The simulation and prediction of glacially derived runoff are significant for water resource management and sustainable development in water-stressed arid regions. However, the application of a hydrological model in such regions is typically limited by the intricate runoff production mechanism, which is associated with snow and ice melting, and sparse monitoring data over glacierized headwaters. To address these limitations, this study develops a set of mathematical models with a certain physical significance and an efficient particle swarm optimization algorithm by applying long- and short-term memory networks on the glacierized Muzati River basin. First, the trends in the runoff, precipitation, and air temperature are analyzed from 1990 to 2015, and differences in their correlations in this period are exposed. Then, Particle Swarm Optimization–Long Short-Term Memory (PSO-LSTM) and Bi-directional Long Short-Term Memory (BiLSTM) models are combined and applied to the precipitation and air temperature data to predict the glacially derived runoff. The prediction accuracy is validated by the observed runoff at the river outlet at the Pochengzi hydrological station. Finally, two other types of models, the RF (Random Forest) and LSTM (Long Short-Term Memory) models, are constructed to verify the prediction results. The results indicate that the glacially derived runoff is strongly correlated with air temperature and precipitation. However, in the study region over the past 26 years, the air temperature was not obviously increasing, and the precipitation and glacially derived runoff were significantly decreasing. The test results show that the PSO-LSTM and BiLSTM runoff prediction models perform better than the RF and LSTM models in the glacierized Muzati River basin. In the validation period, among all models, the PSO-LSTM model has the smallest mean absolute error and root-mean-square error and the largest coefficient of determination of 6.082, 8.034, and 0.973, respectively. It is followed by the BiLSTM model having a mean absolute error, root-mean-square error, and coefficient of determination of 6.751, 9.083, and 0.972, respectively. These results imply that both the particle swarm optimization algorithm and the bi-directional structure can effectively enhance the prediction accuracy of the baseline LSTM model. The results presented in this study can provide a deeper understanding and a more appropriate method of predicting the glacially derived runoff in glacier-fed river basins.

Understanding the main drivers of runoff availability has important implications for water-limited inland basins, where snow and ice melt provide essential input to the surface runoff. This paper presents an analysis on the runoff response to changes in climatic and other controls of water-energy balance in an inland glacierized basin, the Urumqi River basin, located in the arid region of northwest China, and identifies the major control to which runoff is sensitive across the basin’s heterogeneous subzones. The results indicate that the runoff is more sensitive to change in precipitation in the mountainous headwaters zone of the upper reach, and followed by the impact of basin characteristics. In contrast, the runoff is more sensitive to changes in the basin characteristics in the semiarid and arid zones of the mid and lower reaches. In addition, the change in basin characteristics might be represented by the distinct glacier recession in the mountainous upper reach zone and the increasing human interferences, i.e., changes in land surface condition and population growth, across the mid and lower reach zones. The glacier wasting contributed around 7% on average to the annual runoff between 1960 and 2012, with an augmentation beginning in the mid-1990s. Findings of this study might help to better understand the possible triggers of streamflow fluctuation and the magnitude of glacier wasting contribution to runoff in inland glacierized river basins.

The fluctuation of streamflow in snowmelt-dominated watersheds may be an indicator of climate change. However, the relationship between the start of growing season (SOS) and the streamflow in snowmelt-dominated watersheds is not clear. In this study, we update the Coupled Hydro-Ecological Simulation System (CHESS) model by incorporating the Growing Season Index (GSI) module to estimate the start of the growing season. The updated CHESS model is then used to calculate the streamflow in the Cleve Creek, Incline Creek and Twin River watersheds located in Nevada in the United States from 1981 to 2017. This updated CHESS can be applied in any regions that are suitable for deciduous vegetation. The streamflow in the static and dynamic scheme in the three watersheds have been simulated between 1981 and 2017 with the NS of 0.52 and 0.80 in the Cleve Creek, 0.46 and 0.75 in the Incline Creek, and 0.42 and 0.70 in the Twin River watersheds, respectively. The results illustrate that the SOS have come around 3–5 weeks earlier during the last 37 years. The results illustrate a high correlation between the temperature and the timing of the SOS. Early SOS leads to a substantial increase in total annual transpiration. An increase in annual transpiration can reduce aquifer recharge and increase cumulative growing season soil moisture deficit. Comparing to the streamflow without vegetation, the streamflow with vegetation is smaller due to transpiration. As the SOS comes earlier, the peaks of the streamflow with vegetation also come earlier. If the shifts in SOS continue, the effects on annual rates of transpiration can be significant, which may reduce the risk of flooding during snowmelt. On the other hand, earlier SOS may cause soil moisture to decline during summer, which would increase the drought stress in trees and the risk of wildfires and insect infestation.

Various climatic factors and their interactions affect alterations in China's snow depth. However, a measure of the relative importance of climatic factors regarding the variance in reginal snow depth is lacking. Here, we use a large dataset, quality-controlled daily observations from 1,665 meteorological stations, and statistical regression modelling to identify main drivers and their contributions regarding regional variability and trends in China's snow depth and snow depth-climate relationships from 1951 to 2019. Results show that seasonal snow depth alterations are substantial in China and caused by various climatic drivers. In winter, changes in snow depth across the northwest and northeast of China exhibit a more general increasing trend, while changes in snow depth across the Tibetan Plateau and the remaining regions of China show a more common decreasing trend. The observed positive trends in the northwest and northeast of China mainly respond to augmentation in solid precipitation, explaining nearly 60% of the total interannual variance, and snow accumulation in fall. For the Tibetan Plateau, the negative snow depth anomaly seems to respond to the rising air temperature below the freezing point along with some secular changes in solid precipitation and shallower fall snowpacks. All these trends in winter snow depth over the northwest, northeast, and Tibetan Plateau of China translate into the properties of spring snow depth, accounting approximately for 30% of the total interannual variance in spring snow depth in these regions. In the milder central north and south of China, snow accumulation is not evident. Besides solid precipitation and rising air temperature below the freezing point, atmospheric humidity is also notable for its effect on the interannual variance in winter snow depth in these regions. In spring, however, China's snow depth exhibits a general reduction. Particularly, the central north of China, has experienced the greatest reduction of nearly 50% in spring snow depth during the recent two decades. In fall, on the other hand, changes in snow depth are mostly insignificant in China, but the stronger signal of reduction mainly occurs on the Tibetan Plateau. Moreover, in spring and fall, the negative effect of warming climate on snow depth becomes remarkably more important. It seems that about 13–33% (12–50%) of the total interannual variance in the widespread decrease (insignificant perturbation) of China's spring (fall) snow depth are caused by the rising air temperature below the freezing point.