Climate change impacts water supply dynamics in the Upper Rio Grande (URG) watersheds of the US Southwest, where declining snowpack and altered snowmelt patterns have been observed. While temperature and precipitation effects on streamflow often receive the primary focus, other hydroclimate variables may provide more specific insight into runoff processes, especially at regional scales and in mountainous terrain where snowpack is a dominant water storage. The study addresses the gap by examining the mechanisms of generating streamflow through multi-modal inferences, coupling the Bayesian Information Criterion (BIC) and Bayesian Model Averaging (BMA) techniques. We identified significant streamflow predictors, exploring their relative influences over time and space across the URG watersheds. Additionally, the study compared the BIC-BMA-based regression model with Random Forest Regression (RFR), an ensemble Machine Learning (RFML) model, and validated them against unseen data. The study analyzed seasonal and long-term changes in streamflow generation mechanisms and identified emergent variables that influence streamflow. Moreover, monthly time series simulations assessed the overall prediction accuracy of the models. We evaluated the significance of the predictor variables in the proposed model and used the Gini feature importance within RFML to understand better the factors driving the influences. Results revealed that the hydroclimate drivers of streamflow exhibited temporal and spatial variability with significant lag effects. The findings also highlighted the diminishing influence of snow parameters (i. e., snow cover, snow depth, snow albedo) on streamflow while increasing soil moisture influence, particularly in downstream areas moving towards upstream or elevated watersheds. The evolving dynamics of snowmelt-runoff hydrology in this mountainous environment suggest a potential shift in streamflow generation pathways. The study contributes to the broader effort to elucidate the complex interplay between hydroclimate variables and streamflow dynamics, aiding in informed water resource management decisions.

Study region: Urumqi River headwater region in eastern Tianshan, central Asia. Study focus: Climate change is anticipated to accelerate glacier shrinkage and alter hydrological conditions, causing variations in the runoff patterns in the catchment and significantly threatening the regional water resources. However, few models exhibit adequate performance to simulate both surface alterations and glacier/snow runoff. Therefore, this study combined the glacier module with the Soil and Water Assessment Tool (SWAT) model to estimate the effect of climate change on the streamflow in the Urumqi River headwater region. The Urumqi River Headwater region is representative because of its long data series, viatal location, and local water availability, and it contains the longest-observed reference glacier (Urumqi Glacier No.1) in China, which spans the period from 1958 to the present. New hydrological insights for the region: The SWAT model performed satisfactorily for both calibration (1983-2005) and validation (2006-2016) periods with a Nash-Sutcliffe efficiency (NSE) greater than 0.80. The water balance analysis suggested that the snow/glacier melt contributed approximately 25% to the water yield. At the end of the 21st century, the temperature would increase by 2.4-3.8 degrees C while the precipitation would decrease by 1-2% under two future scenarios (ssp245 and ssp585). Thus, a 34-36% reduction in streamflow was projected due to above climate change impacts. This information would contribute to the development of adaptation strategies for sustainable water resource management.

Hydrologic-land surface models (H-LSMs) offer a physically-based framework for representing and predicting the present and future states of the extensive high-latitude permafrost areas worldwide. Their primary challenge, however, is that soil temperature data are severely limited, and traditional model validation, based only on streamflow, can show the right fit to these data for the wrong reasons. Here, we address this challenge by (1) collecting existing data in various forms including in-situ borehole data and different large-scale permafrost maps in addition to streamflow data, (2) comprehensively evaluating the performance of an H-LSM with a wide range of possible process parametrizations and initializations, and (3) assessing possible trade-offs in model performance in concurrently representing hydrologic and permafrost dynamics, thereby pointing to the possible model deficiencies that require improvement. As a case study, we focus on the sub-arctic Liard River Basin in Canada, which typifies vast northern sporadic and discontinuous permafrost regions. Our findings reveal that different process parameterizations tend to align with different data sources or variables, which largely exhibit inconsistencies among themselves. We further observe that a model may fail to represent permafrost occurrence yet seemingly fit streamflows adequately. Nonetheless, we demonstrate that accurately representing essential permafrost dynamics, including the active soil layer and insulation effects from snow cover and soil organic matter, is crucial for developing high-fidelity models in these regions. Given the complexity of processes and the incompatibility among different data sources/variables, we conclude that employing an ensemble of carefully designed model parameterizations is essential to provide a reliable picture of the current conditions and future spatio-temporal co-evolution of hydrology and permafrost.

Cyclonic storms (i.e., hurricanes) are powerful disturbance events that often cause widespread forest damage. Storm-related canopy damage reduces rainfall interception and evapotranspiration, but impacts on streamflow regimes are poorly understood. We quantify streamflow changes in Puerto Rico following Hurricane Maria in September 2017, and evaluate whether forest cover and storm-related canopy damage account for the differences. Streams are particularly vulnerable to flooding in early post-disturbance stages during hurricane season, so we focus on 3 months (Oct-Dec) following the hurricane. To discern changes in rainfall responses, we partitioned streamflow into baseflow and quickflow using a digital filter. We collected 2010-2017 streamflow and rainfall data from 18 watersheds and compared the relative magnitude of post- to pre-hurricane double mass curve slopes of baseflow and quickflow volumes against rainfall. Several watersheds displayed higher post-hurricane quickflow and baseflow, however, the response was variable. The magnitude of quickflow increase was greater in watersheds with high forest damage. Under the same level of relative damage, watersheds with low initial forest cover had greater quickflow increases than highly forested ones. Conversely, baseflow generally increased, but increases were greater in highly forested watersheds and smaller in highly damaged watersheds. These results suggest that post-storm baseflow increases were due to recharge of hurricane-related rainfall, as well as forest transpiration interruption and soil disturbance enhancing recharge of post-hurricane rainfall, while increases to quickflow are related to loss of canopy rainfall interception and higher soil saturation decreasing infiltration. Our research demonstrates that forest damage from disturbance lowers quickflow and elevates baseflow in highly forested watersheds, and elevates quickflow and lowers baseflow in less-forested watersheds. Less-forested watersheds may be closer to the forest cover loss threshold needed to elicit a streamflow response following disturbance, suggesting higher flooding potential downstream, and a lower storm-related forest disturbance threshold than in heavily forested watersheds. We quantify streamflow component changes following a severe hurricane and relate these changes to watershed forest cover and canopy damage. Several watersheds displayed higher post-hurricane quickflow and baseflow, however, the response was variable. Quickflow increases were greater in watersheds with high forest damage. Under the same level of relative damage, watersheds with low forest cover had greater quickflow increases than highly forested ones. Conversely, baseflow increases were greater in highly forested watersheds and smaller in highly damaged watersheds. image

Reliable drought prediction should be preceded to prevent damage from potential droughts. In this context, this study developed a hydrological drought prediction method, namely ensemble drought prediction (EDP) to reflect drought-related information under the ensemble streamflow prediction framework. After generating an ensemble of standardized runoff index by converting the ensemble of generated streamflow, the results were adopted as the prior distribution. Then, precipitation forecast and soil moisture were used to update the prior EDP. The EDP + A model included the precipitation forecast with the PDF-ratio method, and the observed soil moisture index was reflected in the former EDP and EDP + A via Bayes' theorem, resulting in the EDP + S and EDP + AS models. Eight basins in Korea with more than 30 years of observation data were applied with the proposed methodology. As a result, the overall performance of the four EDP models yielded improved results than the climatological prediction. Moreover, reflecting soil moisture yielded improved evaluation metrics during short-term drought predictions, and in basins with larger drainage areas. Finally, the methodology presented in this study was more effective during periods with less intertemporal variabilities.

Climate change and rapid socioeconomic development have exacerbated the damage caused by hydrological droughts. To ensure effective drought defense and infrastructure development, it is essential to investigate variations in hydrological droughts. The primary objective of this study is to reconstruct the natural streamflow by using Soil and Water Assessment Tool (SWAT) hydrological modeling. The hydrological drought at different time scales (1, 3, 6, and 12 months) were measured using the streamflow drought index (SDI). The statistical parameters, including Nash-Sutcliffe Efficiency and the Coefficient of Determination, which yielded values of 0.84 and 0.82 during the calibration period and 0.78 and 0.76 during the validation period, respectively, showed a satisfactory SWAT model performance. Additionally, the Pettit test was used to identify a change point in streamflow within the 1991-2015 timeframe, leading to the division of the study period into two distinct phases: an undisturbed period (1991-1998) and a disturbed period (1999-2015). The SDI index-based analysis revealed 9.39% moderate drought and 3.13% severe drought during the undisturbed period, while 11.76% moderate drought and 7.35% severe drought may happen due to the human influences that occurred in the disturbed period. These findings enhance the understanding of the hydrological drought variations in the Soan River basin for optimizing the water resources management system and effectively preventing and mitigating drought-related damages.

Land-use change may significantly influence streamflow. The semi-empirical model PhosFate was used to analyze the impact of land use and climate change on streamflow by choosing the Guishui watershed as a pilot site and then expanding, applying it to all of North China. The Guishui watershed (North Beijing, China) has experienced a dramatic decline in its streamflow in recent decades. Parallel to this, significant land-use change has happened in this area; afforestation programs have increased forest cover from 41% (1980) to 59% (2013) and a similar increase in forest cover can also be observed in North China. Managing flow decline requires separating climatic and direct human-influenced effects. The results showed the following: (1) Afforestation is a major factor that decreased total flow in the Guishui watershed from 1996 to 2014; total flow increased by around 24% more than the actual dataset in the constant scenario (no afforestation) and decreased by 5% more than the actual dataset in the forest scenario (all agriculture land use transferred to forests). (2) When forest coverage increases, the Qinghai-Tibet Plateau and the Loess Plateau are the most sensitive areas regarding total flow in North China; the total flow change rate increased by up to 25% in these two areas when land use shifted from sparse vegetation to mixed forests. After analyzing the contributions of these two factors, we formulated recommendations on future afforestation practices for North China. In the central-north and northwest districts, the annual precipitation is under 520 mm and 790 mm, respectively, and the practice of afforestation should be more carefully planned to prevent severe damage to streams. This research also proved that the PhosFate model can be used in North China, which would be a practical tool for watershed management.

A climate transition towards warm-wet conditions in Northwest China has drawn much attention. With continuous climate change and universal glacier degradation, increasing water-related hazards and vulnerability have become one of the important problems facing the Tarim Basin. However, the impacts of the climate transition on streamflow abrupt change and extreme hydrological events were less discussed, especially in glacial basins. In the present study, the discharge datasets in four glacial basins of Tarim Basin from 1979 to 2018 were constructed using the GRU-GSWAT thorn model first. The differences in streamflow characteristics, the shift of hydrological extreme pattern, and potential changes of the controlling factors before and after the abrupt changes were investigated. The results indicated that the abrupt change point (ACP) in streamflow occurred in 2000 in the Qarqan River Basin, 2002 in the Weigan River Basin, and 1994 in the Aksu River Basin and the Yarkant River Basin. A general decrease in streamflow before the ACP has shifted to a notable upward trend in the Qarqan River Basin and the Aksu River Basin, while minor upward fluctuations were observed in other basins. Moreover, the hydrological characteristics in extreme events vary dramatically before and after the ACPs, characterized by a pronouncing shift from drought-dominant pattern to wet events dominated pattern. The driven climate factors have been altered after the ACPs with notable spatial heterogeneity, in which temperature remained as the dominant role in meltwater-dominated basins while the influence of precipitation has increased after the ACPs, whereas the sensitivity of temperature on streamflow change has been enhanced in basins dominated by precipitation such as the Qarqan River Basin. Owing to the evident warming-wetting trend and glacier compensation effect, both the inter-annual and intra-annual streamflow fluctuations can be efficiently smoothed in basins with a high glacier area ratio (GAR). These findings provide a further understanding of the abrupt change in streamflow under the exacerbated climate and glacier change in mountainous arid regions.

Determining the age and sources of stream water is critical for understanding the watershed hydrological processes and biogeochemical cycle. In this study, daily isotope data of rainfall and runoff, as well as continuously monitored conductivity data from June to October in 2019 in-Laoyeling(LYL) watershed located in permafrost region of northeastern China were used to separate streamflow components through the application of two independent methods: isotope-based hydrograph separations (IHS) and the conductivity mass balance (CMB) methods. The results showed that stream water in a boreal forest watershed with permafrost of the Daxing'an Mountains is mainly composed of pre-event water. Although the IHS method is more sensitive and provides more details than the CMB method, the results of both methods show a similar trend. The average value of the young water fractions (Fyw) for those aged less than 65 days is 5.6%, while the mean transit time (MTT) was calculated to be 3.33 years. These findings enhance our understanding of the fundamental characteristics of runoff generation mechanisms and changes in runoff components in permafrost regions. Such knowledge is crucial for effective regional water resource management under the context of climate change, such as construction of water conservancy facilities and prediction of flood and drought disasters.

Robust streamflow simulation at glacial basins is essential for the improvement of water sustainability assessment, water security evaluation, and water resource management under the rapidly changing climate. Therefore, we proposed a hybrid modelling framework to link the SWAT+ model considering glacial hydrological processes (GSWAT+) with Gated Recurrent Unit (GRU) neural networks to improve the model simulations and to establish a framework for the robust simulation and forecast of high and low flows in glacial river basins, which could be further used for the explorations of extreme hydrological events under a warming climate. The performance of different models (GSWAT+, GRU, and GRU-GSWAT+, respectively) were thoroughly investigated based on numerical experiments for two data-scarce glacial watersheds in Northwest China. The results suggested that the hybrid model (GRU-GSWAT+) outperformed both the individual deep learning (DL) model (GRU) and the conventional hydrological model (GSWAT+) in terms of simulation and prediction accuracy. Notably, the proposed hybrid model considerably enhanced the simulations of low and high flows that the conventional GSWAT+ failed to capture. Furthermore, utilizing suitable data integration (DI) schemes on feature and target sequences can substantially help to strengthen model stability and representativeness for monthly and annual streamflow sequences. Specifically, introducing one order differential method and decomposition approach, such as ensemble empirical signal decomposition (EEMD) and complete EEMD with adaptive noise (CEEMDAN), into feature and target sequences enriched the learnable ancillary information, which consequently strengthened the predictive performance of the proposed model. Overall, the proposed hybrid model with the suitable DI scheme has the potential to significantly enhance the accuracy of streamflow simulation in data-scarce glacial river basins. This hybrid model not only upheld the fundamental physical principles from the GSWAT+ model, but also considerably mitigated the accumulated bias errors, which caused by the shortage of climate data and inadequate hydrological principles, by using DL based model and DI schemes.