Climate Background, Relative Rate, and Runoff Effect of Multiphase Water Transformation in Qilian Mountains, the Third Pole Region

Science of the Total Environment 663 (2019) 315–328

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Climate background, relative rate, and runoff effect of multiphase water transformation in Qilian Mountains, the third pole region

Zongxing Li a,⁎,RuifengYuana, Qi Feng a,⁎, Baijuan Zhang a, Yueming Lv a,YonggeLia,WeiWeic, Wen Chen b, Tingting Ning a,JuanGuia, Yang Shi a a Key Laboratory of Ecohydrology of Inland River Basin/Gansu Qilian Mountains Eco-Environment Research Center, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Lanzhou 730000, China b Gansu Provincial Hydrographic Resources Bureau, Lanzhou 730000, China c Research Center for Eco-Environment Sciences, Chinese Academy of Sciences, Beijing 100085, China

HIGHLIGHTS GRAPHICAL ABSTRACT

• Finding the lengthening ablation period and the larger warming in cryosphere belt. • Glaciers area retreat rate has accelerated by 50% after 1990. • The percent of snowfall accounting for precipitation has decreased by 7% after 1990. • Contribution from the recycling moisture to precipitation has increased by 60%. • The outlet runoff increased and seasonal pattern changed.

article info abstract

Article history: Multiphase water transformation has great effects on alpine hydrology, but these effects remain unclear in the third Received 18 September 2018 pole region. Taking the Qilian Mountains as an example, the climate background and relative rates of multiphase Received in revised form 7 January 2019 water transformation were analyzed, and the runoff effect was evaluated based on long-term field observations. Accepted 25 January 2019 There are three climatic aspects driving multiphase water transformation, including lengthening ablation period, accel- Available online 28 January 2019 erative warming after 1990, and larger warming in the cryosphere belt than in the vegetation belt. The accelerative fi Editor: Ralf Ludwig multiphase water transformation was quanti ed by three facts: the glacier area retreat rate accelerated by 50% after 1990, the percentage of snowfall in precipitation decreased by 7% after 1990, and the contribution from recycling mois- Keywords: ture to precipitation increased by 60% from 1961–1990 to 1991–2016. Under the multiphase water transformation, the Multiphase water transformation outlet runoff for three inland rivers increased by 5 × 108 m3/10 a after 1990. This runoff increase was concentrated Climate warming mainly in the ablation period. For the seasonal runoff pattern, maximum runoff lagged maximum precipitation by Cryosphere one month under increasing glacier snow meltwater and thickening permafrost active layer. Meltwater from the Runoff effect cryosphere is a crucial runoff component in the Qilian Mountains. At present, these multiphase water transformations Qilian Mountains are accelerating, along with the yearly runoff increase, which will obviously have a profound impact on water resources management and flood control in the third pole region. © 2019 Elsevier B.V. All rights reserved.

⁎ Corresponding authors at: Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, China. E-mail addresses: [email protected] (Z. Li), [email protected] (Q. Feng). https://doi.org/10.1016/j.scitotenv.2019.01.339 0048-9697/© 2019 Elsevier B.V. All rights reserved. 316 Z. Li et al. / Science of the Total Environment 663 (2019) 315–328

1. Introduction area, during the past several decades (Cheng and Jin, 2013), which has changed the water cycle and yearly runoff variation (Yao et al., 2013). In the third pole regions, the prominent hydrological feature is the Based on continuous observations from 8 stations along the Qinghai– existence of multiphase water and its transformations (Li et al., Tibet road, the average permafrost active layer depth increased from 2019a). Solid water reserves include glaciers, snow, and ground ice, liq- 252 cm in 2006 to 276 cm in 2011, with an average rate of increase of uid water mainly includes river water, lake water, marsh water, soil 4.7 cm/a (Liu et al., 2014; Li et al., 2019a). Data from 16 stations in Qing- water, plant water, and groundwater, and gaseous water includes hai Province showed that the maximum freezing depth decreased con- local recycling vapor and advection vapor (Fig. 1). At present, warming tinuously by 4.8 cm/10 a during 1961–2001, while the average depth of the climate system is significant, and global annual mean tempera- decreased from 144 cm in 1961–1970 to 124 cm in 1990–2001 (Wang ture has increased by 0.85 °C in the last 100 years (1880–2012) (IPCC, et al., 2005; Li et al., 2019a). In the Tibet autonomous region, the average 2013). Almost all glaciers worldwide have continued to shrink, as re- maximum freezing depth measured at 17 stations also decreased with a vealed by time series of measured changes in glacier length, area, vol- rate of 5.5 cm/10 a during 1961–2010. The rate of decrease increased ume, and mass, and the total mass loss from all glaciers in the world, after 1990; the freezing depth decreased by 14 cm from 1961–1990 to excluding those on the periphery of ice sheets, was very likely 226 ± 1991–2010, while the start date of thawing advanced with a rate of 135 Gt yr−1 in the period 1971–2009 (IPCC, 2013). Satellite records in- 2.1–5.2 d/10 a during 1971–2010 (Du et al., 2012; Li et al., 2019a). dicate that annual mean snow cover extent also decreased with statisti- In the Qilian Mountains, many studies have also confirmed the sig- cal significance over the period 1967–2012 (Bulygina et al., 2009). nificant warming based on records of tree rings and metrological sta- Permafrost temperatures have increased in most regions since the tions (Chen et al., 2012; Xu et al., 2014; Wang et al., 2018), and early 1980s, although the rate of increase has varied regionally, and ac- precipitation increase has also been demonstrated, especially during re- tive layer thicknesses have increased by a few centimeters to tens of cent decades. Under climate warming, the area of 244 glaciers in the centimeters in many areas since the 1990s (Callaghan et al., 2011). eastern Qilian Mountains had decreased by 24.286 km2 in 2007 com- Annual mean temperature also increased by 0.9–1.5 °C in China in pared to the area in 1972, accounting for 23.57% of the glacierized the last 100 years (Ding and Wang, 2016), and temperature increased area in 1972, and 27 glaciers had disappeared by 2007 (Cao et al., by about 1.80 °C during 1960–2007 in the cold regions of western 2010). In the middle Qilian Mountains, glacier area shrank significantly China (Ding and Wang, 2016), where the major cryosphere in China, in- by a total of 138.9 km2, accounting for 35.6% of the glacier area in 1960 cluding all glaciers and most of the permafrost and stable snow cover, is (Bie et al., 2013). In the western Qilian Mountains, the glacier area and distributed (Chen et al., 2018). Under this change, multiphase water ice volume decreased by 17.21% and 24.1% from 1957/1966 to 2010, re- transformation is becoming progressively more accelerative (Kääb spectively (Yu et al., 2014). Additionally, the average decrease of glacier et al., 2007; Kabel et al., 2012; Malatinszky et al., 2013), which is charac- thickness in the Qilian Mountains from 2000 to 2010 was 5.68 ± 2.76 m, terized by the rapidly shrinking cryosphere. In western China, glacier the average glacier mass balance was −0.48 ± 0.23 m w.e·a−1, and the area decreased by 18% from 1970 to 2010, and 5797 glaciers have disap- change of glacier volume was −1.59 ± 0.72 Gt (Gao et al., 2018). Be- peared according to the Second Chinese Glaciers Inventory (Liu et al., tween 1957 and 2007, the Laohugou No. 12 glacier in the western Qilian 2015). Glacier thickness also decreased by 10.56 m between 1980 and Mountains experienced significant thinning and areal shrinkage in the 2005 (China Meteorological Administration, 2006). Ren et al. (2011) ablation zone, the elevation decreased by 18.6 ± 5.4 m, and the total confirmed that the average yearly meltwater runoff from western volume loss for the entire glacier was estimated to be 0.218 km3 China increased from 51.8 × 109 m3 in the 1960s to 79.5 × 109 m3 be- (Zhang et al., 2012). Based on the altitude-response model, the areas tween 2001 and 2006, which contributed 0.12 mm/a to the global of simulated permafrost distribution in the Qilian Mountains in the mean sea level rise. Permafrost showed widespread degradation, in- 1970s, 1980s, 1990s, and 2000s are 9.75 × 104 km2,9.35×104 km2, cluding increased temperature, deeper active layer depth, and reduced 8.85 × 104 km2, and 7.66 × 104 km2, respectively, showing a decreasing

Fig. 1. Sketch map for multiphase water transformation in the Qilian Mountains. Z. Li et al. / Science of the Total Environment 663 (2019) 315–328 317 trend over the past 40 years (W. Zhang et al., 2014; L. Zhang et al., 2014). The eastern branch is the Heihe River and the western branch is the Under climate warming, the combination of increasing evapotranspira- Taolaihe River. The Shulehe River Basin, mainly including the tion and diminishing permafrost area has resulted in smoother and flat- Changmahe River and Danghe River, is the third largest inland river ter hydrographs and a reduction in total river discharge, which confirms basin in China, with a whole catchment area of approximately 14.21 the influences of permafrost and climate change on hydrological pro- ×104 km2. The Shiyanghe River Basin, including seven branches, is cesses in the Qilian Mountains (Zhang et al., 2017). Atmospheric the fourth largest inland river basin, with an area of 4.16 × 104 km2. water vapor in the Qilian Mountains showed an increasing trend during The annual mean runoff for the Shiyanghe River, Heihe River, and 1979–2016 (Gong et al., 2017). At present, these multiphase water Shulehe River is 1.6 billion m3, 3.2 billion m3, and 1.72 billion m3, re- transformations, along with the yearly runoff, are accelerating, which spectively. As the middle stream of three inland rivers, the Hexi Corri- obviously has a profound impact on hydrologic processes and the dor, located in the north of the Qilian Mountains, includes five cities: water cycle and especially on water resources management and flood Wuwei, Zhangye, Jiuquan, Jinchang, and Jiayuguan. The corridor covers control in alpine basins. an area of 271,100 km2 and crosses an arid to semiarid part of China. It Due to the high altitude and harsh environments in the third regions, has an annual precipitation of 50–200 mm, with precipitation decreas- observation data are very limited, resulting in large uncertainties in cold ing from east to west. Annual potential evaporation increases from region hydrology (Li et al., 2016a, 2016b, 2016c, 2019a). Therefore, N2000 mm in the east to 3500 mm in the west. many researchers have focused on the statistical relationships between cryosphere change and runoff variation (Ye et al., 2003; Niu et al., 2011). 2.2. Data However, cryosphere change is only one part of the multiphase water transformation, which has changed the water cycle and river discharge The snowfall, daily precipitation, temperature, maximum tempera- in cold regions, and particular attention has not been paid to under- ture, and minimum temperature data were provided by the National standing the effects of these transformations. Relative to other cold re- Climate Center, China Meteorological Administration (CMA) (available gions of western China, there are good observation data on the from http://www.nmic.gov.cn/). The modern nationwide network of cryosphere, meteorology, and hydrology in the Qilian Mountains for weather observing stations of China began operation in the 1950s, and the past decades, and the most adequate research results and observed 24 of the 30 stations in the original data set have maintained daily data have been chosen for studies of multiphase water transformation. data since 1961. In this study, 6 of these stations were excluded because Hence, the climate background for multiphase water transformation of data quality problems based on the quality control method from a in the Qilian Mountains has been analyzed. In detail, the relative rate previous study (Li et al., 2012). The time span of the meteorological of the accelerative multiphase water transformation, including glacier data was mainly set from January 1, 1961, to December 31, 2017, to en- change, snowfall change, and moisture recycling change, in the Qilian sure that the length of all data was uniform and stable. In the end, 24 Mountains has been quantified. Additionally, the runoff effect, which meteorological stations remained (Table 1). These stations, which is related to increasing runoff, changing seasonal runoff patterns, and were built before 1961, have data of good quality, relatively even distri- runoff components, has been evaluated under multiphase water trans- bution, and continuous records. Detailed information about the stations formation. This should lead to an overall understanding of the fre- is provided in Table.1. The stations belong to the World Meteorological quency, intensity, and duration of multiphase water transformation Organization (WMO)'s climate data exchange network and each has a and its hydrological effects in third pole regions. WMO number. Also, the indices for FD (frost days), ID (ice days), GSL (growing season length) are from Li et al. (2019b), and R10mm (num- 2. Data and methods ber of heavy precipitation days), R20mm (number of heavier precipitation days), and R25mm (number of heaviest precipitation days) were 2.1. Study region calculated based on the method from the previous study (Li et al., 2012). The daily water vapor flux, wind speed, mean geopotential height, The Qilian Mountains (36–43° N, 92–107° E), with an area of 18.3 wind fields, and relative humidity at 300 and 500 hPa were obtained ×104 km2, lie at the northeastern edge of the third pole regions. The from the National Oceanic and Atmospheric Administration– range consists of several parallel mountains and valleys, extending Cooperative Institute for Research in Environmental Sciences (NOAA– 850 km from northwest to southeast and encompassing a width of CIRES) Climate Diagnostics Center reanalysis R1 dataset (available 200–300 km. About 30% of the mountains in the Qilian Mountains from http://www.cdc.noaa.gov/). This dataset covers the period from have altitudes higher than 4000 m. The highest peak is Tuanjie Peak, January 1948 to present with a spatial resolution of 2.5° × 2.5° and with an altitude of 5826.8 m. Based on the Second Glaciers Inventory, with continuous global coverage (Kalnay et al., 1996; Kistler et al., there are 2859 glaciers, with a total area of 1597 km2, and the total gla- 2001). On the basis of the reanalysis R1 dataset, Liu's (1997) atmo- cier reserve and permafrost area are 84.48 × 108 m3, and 9.39 × 104 km2, spheric hydrological cycle model is used to explore the contribution respectively (Sun et al., 2015). The temperature varies with the altitude from recycling moisture to precipitation in the Qilian Mountains, and change from valley to mountain. The annual precipitation exhibits a de- the data during 1961–2010 is from W. Zhang et al. (2014) and L. creasing trend from east to west and from south to north. The Qilian Zhang et al. (2014) and during 2011–2016 has been calculated in this Mountains are the source region of three inland river basins, the study. The runoff data for the inland rivers (shown in Table 2)are Shiyanghe, Heihe, and Shulehe, from southeast to northwest, respec- from the Gansu Provincial Hydrographic Resources Bureau (http:// tively, as shown in Fig. 2. The river source region is the cryosphere www.gssl.gov.cn/slzx/tzdt/sswszyj/). The glacier area data for 1958 belt, which is distributed with glaciers, snow, and permafrost. This up- are from the First Chinese Glacier Inventory (Wang et al., 1981), those stream mountainous region (altitude of 2000–5000 m) in the Qilian for 1990 are from Tian (2013), and those for 2010 are from the Second Mountains has a cold semi-arid to semi-humid climate. The annual Chinese Glacier Inventory (Sun et al., 2015). mean air temperature is b2 °C, and the annual precipitation increases from about 200 mm in the low-mountain zone to about 500 mm in 3. Results the high-mountain zone. The ecosystem changes vertically, moving through dry shrubbery grassland, forest grassland, sub-alpine shrub- 3.1. Climate background for multiphase water transformation bery meadow, alpine cold-desert meadow, and alpine permafrost and glaciers from low altitude to high altitude. 3.1.1. Lengthening ablation period The Heihe River Basin, with an area of 14.29 × 104 km2, including the In the Qilian Mountains, annual average temperature increased at river's two branches, is the second largest inland river basin in China. 0.35 °C/10 a during 1961–2016 (Li et al., 2019b), and the amplitudes 318 Z. Li et al. / Science of the Total Environment 663 (2019) 315–328

Fig. 2. Study region. of increase of warming increased from east to west, with larger (Fig.3), and larger increases occurred mainly in the eastern and western warming at higher altitude regions (Figs. 3 and 4). This caused a gradual Qilian Mountains (Fig. 4). Annual precipitation increase was also signif- extension of the ablation period and a continuous reduction of the freez- icant, with a rate of 14.7 mm/10 a during 1961–2016 (Li et al., 2019b) ing period. In the study region, the ID (annual count when daily maxi- (Fig. 5). The R10mm, R20mm, and R25mm (annual count of days with mum temperature b 0 °C) decreased signiﬁcantly by 3.3 d/10 a during daily precipitation ≥ 10 mm, 20 mm, and 25 mm) increased by 1961–2016, while the FD (annual count when daily minimum 0.4 d/10 a, 0.1 d/10 a, and 0.05 d/10 a during 1961–2016 (Fig. 5), respec- temperature b 0 °C) decreased statistically by 3.86 d/10 a (Li et al., tively, but the amplitudes of increase were relatively lower. For annual 2019b)(Fig. 3). In the spatial pattern, the amplitudes of decrease of precipitation, the largest increases occurred mainly in the middle Qilian these were larger at higher altitude regions (Fig. 4). Meanwhile, the Mountains, and the amplitudes of increase were higher in the eastern GSL increased at the signiﬁcant level of 3.48 d/10 a (Li et al., 2019b) than the western region (Fig. 6). The spatial patterns for R10mm, R20mm, and R25mm were similar to that of annual precipitation

Table 1 (Fig. 6). The selected weather stations in study region. 3.1.2. Accelerative warming after 1990 WMO number Name Latitude Longitude Altitude (m) Under global warming, especially in northwestern China, the climate 52418 Dunhuang 40.09 94.41 1139.0 was basically characterized by the development of warm-dry conditions 52424 Anxi 40.32 95.46 1170.9 from the end of the Little Ice Age until the 1980s, but a climatic mutation 52436 Yumenzhen 40.16 97.02 1526.0 52447 Jinta 40.00 98.54 1270.5 to warm-wet conditions occurred around 1990 (Shi et al., 2003). In the 52533 Jiuquan 39.46 98.29 1477.2 Qilian Mountains, the warming amplitude for annual mean temperature 52546 Gaotai 39.22 99.50 1332.2 in 1991–2016 is two times higher than that in 1961–1990 (Fig. 7). Cor- 52652 Zhangye 39.05 100.17 1461.1 respondingly, the decreasing trends for ID and FD are 1.6 d/10 a and 52661 Shandan 38.48 101.05 1764.6 – – 52674 Yongchang 38.14 101.58 1976.9 4.7 d/10 a higher in 1991 2016 than in 1961 1990 (Fig. 7). GSL 52679 Wuwei 37.55 102.40 1531.5 displayed an obvious increasing trend during 1961–1990, but it in- 52797 Jingtai 37.11 104.03 1630.9 creased statistically by 5.8 d/10 a during 1991–2016. Meanwhile, 52876 Minhe 36.20 102.50 1813.9 52602 Lenghu 38.45 93.20 2770.0 52657 Qilian 38.11 100.15 2787.4 Table 2 52737 Delingha 37.22 97.22 2981.5 The selected hydrological stations in the study region. 52765 Menyuan 37.23 101.37 2850.0 River basin Station Lon Lat 52856 Gonghe 36.16 100.37 2835.0 52866 Xining 36.44 101.45 2295.2 Changma river Changmabao 96.85 39.83 52868 Guide 36.01 101.22 2237.1 Danghe river Dangchengwan 94.87 39.51 52633 Tuole 38.48 98.25 3367.0 Heihe river Yingluoxia 100.18 38.81 52645 Yeniugou 38.25 99.35 3320.0 Taolaihe river Jiayuguan 98.26 39.75 52713 Dachaidan 37.51 95.22 3173.2 Zamu river Zamusi 102.58 37.70 52754 Gangcha 37.20 100.08 3301.5 Xiying river Jiutiaoling 102.05 37.86 52842 Chaka 36.47 99.05 3087.6 Nanying river Nanyingshuiku 101.52 37.70 Z. Li et al. / Science of the Total Environment 663 (2019) 315–328 319

Fig. 3. Interannual variation of temperature (a), ID (b), FD (c), and GSL (d). precipitation increased by 24 mm/10 a during 1991–2016, which is two 3.1.3. Larger warming in cryosphere belt than in vegetation belt times larger than that during the period of 1961–1990 (Fig. 7). The in- In the Qilian Mountains, the lower boundary of permafrost is at the creasing trend of R10mm during 1991–2016 is 0.1 d/10 a larger than altitude of 3600 m (Zhou et al., 2000), and the region can be divided that during 1961–1990. However, R20mm and R25mm show no in- into two belts according to the permafrost boundary: cryosphere belt creasing trend during 1961–1990, but these do display a signiﬁcant in- (above 3600 m), which is covered by glaciers, permafrost, and snow, creasing trend during 1991–2016. and vegetation belt (below 3600 m), with growing forests and

Fig. 4. Spatial pattern for variation amplitudes of temperature (a), ID (b), FD (c), and GSL (d). 320 Z. Li et al. / Science of the Total Environment 663 (2019) 315–328

Fig. 5. Interannual variation of precipitation (a), R10mm (b), R20mm (c), and R25mm (d). meadows (Fig. 8). As a whole, the amplitudes of warming display an in- temperature in the cryosphere belt is 0.05 °C/10 a higher than that in creasing trend with increasing altitude in the study region, and this can the vegetation belt (Fig. 8). Correspondingly, the ID, FD, and GSL in the be conﬁrmed by the larger warming in the cryosphere belt. These vari- cryosphere belt are 1.5 d/10 a, 1.7 d/10 a, and 0.2 d/10 a greater than ations reﬂect the more obvious extension of the ablation period in the those in the vegetation belt, respectively (Fig. 8). The precipitation in- cryosphere belt. The amplitude of warming of annual mean crease in the cryosphere belt is obviously larger than that in the

Fig. 6. Spatial pattern of variation amplitudes for precipitation (a), R10mm (b), R20mm (c), and R25mm (d). Z. Li et al. / Science of the Total Environment 663 (2019) 315–328 321

belt than in the vegetation belt, respectively. These facts reﬂect the larger warming and more humid climate of the cryosphere belt in the Qilian Mountains.

3.2. Relative rate of the accelerative multiphase water transformation

3.2.1. Accelerated glacier area retreat rate after 1990 In the Qilian Mountains, glaciers are the sources of runoff for the upstream regions of the Yellow River, Hexi inland rivers, and Qinghai inland rivers. Based on the Second Chinese Glaciers Inventory, the region included 2684 glaciers with an area of 1597 km2 and had an ice volume of 84.48 km3 in 2010 (Sun et al., 2015). In 1956 and 1990, the glaciated area was 2017 km2 and 1802 km2, respectively. From 1956 to 2010, glaciers retreated by 420.81 km2 (−20.88%). From further Fig. 7. Variation amplitudes for climate indices during 1961–1990 and 1991–2016. study (Fig. 9), the glacier retreat rates for the upstream of the Yellow River region, Hexi inland river region, and Qinghai inland river region vegetation belt, and the variation trend is also higher by 9 mm/10 a are 2.93 km2/10 a, 41.83 km2/10 a, and 9.74 km2/10 a during (Fig. 8). The increasing trends of R10mm, R20mm, and R25mm are 1956–1990, respectively; however, the corresponding values in also 0.3 d/10 a, 0.17 d/10 a, and 0.1 d/10 a higher in the cryosphere 1990–2010 are 4.4 km2/10 a, 62.74 km2/10 a, and 14.61 km2/10 a,

Fig. 8. Cryosphere belt (above 3600 m) and vegetation belt (below 3600 m) in the Qilian Mountains (a) and variation amplitudes in the cryosphere belt and vegetation belt (b). 322 Z. Li et al. / Science of the Total Environment 663 (2019) 315–328 respectively. Compared with the period of 1956–1990, the glacier re- processes. In the Qilian Mountains, the precipitation sourced by mois- treat rate during 1990–2010 is accelerated by about 50%, reflecting the ture recycling showed a sustentative increasing trend of 5.1 mm/10 a accelerative transformation from solid water to liquid water in the during 1961–2016, and the increase has accelerated since 1990. Addi- Qilian Mountains. tionally, the average precipitation from moisture recycling increased by 60% from 1961–1990 (49.5 mm) to 1991–2016 (80 mm) (Fig. 11). 3.2.2. Decreased percentage of snowfall in annual precipitation after 1990 This reflects the accelerating transformation of gaseous water to liquid Snow is the main type of solid precipitation in the Qilian Mountains, water. and the average snowfall during 1961–2016 was 20 mm, accounting for 7% of the total annual precipitation. Snowfall displayed an increasing trend with fluctuation during 1961–1990, but it decreased by 4. Discussion 2.2 mm/10 a during 1991–2016 (Fig. 10). In 1961–1990, snowfall accounted for 13% of the average annual precipitation during 4.1. Increasing runoff after 1990 1961–2016, whereas this ratio decreased to 6% in 1991–2016 (Fig. 10). Precipitation maintained an increasing trend during the entire Under the multiphase water transformation, increasing rainfall, ab- period of 1961–2016 (Fig. 5). In the spatial pattern, the amplitudes of lating glaciers, and degrading permafrost, the outlet runoff in the Qilian decrease of snowfall during 1991–2016 increase from west to east, Mountains has also changed. In the Shulehe River Basin, the annual av- and larger amplitudes occur mainly in the Shiyanghe River Basin and erage runoff in the Changmahe River increased by 0.993 × 108 m3/10 a Qinghaihu Lake Basin (Fig. 10). The percentage of snowfall in annual during 1953–2016, and the increasing trend during 1991–2016 was 7.5 precipitation has decreased by 7% since 1990. These variations reflect times higher than that of 1953–1990, reflecting accelerative runoff in- the decrease of solid precipitation and the increase of liquid precipita- crease (Fig. 12). Additionally, winter, spring, summer, and autumn run- tion, which also confirms the accelerative transformation from solid off increased by 52%, 30%, 35%, and 48% from 1953–1990 to 1991–2016, water to liquid water in the Qilian Mountains. and the corresponding increase values of discharge were 4.5, 3.9, 23, and 9.7 m3/s, indicating that although the runoff increase occurred 3.2.3. Increased contribution from recycling moisture to precipitation mainly in summer and autumn the relative rate of increase was higher Moisture recycling refers to the contributions of moisture from ter- in autumn and winter (Fig. 12). For the Danghe River, annual average restrial evaporation and transpiration to precipitation, which includes runoff increased by 0.17 × 108 m3/10 a during 1966–2016 (Fig. 12). Dif- moisture evaporated from the surfaces of soil and water and moisture ferent from the case of the Changmahe River, winter discharge in the from plant transpiration. Based on many previous studies (Kong et al., Danghe River decreased by 0.2 m3/s from 1966–1990 to 1991–2016, 2013; Schlesinger and Jasechko, 2014; Cui and Li, 2015; Li et al., whereas spring, summer, and autumn discharge increased by 9%, 24%, 2016a, 2016b), moisture recycling, with strengthening evapotranspira- and 14%, and the corresponding increase values were 1.1, 3.4, and tion, has been the crucial component of local precipitation under cli- 1.3 m3/s, respectively, also confirming larger summer and autumn dis- mate warming, and it has made a great impact on hydrological charge increases.

Fig. 9. Hexi inland river region, Qinghai inland river region, and upstream of the Yellow River region in the Qilian Mountains (a); glacier retreat rate in the Hexi inland river region (b); glacier retreat rate in the Qinghai inland river region (c); glaciers retreat rate in the upstream of Yellow River region (d). Z. Li et al. / Science of the Total Environment 663 (2019) 315–328 323

Fig. 10. Annual variation of snowfall during 1961–2016 (a); annual variation of snowfall during 1991–2016 (b); percentage of snowfall in annual precipitation during 1961–1990 and 1991–2016 (c); spatial pattern for variation amplitudes of snowfall during 1991–2016 (d).

In the Heihe River Basin, the outlet runoff of the Taolaihe River autumn runoff increased by 1.8, 3.8, 9.6, and 5.9 m3/s, with correspond- displayed a fluctuating decreasing trend during 1972–2016, but it in- ing increase rates of 13%, 14%, 9%, and 14%, respectively, which also con- creased at 0.38 × 108 m3/10 a during 1991–2016 (Fig. 13). In its seasonal firms that larger runoff increases occurred mainly in summer and variation, summer and autumn runoff maintained a stable state from autumn (Fig. 13). 1972–1990 to 1991–2016, whereas winter and spring runoff decreased In the Shiyanghe River Basin, the outlet runoff for the Zamu River by 1.42 m3/s and 1.59 m3/s, respectively (Fig. 13). The outlet runoff in- displayed a fluctuant variation with a slight decrease of 0.08 creased by 0.64 × 108 m3/10 a from 1945 to 2016 and, additionally, ×108 m3/10 a during 1952–2016, whereas it increased by 0.14 the rate of increase (2.84 × 108 m3/10 a) during 1991–2016 was 7.5 ×108 m3/10 a during 1991–2016; autumn and winter runoff in- times higher than that during 1945–1990 (0.38 × 108 m3/10 a) creased obviously after 1990 (Fig. 14). For the Nanying River, the (Fig. 13). From 1945–1990 to 1991–2016, winter, spring, summer, and outlet runoff displayed a fluctuant decrease of 0.055 × 108 m3/10 a

Fig. 11. Annual variation of precipitation from recycling moisture during 1961–2016 (a); average precipitation from the recycling of moisture during 1961–1990 and 1991–2016 (b). 324 Z. Li et al. / Science of the Total Environment 663 (2019) 315–328

Fig. 12. Annual and seasonal variation of runoff in the Shulehe River Basin: Changmahe River (a, b) and Danghe River (c, d). during 1955–2016, but the rate of decrease was reduced during 1956–2016, whereas it increased by 0.29 × 108 m3/10 a during 1991–2016 (0.021 × 108 m3/10 a) (Fig. 14). The outlet runoff for 1991–2016; winter and autumn runoff increased from 1956–1990 the Xiying River decreased by 0.03 × 108 m3/10 a during to 1991–2016 (Fig. 14).

Fig. 13. Annual and seasonal variation of runoff in the Heihe River Basin: Heihe River (a, b) and Taolaihe River (c, d). Z. Li et al. / Science of the Total Environment 663 (2019) 315–328 325

In summary, the outlet runoff for three inland rivers displayed an obvious increasing trend during 1991–2016 with a rate of 5 × 108 m3/10 a (Fig. 15), and the increasing trend is signiﬁcant at the 0.05 level, indicating the obvious runoff effect of the multiphase water transformation.

4.2. Lag between precipitation and runoff after 1990

The multiphase water transformation, especially the increasing rainfall, heavy ablation of glaciers, and degradation of permafrost, has had a great influence on the seasonal pattern of runoff. In their seasonal variation, six of the rivers displayed a “single peak” with maximum runoff occurring in July, because precipitation was concentrated mainly in summer, whereas only the Danghe River had a “double peak” of April and July (Fig. 16). For the Danghe River, the first runoff peak occurred in spring and was dominated by spring snow meltwater, whereas the second runoff peak was contributed mainly by precipitation. Considering the relationship between precipitation and runoff, the maximum runoff lagged the maximum precipitation by one month for Fig. 15. Annual total runoff variation for three inland rivers during 1991–2016. the Changmahe River and Taolaihe River (Fig. 16). There were two rea- sons for this lag. One is the influence of the increasing glacier and snow meltwater, because in contrast to precipitation glacier and snow can (Zhang et al., 2015). The other is the permafrost degradation because supply runoff after a specific ablation period. The other is the thickening this would result in the decline or disappearance of soil impermeability permafrost active layer, which would be an “underground reservoir” and therefore advance the confluence runoff time. However, the perma- supplying the groundwater with additional surface water and extend- frost area in the Shiyanghe River Basin accounts for only 18% of the basin ing the confluence runoff time. (Fig. 17). For the Xiying River and Nanying River, the maximum runoff was For the three inland rivers, summer and autumn runoff accounted one month earlier than the maximum precipitation (Fig. 16). Two rea- for about 80% of annual total runoff (Table 3). As shown in Table 3,the sons can account for this difference. One is the attenuated regulatory ef- percentage of summer runoff in annual total runoff was higher after fects of glacier and snow meltwater to runoff due to the smaller 1990 than before 1990 in the Danghe River and Taolaihe River. The per- glaciated area of only 39.94 km2 in the Shiyanghe River Basin. The centage of autumn runoff in annual total runoff has also commonly in- total glacier runoff over glacier cover observed since 1970 reached its creased from before 1990 to after 1990 in the Changmahe River, peak in the 2000s (Zhang et al., 2015), and the decreasing glacier accu- Taolaihe River, Heihe River, Xiying River, Nanying River, and Zamu mulation and glacier mass balance also reflect this decreasing effect River. These facts confirm that the increasing runoff under multiphase

Fig. 14. Annual and seasonal variation of runoff in the Shiyanghe River Basin: Zamu River (a, b), Nanying River (c, d), and Xiying River (e, f). 326 Z. Li et al. / Science of the Total Environment 663 (2019) 315–328

Fig. 16. Average monthly runoff variation in the Danghe River (a) and relationship between precipitation and runoff in the Changmahe River (b), Taolaihe River (c), Xiying River (d), and Nanying River (e). water transformation has occurred mainly in the ablation period (sum- rates are 28%, 22%, and 13%, respectively (Li et al., 2016b). Glacier mer and autumn). It can be concluded that the variation of multiphase snow meltwater accounted for 6% of the outlet river water in the water transformation has a great inﬂuence on the seasonal runoff Taolaihe River Basin, while the contribution rate from supra- variation. permafrost water was 15% (Li et al., 2016b). The results indicated that supra-permafrost water and glacier snow meltwater have contributed 4.3. Change of runoff components 28% and 7%, on average, to the outlet river water in the Heihe River Basin, respectively (Li et al., 2016c). Based on the newest study (Li Under multiphase water transformation, the runoff components et al., 2019b), it was found that the cryosphere belt accounts for 44% have also changed for inland rivers. Because of heavy glacier ablation, of the area of the upstream mountainous region, but that it contributes the contribution of glacier snow meltwater to runoff was 13.4% in the to about 80% of the water resources in the Qilian Mountains and Hexi headwaters of the Changmahe River in the western Qilian Mountains corridor. Glacier snow meltwater from the Qiyi Glacier in the Qilian with the relatively larger glacier cover (Zhou et al., 2015). Glacier Mountains increased by 0.41 × 106 m3 from 1960–1995 to snow meltwater only accounted for 3% of the outlet runoff in the 1996–2004, alongside a 0.41 °C temperature increase in the glacial re- Shiyanghe River Basin, whereas the contribution rate from supra- gion (Song et al., 2010). Based on research by Gao et al. (2011), the av- permafrost water was 20% (Li et al., 2016a). The contributions of glacier erage glacier mass balance in the Qilian Mountains during 1961–2006 is snow meltwater to outlet runoff for the three tributaries, the Zamu −49.5 mm/a, and the average contribution rate of meltwater to runoff is River, Xiying River, and Nanying River, are 7%, 5%, and 1%, respectively. 14.1%, with an annual mean meltwater runoff of 10.2 × 108 m3.Al- However, with respect to supra-permafrost water, these contribution though local water resources have been dominated mainly by Z. Li et al. / Science of the Total Environment 663 (2019) 315–328 327

Fig. 17. Conceptual model for multiphase water transformation in the Qilian Mountains. precipitation, meltwater from the cryosphere has been an important One is that the maximum runoff lagged the maximum precipitation runoff component in the Qilian Mountains. by one month due to the influence of increasing glacier snow meltwater and thickening permafrost active layer. The other is that the maximum 5. Conclusions runoff was advanced of the maximum precipitation by one month, which was caused by attenuated regulatory effects from glacier snow The warming and moistening climate in the Qilian Mountains is meltwater and permafrost degradation. Although local water resources mainly driving the multiphase water transformation. Annual average were mainly dominated by precipitation, meltwater from the temperature and precipitation increased by 0.35 °C/10 a and cryosphere has been an important runoff component in the Qilian 14.7 mm/10 a during 1961–2016, which caused a gradual extension of Mountains. the ablation period, significant decreases in ice days and frost days of Under climate warming, these accelerating multiphase water trans- 3.3 d/10 a and 3.9 d/10 a, respectively, and a growing season length in- formations, along with the annually increasing water, will obviously crease of 3.5 d/10 a. The warming amplitude and precipitation increase have a profound impact on water resources management and flood con- during 1991–2016 are two times higher than those during 1961–1990. trol in cold basins. Therefore, in the future, particular attention should Additionally, the amplitude of warming in the cryosphere belt is 0.05 °C/ be paid to understanding the runoff effects of these transformations. Es- 10 a higher than that in the vegetation belt, and the precipitation in- pecially the contribution from multiphase water transformation to increase rate is 9 mm/10 a higher in the cryosphere belt. For transforma- creased runoff should be quantified, the transformation from liquid tion of solid water to liquid water, the glacier retreat rate during water to gaseous water should be explored, and future runoff changes 1956–1990 was 18.2 km2/10 a; however, the corresponding value in should be forecast for periods of 30–50 years. Furthermore, it is urgently 1990–2010 was 27.3 km2/10 a, representing an acceleration of 50%. In necessary to reduce the potential threats to the water resources supply addition, the percentage of snowfall in precipitation decreased by 7% and ecology in the Qilian Mountains and Hexi Corridor. after 1990, and snowfall decreased by 2.2 mm/10 a during 1991–2016. Precipitation sourced by moisture recycling showed a sustentative in- Acknowledgments crease of 5.1 mm/10 a during 1961–2016, and the contribution from the recycling of moisture to precipitation increased by 60% after 1990, This study was supported by the Youth Innovation Promotion Asso- reflecting accelerating transformation of gaseous water to liquid ciation, CAS of China (2013274), the Strategic Priority Research Program water. Under multiphase water transformation, the outlet runoff for of the Chinese Academy of Sciences of China (XDA19070503), National three inland rivers displayed an obvious increasing trend of 5 Natural Science Foundation of China (91547102, 41771077), National ×108 m3/10 a after 1990, and the runoff increased occurred mainly in Key R&D Program of China (2017YFC0404305), the open funding from the ablation period. There were two different seasonal runoff patterns. State Key Laboratory of Urban and Regional Ecology (SKLURE2018-2-

Table 3 Comparison on the percent of seasonal runoff accounting for annual total runoff between before 1990 and after 1990.

Rivers Period Spring Summer Autumn Winter Period Spring Summer Autumn Winter

Danghe river 1966–1990 28.9% 33.2% 21.0% 17.0% 1991–2016 27.8% 36.4% 21.2% 14.6% Changmahe river 1953–1990 14.5% 58.7% 18.7% 8.0% 1991–2016 13.1% 58.0% 20.1% 8.8% Taolaihe river 1972–1990 17.0% 44.2% 23.0% 15.7% 1991–2016 15.7% 45.8% 23.9% 14.6% Heihe river 1945–1990 14.6% 56.2% 22.0% 7.2% 1991–2016 14.9% 55.1% 22.6% 7.3% Zamu river 1952–1990 17.9% 55.2% 23.1% 3.8% 1991–2016 19.2% 50.7% 25.7% 4.5% Xiying river 1956–1990 18.2% 56.1% 21.5% 4.2% 1991–2016 18.3% 53.1% 23.9% 4.8% Nanying river 1955–1990 13.4% 63.3% 19.6% 3.8% 1991–2016 13.0% 61.8% 21.1% 4.1% 328 Z. Li et al. / Science of the Total Environment 663 (2019) 315–328

6) and from the Key Laboratory of Mountain Hazards and Earth Surface Li, Zongxing, Qi, Feng, Wang, Q.J., Song, Yong, Li, Jianguo, Li, Yongge, 2016c. The influence from the shrinking cryosphere and strengthening evopotranspiration on hydrologic Process (SKLMHESP20181207) of China, and Longyuan Innovation and process in a cold basin, Qilian Mountains. Glob. Planet. Chang. 144, 119–128. Entrepreneurship Project in Gansu Province of China (LIEP-2018-35). Li, Zongxing, Qi, Feng, Li, Zongjie, Yuan, Ruifeng, Gui, Juan, Lv, Yuemin, 2019a. Climate We greatly appreciate suggestions from anonymous referees for the im- background, fact and hydrological effect of multiphase water transformation in cold regions of the Western China: a review. Earth Sci. Rev. 190, 33–57. provement of our paper. Thanks also to the editorial staff. Li, Zongxing, Gui, Juan, Wang, Xufeng, Feng, Qi, Zhao, Tongtiegang, Ouyang, Chaojun, Guo, Xiaoyan, Zhang, Baijuan, Shi, Yang, 2019b. Water resources in inland regions of cen- References tral Asia: evidence from stable isotope tracing. J. Hydrol. https://doi.org/10.1016/j. jhydrol.2019.01.003. Bie, Qiang, Qiang, Wenli, Wang, Chao, He, Lei, Zhao, Chuanyan, 2013. Monitoring glacier Liu, Guowei, 1997. Atmosphere Process in Hydrological Cycle. Science Press, Beijing, variation in the upper reaches of the Heihe river based on remote sensing in pp. 25–55 (in Chinese). 1960–2010. J. Glaciol. Geocryol. 35 (3), 574–582 (in Chinese). Liu, M.H., Sun, Z.Z., Niu, F.J., Wu, G.L., Yun, Humber, 2014. Variation characteristics of per- Bulygina, O.N., Razuvaev, V.N., Korshunova, N.N., 2009. Changes in snow cover over mafrost along the Qinghai Tibet Railway under the background of climate change. Northern Eurasia in the last few decades. Environ. Res. Lett. 4, 045026. J. Glaciol. Geocryol. 36 (5), 1122–1130 (in Chinese). Callaghan, T.V., Johansson, M., Anisimov, O., Christiansen, H.H., Instanes, A., Romanovsky, Liu, Shiyin, Yao, Xiaojun, Guo, Wanqin, Xu, Junli, Shangguan, Donghui, Wei, Junfeng, Bao, V., Smith, S., 2011. Changing permafrost and its impacts. Snow, Water, Ice and Perma- Weijia, Wu, Lizong, 2015. The contemporary glaciers in China based on the second frost in the Arctic (SWIPA). Arctic Monitoring and Assessment Program (AMAP). Chinese glacier inventory. Acta Geograph. Sin. 70 (1), 3–16 (in Chinese). Cao, Bo, Pan, Baotian, Gao, Hongshan, Jiang, Shaofei, Wen, Yuhua, Shangguan, Donghui, Malatinszky, A., Adam, S., Falusi, E., et al., 2013. Climate change related land use problems 2010. Glacier variation in the Lenglongling Range of eastern Qilian Mountains from in protected wetlands: a study;in a seriously affected hungarian area. Clim. Chang. 1972 to 2007. J. Glaciol. Geocryol. 32 (2), 242–248 (in Chinese). 118 (3-4), 671–682. Chen, Zhikun, Zhang, Shuyu, Luo, Jiali, Li, Zhaorong, Sun, Yan, 2012. Analysis on the Niu, Li, Ye, Baisheng, Li, Jing, Sheng, Yu, 2011. Effect of permafrost degradation on hydro- change of precipitation in the Qilian Mountains. Arid Zone Res. 29 (5), 847–853 (in logical processes in typical basins with various permafrost coverage in western China. Chinese). Sci. China Earth Sci. 54 (4), 615–624. Chen, Rensheng, Wang, G., Yang, Y., Liu, Jonhson, Han, C., Song, Y., Liu, Zhangwen, 2018. Ren, Jiawen, Ye, Baisheng, Ding, Yongjian, Liu, Shiyin, 2011. Initial estimate of the contri- Effects of cryospheric change on alpine hydrology: combining a model with observa- bution of cryospheric change in China to sea level rise. Chin. Sci. Bull. 56 (14), tions in the upper reaches of the Hei River, China. J. Geophys. Res. Atmos. 123, 1084–1087 (In Chinese). 3414–3442. Schlesinger, W.H., Jasechko, S., 2014. Transpiration in the global water cycle. Agric. For. Cheng, Guodong, Jin, Huijun, 2013. Permafrost and groundwater on the Qinghai-Tibet Meteorol. 189-190, 115–117. Plateau and in northeast China. Hydrogeol. J. 21 (1), 5–23. Shi, Yafeng, Shen, Yongping, Li, Dongliang, Zhang, Guowei, Ding, Yongjian, Hu, Ruji, Kang, China Meteorological Administration (CMA), 2006. Climate and Environment in China. Ersi, 2003. Discussion on the present climate change from warm-dry to warm-wet in Meteorology Press, Beijing. northwest China. Quat. Sci. 23 (2), 152–164 (in Chinese). Cui, B.L., Li, X.Y., 2015. Runoff processes in the Qinghai Lake Basin, Northeast Qinghai- Song, Gaoju, Wang, Ninglian, Jiang, Xin, He, Jianqiao, Wu, Xiaobo, 2010. Study on glacial Tibet Plateau, China: Insights from stable isotope and hydrochemistry. Quarter. melt-water change of Qiyi glacier in climate warming of Qilian Mountain. J. China Inter. 1–10. Hydrol. 30 (2), 84–88 (in Chinese). Ding, Y.H., Wang, H.J., 2016. New understanding of the scientific issues of climate change Sun, Meiping, Liu, Shiyin, Yao, Xiaojun, Guo, Wuanqin, Xu, Junli, 2015. Glacier changes in in China during the past 100 years. Chin. Sci. Bull. 61, 1029–1041 (In Chinese). the Qilian Mountains in the past half century: based on the revised First and Second Du, J., Hong, J.C., Lu, J.J., Hongya, Chen, D.M., 2012. Response of seasonal frosen soil to cli- Chinese Glacier Inventory. Acta Geograph. Sin. 70 (9), 1402–1414 (in Chinese). mate change on Tibet region from 1961 to 2010. J. Glaciol. Geocryol. 34 (3), 512–521 Tian, Hongzhen, 2013. Glacier Area Changes in the Qilian Mountains, China (Doctoral the- (in Chinese). sis of Lanzhou University, in Chinese). Gao, Xin, Zhang, Shiqiang, Ye, Baisheng, Gao, Hongkai, 2011. Recent changes of glacier Wang, Zongtai, Liu, Chaohai, You, Gengxia, et al., 1981. Glacier Inventory of China I Qilian runoff in the Hexi Inland River Basin. Adv. Water Sci. 22 (3), 344–350 (in Chinese). Mountains. Lanzhou Institute of Glaciology and Cryopedology, Chinese Academy of Gao, Yongpeng, Yao, Xiaojun, An, Lina, Li, Xiaofeng, Gong, Peng, Qi, Miaomiao, 2018. Sciences, Lanzhou (in Chinese). Change of ice volume in the Qilian Mountains during the period from 2000 to Wang, Q.C., Li, L., Li, D.L., Qin, N.S., Wang, Z.Y., Zhu, X., 2005. Response of permafrost over 2010. Arid Zone Res. 35 (2), 325–333 (in Chinese). Qinghai Plateau to climate warming. Plateau Meteorol. 24 (5), 708–713 (in Chinese). Gong, Ninggang, Sun, Meiping, Yan, Luxia, Gong, Peng, Ma, Xing-gang, Mou, Jianxin, 2017. Wang, Zhongwu, Qi, Weixiu, Bai, Lin, Li, Quanping, Yan, Yuxia, 2018. Reanalysis of climate Temporal and spatial characteristics of atmospheric water vapor and its relationship change in Qilian Mountain area. Qinghai Prataculture 27 (2), 42–48 (in Chinese). with precipitation in Qilian Mountains during 1979–2016. Arid Land Geogr. 40 (4), Xu, Haojie, Yang, Taibao, Zhang, Xiaoxiao, 2014. Climate change in upstream area of Shule 762–771 (In Chinese). river and its effects on runoff-yield in last 50 years. Bull. Soil Water Conserv. 34 (4), IPCC, 2013. Climate Change 2013: The Physical Science Basis. Cont ribution of Working 39–52 (in Chinese). Group I to the Fifth Assessment Report of the intergovernmental panel on climate Yao, Tandong, Qin, Dahe, Shen, Yongping, Zhao, Lin, Wang, Ningliang, Lu, Anxin, 2013. change. Cambridge University Press, Cambridge, United Kingdom and New York, Cryospheric changes and their impacts on regional water cycle and ecological condi- NY, USA. tions in the Qinghai-Tibetan Plateau. Chin. J. Nat. 35, 179–186 (in Chinese). Kääb, A., Chiarle, M., Raup, B., Schneider, C., 2007. Climate change impacts on mountain Ye, Baisheng, Ding, Yongjian, Liu, Fengjing, Liu, Caohai, 2003. Responses of various-sized glaciers and permafrost. Glob. Planet. Chang. 56 (1) (vii-ix). alpine glaciers and runoff to climatic change. J. Glaciol. 49 (164), 1–7. Kabel, K., Moros, M., Porsche, C., Neumann, T., Adolphi, F., Andersen, T.J., et al., 2012. Im- Yu, Guobing, Li, Zhongqin, Wang, Puyu, 2014. Glaciers changes at Daxue Mountain and pact of climate change on the baltic sea ecosystem over the past 1,000 years. Nat. Danghenan Mountain of west Qilian Mountains in the recent 50 years. Arid Land Clim. Chang. 2 (12), 871–874. Geogr. 37 (2), 299–309 (in Chinese). Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S., Zhang, Yingsong, Liu, Shiyin, Shangguan, Donghui, Li, Jing, Zhao, Jingdong, 2012. Thinning White, G., Woollen, J., Zhu, Y., Leetmaa, A., Reynolds, R., Chelliah, M., Ebisuzaki, W., and shrinkage of Laohugou no. 12 glacier in the Western Qilian Mountains, China, Higgins, W., Janowiak, J., Mo, K.C., Ropelewski, C., Wang, J., Jenne, Roy, Joseph, from 1957 to 2007. J. Mt. Sci. 9 (03), 343–350. Dennis, 1996. The NCEP/NCAR 40-year reanalysis project. Bull. Am. Meteorol. Soc. Zhang, Wenjie, Chen, Weiming, Li, Baolin, Tong, Chiming, Zhao, Min, Wang, Nan, 2014a. 77, 437–471. Simulation of the permafrost distribution on Qilian Mountains over past 40 years Kistler, R., Kalnay, E., Collins, W., Saha, S., White, G., Woollen, J., Challiah, M., Ebisuzaki, W., under the influence of climate change. Geogr. Res. 33 (7), 1275–1284 (in Chinese). Kanamitsu, M., Kousky, V., van den Dool, H., Jenne, R., Fiorino, M., 2001. The NCEP- Zhang, Liang, Zhang, Qiang, Feng, Jianying, Bai, Huzhi, Zhao, Jianhua, Xu, Xiaohua, 2014b. NCAR 50-year reanalysis: monthly means CD-ROM and documentation. Bull. Am. A study of atmospheric water cycle over the Qilian Mountains (II): analysis of hydro- Meteorol. Soc. 82, 247–267. logical cycle. J. Glaciol. Geocryol. 36 (5), 1092–1100. Kong, Y.l., Pang, Z.H., Froehlich, K., 2013. Quantifying recycled moisture fraction in precip- Zhang, Shiqiang, Gao, Xin, Zhang, Xiaowen, 2015. Glacial runoff likely reached peak in the itation of an arid region using D-excess. Tellus B 65, 19251. mountainous areas of the Shiyang River Basin, China. J. Mt. Sci. 12 (2), 382–395. Li, Zongxing, He, Yuanqing, Wang, Puyu, Theakstone, Wilfred H., An, Wenling, Wang, Zhang, Yanlin, Cheng, Guodong, Li, Xin, Jin, Huijun, Yang, Dawen, Flerchinger, Gerald N., Xufeng, Lu, Aigang, Zhang, Wei, Cao, Weihong, 2012. Changes of daily climate ex- Chang, Xiaoli, Bense, Victor F., Han, Xujun, Liang, Ji, 2017. Influences of frozen ground tremes in southwestern China during 1961–2008. Glob. Planet. Chang. 80–81, and climate change on hydrological processes in an alpine watershed: a case study in 255–272. the upstream area of the Hei'he River, Northwest China. Permafr. Periglac. Process. Li, Zongxing, Qi, Feng, Wang, Q.J., Song, Yong, Cheng, Aifang, Li, Jianguo, 2016a. Contribu- 28, 420–432. tion from frozen soil meltwater to runoff in an in-land river basin under water scar- Zhou, Youwu, Guo, Dongxing, Qiu, Guoqing, Cheng, Guodong, Li, Shude, 2000. Frozen Soil city by isotopic tracing in northwestern China. Glob. Planet. Chang. 136, 41–51. in China. Science Press, Beijing (in Chinese). Li, Zongxing, Qi, Feng, Wang, Q.J., Song, Yong, Li, Jianguo, Li, Yongge, Yamin, Wang, 2016b. Zhou, Jiaxin, Wu, Jinkui, Liu, Shiwei, Zeng, Guoxiong, Qin, Jia, Wang, Xiuna, Zhao, Qiudong, Quantitative evaluation on the influence from cryosphere meltwater on runoff in an 2015. Hydrograph separation in the headwaters of the Shule river basin: combining inland river basin of China. Glob. Planet. Chang. 143, 189–195. water chemistry and stable isotopes. Adv. Meteorol. 2015, 1–10.