and glacier melt are most influenced by temperature. The annual discharge in LRB is projected to decrease, ranging from 20% to 39%, in response to the total permafrost degradation. Finally, the apparent variations in LULC exhibited in this study reveal a variety of water cycle and hydrological implications.