Watershed water policy based on a glacier water service perspective in the typical glacier basins, Western China

SDGs

sustainable development goals

GGAAs

global glacier-covered arid areas

SRB

Shule River Basin

YRSB

upstream region of the Yangtze River Basin

LWS

level of water stress

UN-FAO

United Nations Food and Agriculture Organization

EFR

Environmental Flow Requirement

VIC-CAS

Variable Infiltration Capacity-Chinese Academy of Sciences

IPCC

Intergovernmental Panel on Climate Change

CMIP

Coupled Model Inter-comparison Project

GCM

global climate model

RCP

representative concentration pathway

WWQ

water withdrawal quantity

PWD

primary industry water demand

SWD

secondary industry water demand

TWD

tertiary industry water demand

RWD

residents' domestic water demand

AEWD

artificial ecosystem water demand

Water shortage is a challenging environmental problem in arid regions around the world in the 21st century (World Economic Forum 2015; Doungmanee 2016). Since 1960, the amount of globally available fresh water per capita has decreased by 55%. The global demand for fresh water is expected to increase by 50% by 2030. By 2050, 2.3 billion people are expected to live in areas with severe water scarcity, and agriculture will account for 70% of all water withdrawal globally. The total economic cost of the global water crisis is estimated to be $500 billion per year (WWAP/UN-WATER 2018). In particular, in arid regions, global climate change and urban economic development have exacerbated the problems caused by the limited water supply and water shortages (Vorosmarty et al. 2000; IPCC 2014; Ercin & Hoekstra 2014). The Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report (AR6, 2021) stated that without significant reductions in greenhouse gas emissions, global surface temperatures are projected to rise by at least 2.1 °C by 2,100 (IPCC 2023). Under future climate change scenarios, the glacier retreat in global arid regions will intensify further. Glaciers, acting as a ‘water tower’ or solid reservoir, provide a copious amount of freshwater to both the natural environment and humans, playing a crucial role in the sustainable utilization of water resources in arid regional oases across the world (Immerzeel et al. 2020). The global population inhabiting arid regions, which is highly reliant on glacier water resources, amounts to nearly 2 billion. The expeditious shrinkage of glaciers has had far-reaching effects on both socioeconomic and natural ecosystems in areas downstream of glaciers (Ding & Zhang 2015; Ding & Zhang 2018; Ding et al. 2020), encompassing food production, industrial and agricultural water requirements, ecosystem services, and economic activities, all of which extensively depend on glacier runoff (Milner et al. 2017; Chen et al. 2023). Due to fragile natural conditions, very limited water resources, and serious human disturbances, the risk of desertification in arid regions will continue to be an important negative factor affecting local socioeconomic activities (Allen et al. 2010). On the contrary, in the river source region of the alpine region, the glacier scale is larger, the glacier runoff changes are more stable, but the social water demand is less, and the water stress degree is lower. In view of the differences between the glaciers and their runoff and the social water demand in the lower reaches of the arid region and the alpine river source region, the watershed water resources policy needs to be treated differently (Wang et al. 2021).

The United Nations Food and Agriculture Organization (UN-FAO) and the Water Resources Organization (UN-Water) have used sustainable development goal (SDG) 6.4.2 indicators (level of water stress (LWS)) and their accounting methods to produce water stress progress reports in five pilot countries (Jordan, the Netherlands, Peru, Senegal, and Uganda) (FAO and UN-Water 2018). However, the current LWS method and its assessment progress are mainly concentrated on arid areas that have no glaciers, and the assessment cases are limited to the national scale (Gleick 2014; Vanham et al. 2018; Hannah & Max 2020) and lack the scale of global glacier-covered arid areas (GGAAs). It is necessary to adapt the LWS assessment method for use at the GGAA scale (Wang et al. 2021).

The Shule River Basin (SRB) and Yangtze River Source Basin (YRSB) in western China are representative inland river basins in arid regions, and the upstream areas of these great rivers contain glacier runoff. The natural and social environments in these two basins are significantly different, and their glacier water resource services also exhibit significantly different characteristics, thus, these two basins may require the implementation of completely different sustainable development strategies in the future. Therefore, it is necessary to conduct a comparative study of the glacier water resource services in these two basins. On the basis of the research on the impact and adaptation of cryosphere changes, Chinese scientists were the first to propose the definition of cryospheric services: the benefits that the cryosphere provides to human society, including the resources, products, and benefits obtained directly or indirectly from the cryosphere system (Xiao et al. 2015). Therefore, glacier water resource services can be defined as the benefits that glacier water resources provide to human society, including resources, products, and benefits. Previous research related to glacier water resource services has mainly concentrated on the functional regionalization of glacier water resource services (Liu et al. 2018; Lin et al. 2019) and valuation of glacier services (Zhang et al. 2009; Zhang et al. 2019b; Sun et al. 2021a), and the corresponding research results in other fields are very limited. Research on cryospheric services has produced a series of results. Although these studies still have many problems related to scientific understanding, research methods, and uncertainties (Qin et al. 2020), they still have great significance to the study of glacier water resource services in terms of theoretical guidance and methodological references.

Accurate characterization of hydrological processes in glacier basins is a prerequisite for ensuring accurate assessment of glacier water services. In this area, many studies have emerged, especially the work on simulation and prediction of glacier runoff based on the cryosphere hydrological model, which has produced many fruitful results in recent years. It is found that although the glacier ablation is more intense in the Kyanjing basin in the southern part of the Tibetan Plateau affected by the monsoon, due to the small scale of the glacier, the increased glacier meltwater cannot offset the runoff reduction effect led by the decrease in precipitation. In the Tolle River Basin, which is affected by subpolar westerlies, river runoff increases due to precipitation and glacier melting (Yao et al. 2023). In the Jiemayangzong Glacier basin, the source of the Yarlun Zangbo, the simulation results of the groundwater model HydroGeoSphere suggested that subglacial meltwater recharge to groundwater accounted for about more than 60% of total groundwater recharge, and continuously increasing precipitation recharge to groundwater in the future could not counteract the loss of subglacial meltwater recharge to groundwater (He et al. 2023). Recently, a developed semi-distributed glacio-hydrological conceptual model (FLEX^G) was successfully applied in the simulation of glacier mass balance and runoff in the Torne River Basin, northern Sweden, and obtained better simulation results than the hydrological HBV (Mohammadi et al. 2023). Machine learning techniques have been applied to glacier runoff simulation. In the Manas River Basin of Central Asia, the challenge of accurately simulating and predicting glacier and snow runoff processes, especially simulating flood events, has been well addressed by combining physically based hydrological model and machine learning techniques (Ren et al. 2023). A recent study combined the conceptual hydrological model with the time series predictor model and optimization-driven parameter tuning of the firefly algorithm, uniquely captured the complex interplay between meteorological variables, glacier processes, and hydrological responses, and obtained better simulation results in the glacier and snow runoff (Mohammadi et al. 2024). These aforementioned studies provide a solid theoretical foundation and good data support for the accurate assessment of glacier water resource services.

The study improves the existing method of calculating the LWS, and the improved method considers the impact of glacier runoff. Based on the calculated water withdrawal quantities (WWQs) of the socioeconomic systems in the YRSB and SRB from 2,000 to 2030, as well as the LWS values of these two basins, the sustainability of the water resources in the two basins were evaluated, and corresponding strategies for sustainable development in these two basins were identified. The results of the study provide a reference for the assessment of water resource sustainability in glacial basins worldwide (Figure 1, 2, 3).

The Shule River originates in the Qilian Mountains, northeastern QTP. The area of the SRB (92°11′–98°30′E; 38°00′–42°48′N) is about 5.28 × 10⁴ km². The upstream area of the SRB mainly contains mountain valleys at the western end of the Qilian Mountains, while its midstream and downstream areas are in the western section of the Hexi Corridor, and the terrain in these regions is flat. The SRB is the second-largest inland river basin in the Hexi Corridor, situated in northwestern China. The runoff of the mountain river is influenced by the snow cover, frozen ground, and glaciers. The average temperature per year ranges from 6.9 to 8.8 °C, while the precipitation averages between 47 and 63 mm. In addition, the mean pan evaporation averages from 2,897 to 3,042 mm per year (Wang et al. 2021). Therefore, the SRB is subject to a standard continental drought climate. There are 510 glaciers in the SRB, with a glacier coverage area of 409 km², accounting for about 0.8% of the basin area, total glacier reserves of about 134 km³, an equivalent water volume of 1500 × 10⁸ m³, and a glacier melting volume of about 1.65 × 10⁸ m³ (Figure 1) (Gen & Geocryology 2001; Liu et al. 2015). The average annual amount of runoff in the SRB was about 10.48 × 10⁸ m³, including an average annual amount of glacial runoff of about 2.64 × 10⁸ m³, and glacier runoff accounted for 25.2% of the total annual runoff in the SRB. The SRB in northwestern China is representative of GGAAs (Wang et al. 2021). In the global arid areas, ‘If there is water, there is an oasis. Without water, the oases cannot survive and become desertified.’ Water is the most critical controlling factor for ecosystems in arid inland watersheds (Cheng et al. 2014). Numerous research endeavors concerning agriculture, industry, ecology, and hydroelectric water usage have been implemented in the SRB, and the outcomes of these investigations hold paramount significance for the region's sustainable socioeconomic advancement. As of 2017, the region presented a permanent population of approximately 315,200 individuals, an urbanization rate of 38.50%, and an annual GDP of 20 billion yuan (Table 1). The proportion of the three industries was 9.27:52.43:38.48. Given the substantial migration of the population, the urbanization rate within the basin has transitioned from its prior regression to a gradual increase throughout the last two decades.

The YRSB has a single industry centering around animal husbandry, while the SRB has a variety of industries. Although irrigation agriculture has been developed in the SRB, there is also a high level of industrial development. The level of socioeconomic development in the YRSB is lower than that in the SRB. From 2000 to 2018, the population in the YRSB increased by 2.7%, which was much higher than that in the SRB (0.6%), but the population density was still much lower than that in the SRB (less than 1/3 that in the SRB). The area of sown crops in the YRSB was small, accounting for about 0.04% of the total basin area, whereas that in the SRB was large, accounting for 0.6–1.4% of the total basin area. The differences in the socioeconomic indicators in the two basins directly lead to obvious differences in the water consumption of their socioeconomic systems, which is also the main reason for the huge difference in the water stress in these two basins.

Freshwater resources are what we usually call water resources that are directly utilized on land. They are composed of the water in rivers and lakes, mountain snow, glaciers, and groundwater. In arid and semi-arid areas, precipitation provides little recharge to the groundwater, and therefore, the sustainable exploitation of groundwater is limited. Because there is continuous mutual transformation between surface water resources and groundwater resources, there is very little nonduplicative quantity between groundwater resources and surface water resources in the YRSB and SRB (http://slt.gansu.gov.cn/xxgk/gknb/, http://slt.qinghai.gov.cn/), so the freshwater resources mainly consist of surface water in this study.

The water resource data used in this study included runoff data and water consumption data for the YRSB and SRB. Among them, the runoff simulation data were the output results of the Variable Infiltration Capacity-Chinese Academy of Sciences (VIC-CAS) model of Zhao et al. (2019b) and Zhang et al. (2019a). The input parameters and model data included the following: meteorological data from 52 national meteorological stations on and around the QTP, soil data, digital elevation model data, monthly river runoff data collected at the Changma and Zhimenda hydrological stations (Figures 1(b) and 1(c)) between 1971 and 2013, 1 km resolution global surface coverage data products, glacier distribution data from the Second Chinese Glacier Inventory, and the downscaling scheme of the data outputs from eight Coupled Model Inter-comparison Project five (CMIP5) global climate models (GCMs) under the representative concentration pathway (RCP) 2.6 and 4.5 (Supplementary Table 1). The runoff simulation data have been scientifically verified in many studies (Zhang et al. 2019a; Zhao et al. 2019b; Wang et al. 2021). We calculated the Nash–Sutcliffe efficiency coefficient (NSE) and correlation coefficient (R²) to evaluate the model simulation effectiveness using the observed runoff data of the two basins and the runoff simulation results of the VIC-CAS model. The model achieved satisfactory results in terms of NSE and R², with NSE of 0.77 and 0.89 and R² of 0.81 and 0.89 for the two basins in the validation period and NSE of 0.76 and 0.82 and R² of 0.81 and 0.86 for the two basins in the calibration period.

The results of the previous studies showed that the sources of uncertainty in the runoff data include (1) different model parameters and (2) GCMs, emission scenarios, and downscaling methods used (Chiew et al. 2010; Huss et al. 2014). In this study, the uncertainties associated with the VIC model soil parameters and downscaling methods were much smaller than those of climate change projections (Zhao et al. 2019b). In this study, we compared the predictions from 19 GCMs and selected the eight GCMs that were closest to the average of the 19 GCMs. The range of predicted total annual runoff provided by these GCMs was large, and under the RCP 2.6 (RCP 4.5) scenario, in the YRSB and SRB, relative to the average of the total runoff provided by all the eight GCMs, the mean annual total runoff provided by the different GCMs varied from −17.2 to 13.5% (−14.8 to 19.4%) and −22.7 to 24.2% (−26.4 to 25.6%) and from −21.7 to 17.3% (−22.6 to 14.6%) and −23.2 to 23.8% (−13 to 13.3%) for the 2020s. To attenuate the negative effect of the uncertainty introduced by the GCMs on the results of this article, we conducted the study using the average of the predicted annual total runoff provided by the eight GCMs in the RCP 2.6 and 4.5 scenarios.

The environmental flow requirement (EFR) is important to maintaining the sustainability of a basin's natural ecosystem and human development. Considering the available data, in this study the hydrology method was used, namely, the minimum monthly average flow method (Q90), to calculate the EFR (Wang et al. 2021). Based on the actual runoff data for 2000–2010, the average annual EFRs in the YRSB and SRB were calculated to be 19.11 × 10⁸ m³ and 4.20 × 10⁸ m³, accounting for 12.58 and 40.60% of their average annual surface runoff, respectively. Due to the reduction in glacial runoff, the basins' ecological base flows have exhibited downward trends. During 2020–2030, their EFRs will reach 16.80 × 10⁸ m³ and 3.84 × 10⁸ m³, respectively.

The WWQ involves many indicators of the water resource system. The WWQ is composed of the primary industry water demand (PWD), the secondary industry water demand (SWD), the tertiary industry water demand (TWD), the urban and rural residents' domestic water demand (RWD), and the artificial ecosystem water demand (AEWD). In this study, the WWQ between 2000 and 2017 was obtained from the 2000 to 2018 Water Resources Bulletin of Gansu and Qinghai (http://slt.gansu.gov.cn/xxgk/gknb/, http://slt.qinghai.gov.cn/).

Natural resources are scarce. The premise of value accounting is that it can bring economic benefits and be measured reliably (United Nations 2012; El Serafy 2015). Glaciers are an important component of natural resources. As an important freshwater reservoir, glacier water resources undoubtedly provide economic benefits to humans, including a static frozen ice body and dynamic glacial meltwater. Static glacial solid ice produces considerable economic benefits including tourism and other aspects (Yuan & Wang 2018; Wang et al. 2020), while dynamic glacial meltwater also produces huge economic benefits including agricultural irrigation, supplying fresh water, water conservancy, and power generation (Wang 1996; Wang et al. 2021). Therefore, glacial water resources are an important component of water resource assets and can be included in natural resource assets (Sun et al. 2021b). In contrast, some cryospheric elements do not meet the strict definition of natural resource assets, such as snow water resources (El Serafy 2015). Some cryospheric elements are rarely directly used by humans and even cause disasters, such as ice avalanches, avalanches, and hail (Wang & Xiao 2019). Basic data for some cryospheric elements are lacking, such as data for frozen soil water resources (Zhao et al. 2019a).

Overall, the runoff in the upstream area of the YRSB is 10 times that in the SRB. Under the context of global warming, increasing precipitation leads to an upward trend in the total upstream surface runoff, while the glacier runoff changes little during 2000–2030. From 2000 to 2010, the average annual amounts of runoff in the YRSB and SRB were about 134.02 × 10⁸ m³ and 10.48 × 10⁸ m³, respectively, including average annual amounts of glacial runoff of about 5.54 × 10⁸ m³ and 2.64 × 10⁸ m³, respectively. From 2000 to 2010, glacier runoff accounted for 25.2% of the total annual runoff in the SRB, while it only accounted for 4.1% in the YRSB (Figure 4, left and right).

Under the RCP 2.6 (RCP 4.5) scenario, surface runoff in the SRB will maintain an increasing trend from 2010 to 2030. The surface runoff in the YRSB will maintain an increasing trend during 2010–2025; but after 2025, the surface runoff will decrease slightly. The variation trends of the available freshwater resources in these two basins are different due to the difference in the runoff variations. Forecasts for 2020–2030 have indicated that compared with 2000–2010, the average runoff in the YRSB and SRB will increase by 27% (5.9%) and 25.1% (22.5%), respectively, whereas the average glacial runoff will decrease by 3.7% (8%) and 11.7% (9.8%), respectively (Figure 4, left and right).

From 2000 to 2018, the WWQ in the YRSB decreased from 0.18 × 10⁸ m³ in 2000 to 0.16 × 10⁸ m³ in 2018, whereas in the SRB, it increased from 5 × 10⁸ m³ in 2000 to 8.39 × 10⁸ m³ in 2018. Under the current trends of the climate and economic changes, during 2019–2030, the WWQ in the YRSB will gradually decrease, while in the SRB, it will continue to increase (Figure 5, left and right). By 2030, the WWQ of the two watersheds will reach 0.18 × 10⁸ m³ and 12.01 × 10⁸ m³, respectively. In the YRSB, the PWD significantly decreased from 2003 to 2030 mainly because of water use efficiency improvements and enhanced ecological protection. In contrast, the PWD in the SRB continuously increased from about 5.27 × 10⁸ m³ in 2000 to 10.99 × 10⁸ m³ in 2030 due to the continuous expansion of the farmland area and the increased water demand for irrigation (Ma et al. 2019).

In the YRSB, under the RCP 2.6 (RCP 4.5) scenario, the LWS only increases by an average of 0.000 1 per year if we subtract the glacial runoff from the available freshwater resources during 2000–2030. If the proportion of glacial runoff in the total runoff continues to decrease from 4.00% in 2000–2017 to 3.20% in 2018–2030, the inhibitory effect of the glacial runoff on the LWS will decrease slightly, which also shows that glacial runoff has very little influence on the LWS in the YRSB (Figure 6, left). The majority of glacial water resources in the YRSB flow to downstream areas, the contribution of the glacial water resources to the water stress in a basin is small, but it makes a large contribution to the water resources across basins. In the SRB, under the RCP 2.6 (RCP 4.5) scenario, if the glacial runoff is deducted from the available freshwater resources, the LWS in the SRB rises by an average of 0.71(0.68) annum (Figure 6, right). This reflects the regulating service of glaciers on the balance of regional water resource supply and demand. A decrease in the proportion of glacial runoff in total runoff from 23.86% in 2000–2017 to 18.47% in 2018–2030 significantly diminishes the inhibitory effect of glacier runoff on LWS, exacerbating the future water crisis. Regardless of glacier runoff, SRB will struggle to achieve the water supply and demand balance target of SDG 6.4.2 by 2030.

Based on the aforementioned analysis, glaciers are crucial to the sustainable development of the water resources in these two basins. The spatial flow of the glacier water resources and the related services are the basis of water resource compensation, and the principle of compensation is that the user and beneficiary/owner pay. The value of the glacier water resources and their total service supply is the maximum amount of compensation, and the theoretical amount is based on the actual use of the glacier water resources. In the source region of the YRSB, the glacier water resources flow to the middle and lower reaches of the Yangtze River, mainly producing benefits there, while the upper reaches receive little benefit. Therefore, the compensation of the glacier water resources in the source region of the YRSB should follow the outside-in compensation mode and the national vertical compensation mode, that is, the inter-basin compensation mode from outside the source region to the source region.

The beneficiaries of the glacial water resources in the SRB include the upper, middle, and lower reaches. Among them, the middle and lower reaches are the main beneficiary areas. The glacial water resources and their services are mostly consumed within the basins. According to the principle of user pays and beneficiary pays, the compensation of glacier water resources in the SRB should mainly follow the horizontal compensation from the middle-lower reaches of the basin to the upper reaches. For the aforementioned two types of basins, in addition to clear compensation subjects, for the other beneficiary areas that are difficult to define, the compensation amount for the glacier water resources should be borne by the state (Xu et al. 2019). Generally, the inter-basin compensation of the glacial water resources in the source region of the YRSB should be adopted to ensure the sustainable utilization of the basin's water resources and to improve the ability to achieve synergistic ecological environmental protection and social and economic development. Internal regulation and compensation are mainly used to improve water resource sustainability and to promote sustainable development of the social and economic system in the SRB.

By comparing the two basins, we identified several characteristics. As the most important ecological barrier area (Sun et al. 2012) and restricted development area in China (Fan 2015), the level of socioeconomic development in the YRSB was very low, its industrial structure was simple, and its proportion of animal husbandry was very high. The level of socioeconomic development in the SRB was higher than that in the YRSB, its industry was varied, and its irrigated agriculture was developed. The WWQ in the YRSB was very small, and it was mainly composed of the water demand for animal husbandry and the RWD. The WWQ in the SRB was very large, and the water demand for irrigated agriculture accounted for the majority of the WWQ. The LWS in the YRSB was very low, and the glacier runoff only reduced the LWS in the YRSB slightly. However, in the SRB, the LWS was very high, and the glacial runoff reduced the LWS significantly.

Based on these results, as typical representatives of inland basins in northwestern China's arid region and the upstream areas of many large rivers in China, the glacier water services in these two basins were significantly different. The upstream areas of the large rivers in China represented by the YRSB are important supply areas for the ecosystem services in China, their level of socioeconomic development is low, and their WWQs are limited. Affected by governmental policies, restrictions, and other factors, these basins are unlikely to contain large-scale economic development activities in the future, so their WWQs will be maintained at a very low level. Correspondingly, the sustainability of the water resources in these basins will always be at a very high level, and they are the output areas of the glacier water resources and their services. The inland basins in northwestern China's arid region represented by the SRB have a high level of socioeconomic development, which results in their WWQs being at a high level. Therefore, the sustainability of the water resources in these basins will remain at very low levels, and most of their glacier water resources and their services are mainly consumed within the same basin.

The inland basins in northwestern China's arid region represented by the SRB face the problem of high LWS values, which are mainly caused by irrational water withdrawal, such as inefficient and excessive withdrawal of the basins' water resources. Therefore, for such basins, internal control measures should immediately be adopted to control the PWD, especially to control the water withdrawal for farmland irrigation, to guide the agricultural population transfer a transition to non-agricultural industries, resulting in a lower water demand and more water-efficient industries. This plan would improve the water resource-carrying capacities of these basins and support the sustainable development of the socioeconomic systems in these basins (Wang & Wei 2019). The upstream areas of the great rivers in China represented by the YRSB are the supply regions of glacier water resources and their services. Most of their glacier water resources and services flow to their midstream and downstream areas and support the socioeconomic development of their midstream and downstream areas. According to the Chinese law, the property rights of natural resources belong to the country and are supervised and used by local governments or relevant departments, and local residents have the right to use them (Diao 2011). Therefore, these basins have experienced the phenomenon of the loss of the right to use the glacier water resources due to the spatial flow of the glacier water resources and their services, as well as the phenomenon of deviation between the benefits and costs (benefits: the benefits provided by the glacier water in these basins to the entire society; costs: the social cost required to maintain the stability of the quantity and quality of the glacier water resources in these basins). A cross-basin compensation system should be implemented in these basins to make up as much as possible for the loss of glacier water resources and their related cost deviations and to improve the ability to achieve synergistic ecological environmental protection and socioeconomic development in the QTP water tower areas such as the upstream areas of great rivers.

Previous studies have shown that under the background of global warming, the cryospheric elements in the alpine mountainous regions of western China have undergone significant changes. In particular, in recent decades, mountain glaciers have rapidly melted, which has led to significant changes in the river runoff in the mountainous basins of western China (Ding & Zhang 2015; Ding & Zhang 2018; Chen et al. 2019; Ding et al. 2020). These changes were specifically manifested as follows. In the basins in the western part of the Heihe River Basin, the runoff increased, and the drier the basin, the greater the proportion of glacier runoff. In the basins in which the number and areas of small glaciers are large, the glacier runoff and its proportion have continued to decrease. In the basins in which the number and areas of small glaciers are large, the glacier runoff and its proportion initially increased and then decreased. In some basins, the peak of the glacier runoff has or is about to occur (Ding & Zhang 2015; Ding & Zhang 2018; Chen et al. 2019; Ding et al. 2020). After 2030, in some river basins of western China, especially the arid regions in northwestern China, their surface water resources will continue to decrease, their glacier water resource services will continue to decrease or even disappear completely, and the problem of the increasing LWS will gradually occur. Similarly, the impacts of accelerated glacier melt on runoff are being shown in some glacier basins in other regions of the planet. In the Torne River Basin, located in the north of Sweden, in northern Europe, the simulation results from the FLEX^G model show that glaciers and their changes have had a significant impact on the runoff in the basin from 1989 to 2018, with an upward trend in simulated runoff and a negative and decreasing glacier mass balance (Mohammadi et al. 2023). In the Nooksack River Basin of the United States, the continued loss of glacier area has resulted in reduced runoff and increased water temperatures, potentially threatening local salmonid populations (Pelto et al. 2022). In the Naryn Basin, Central Asia, the results indicate a decrease in ice volume by −15% for the time period 1982–2019, and in the Small Naryn catchment, a significant increase in glacier melt and meltwater contribution to runoff occurs (Saks et al. 2022). In the western Juneau Icefield, Southeast Alaska, from 1980 to 2016, runoff of the basin is still in an increasing glacier runoff period prior to reaching ‘peak water’. Although higher proportions of precipitation have controlled the runoff change of the area, it is experiencing hydrological regime change driven by ongoing glacier mass loss (Young et al. 2021). It is urgent that various measures be implemented to reduce the various negative impacts brought about by the reduction of glacier water resources and the weakening of their services to improve the adaptability of global residents to climate change and its impacts.

The upstream areas of the great rivers on the QTP are the debt region for glacial resources, and this debt benefits the midstream and downstream areas. Therefore, it is necessary to establish a national or downstream-led water resource compensation mechanism. Based on the changes in the flow of the glacier water resources, the river basin water resources on the QTP should be managed comprehensively, and the source area should be appropriately compensated for the loss of glacial water resources.

The proportion of glacial runoff in the YRSB is very small and will continue to decrease in the future. The glacial runoff in the SRB accounts for a large proportion, and it may increase in the next few years. The level of social and economic development in the YRSB is very low, the industry is singular, and the proportion of animal husbandry is very high. The SRB has diversified industries and relatively developed irrigation agriculture. The water consumption of the social and economic systems in the YRSB is very small and is mainly used for animal husbandry and residential life. The water consumption of the social and economic system in the SRB is very large, and the water consumption for farmland irrigation accounts for the majority. The water stress in the YRSB is very low, and the glacial water resources have a very weak alleviating effect on the water stress, while the water stress in the SRB is very high, and the glacial water resources have a large alleviating effect on the water stress.

Most of the glacier water resources and services in the SRB are consumed within the basin, and its sustainable development strategy should focus on internal control. In the future, the improvement of irrigation efficiency and efficient water-saving measures for use in agriculture will be the key measures needed to reduce the LWS and ensure sustainable development in the SRB. The majority of glacial water resources and their services in the YRSB flow to the downstream areas and a cross-basin water compensation system should be implemented to enhance the coordination of ecological environmental protection and social and economic development.

Overall, the YRSB is rich in freshwater resources, while these resources are very limited in the SRB. The glacial water resources and their services differ significantly in the source regions of major rivers and their inland river basins in arid regions. The YRSB and other major rivers have high water resource sustainability and are the supply and export regions of glacier water resources and their services. The sustainability of the water resources in the SRB and even in other inland river basins in arid areas is very low, and the glacial water resources and their services are mainly supplied from within the basin.

This research was supported by the Key Science and Technology Project of Gansu Province (22ZD6FA005) and the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA19070503). The authors wish to thank the editors and the anonymous reviewers for all their helpful discussions and advice. The authors also would like to thank Dr Yan Xingguo, Dr Cheng Wenju, Dr Wu Rui, Dr Ji Qin of the Northwest Institute of Ecology and Environmental Resources, Chinese Academy of Sciences, Dr Li Yao of Sun Yat-sen University, and Dr Zhou Lanyue of the University of Munich for their many helpful suggestions during the revision of this paper.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.