Effects of season and slope orientation on stable isotopes of different water bodies in the subalpine shrub zone of the eastern Qilian Mountains in China

Water in nature is connected and transformed through the water cycle process, including water surface evaporation, water vapor transport, condensation precipitation, plant transpiration, runoff, infiltration, etc. (Vreča & Kern 2020). As the main bodies of water transformation, precipitation, soil water, plant water, river water, and groundwater play important roles in the water cycle. Therefore, it is important for the rational utilization of regional water resources to track and study the patterns and processes of the regional water cycle (Sun et al. 2016).

As the constituent elements of H₂O, ¹⁸O and D can reflect the various links of the water cycle and are used as effective indicators to study the regional water cycle (Qiu et al. 2019; Pu et al. 2020). Deuterium excess (d-excess) is also frequently used to study meteorological conditions in source areas of water vapor and below-cloud evaporation (Malik et al. 2022). Therefore, stable isotopes of hydrogen and oxygen as well as d-excess have been applied to study the isotopic characteristics of different water bodies (Liu et al. 2019), water transport and replenishment mechanisms (Malik et al. 2022; Zhang et al. 2022), the hydrological processes at the soil–plant–atmosphere continuum (SPAC) (Sprenger et al. 2016; Che et al. 2019), and the contribution ratios of plant water source, etc. (Phillips & Gregg 2003). This can effectively reveal water transport relationships between different water bodies and trace water cycle processes at local and global scales. Derived from the global meteorological precipitation line (GMWL) (Craig 1961), the local meteorological water line (LMWL) reveals regional environmental information. In order to more accurately understand regional precipitation characteristics, scholars generally prefer to understand the LMWL, with its slope and intercept reflecting the fractionation of precipitation and regional environmental information. As temperature increases and relative moisture decreases, the slope and intercept of the LMWL tend to decrease (Peng et al. 2010; Yeh et al. 2014). Currently, scholars can qualitatively determine the replenishment relationships between different water bodies using their waterline relationships and isotopic differences (Dutton et al. 2005; Yeh et al. 2014; Liu et al. 2019; Zhu et al. 2022), and thus determine local water movement patterns. This is an important base for quantitative studies of the local water cycle.

Atmospheric precipitation is the main input source of the water cycle. The stable isotopic composition of precipitation is mainly influenced by temperature, precipitation amount, altitude, the source of water vapor (sea or land source), latitude, below-cloud secondary evaporation, etc. (Dansgaard 1964; Kumar et al. 2010; Ma et al. 2012; Jia et al. 2018). In arid areas, the stable isotopes of precipitation are primarily influenced by temperature (Ma et al. 2018) and become enriched under very low precipitation conditions (Ma et al. 2012). The stable isotope composition of soil water is influenced by seasonal precipitation, plant type and cover, topography, and soil texture (Wan & Liu 2016). Shallow soil water is the primary source of water for most plants, and plant water uptake can affect the isotopic composition of this shallow soil water (Jackson 2000). The stable isotopic composition of plant water is mainly influenced by water sources and transpiration (Dai et al. 2015; Marttila et al. 2018). Those who use groundwater replenishment have a small variation in plant water isotopes and are close to the isotope values of groundwater (Dai et al. 2015). The stable isotopes of river water are mainly influenced by replenishment sources such as glacial snowmelt, precipitation, and groundwater (Dutton et al. 2005). Environmental information of the basin can influence the evaporative enrichment of a river. The magnitude of river flow can affect its replenishment source and then affect its isotopic abundance. For example, groundwater can replenish the river when river flow is low (Gurumurthy et al. 2015). The isotope values of groundwater are usually low and unaffected by evaporation effects (Komor & Emerson 1994; Sprenger et al. 2016), it is usually a mixture of precipitation and soil water and is replenished by high intensity precipitation (Sprenger et al. 2016). The stable isotopes in different water bodies are influenced by a variety of factors, which can be attributed to environmental factors and hydrological processes. These factors can lead to differences of stable isotopes in various study areas and various water bodies.

The Qilian Mountains, located in the northeast of the Tibetan Plateau, are an important ecological barrier in the northwest arid area and the origin of many inland rivers of the Hexi area in China (Ma et al. 2018). Based on the stable isotope techniques, many studies on the hydrological processes have been carried out in the Qilian Mountains (Jia et al. 2018; Yong et al. 2020). The subalpine shrub forest is an important part of the Qilian Mountain ecosystem, which has a high water-conserving function (Qin et al. 2007). Therefore, the study of hydrological processes in the subalpine shrub zone of the Qilian Mountains is of great importance for the conservation of forest eco-hydrological systems. In addition, the differences in solar radiation time, vegetation type, and cover together with soil texture in different slope orientations of the gully can also influence its hydrological processes. However, at present, there are fewer studies on the differences in isotopic composition between different seasons, different slope orientations, and different water bodies in the subalpine scrub area of the Qilian Mountains from the perspective of water cycling. Therefore, based on the stable isotope values of hydrogen and oxygen of precipitation, soil water, plant water, river water, and groundwater collected in the subalpine shrub zone of the eastern Qilian Mountains from May to October 2019, the characteristics of hydrogen and oxygen isotopes in five types of water bodies were studied, and then the replenishment relationships among them were qualitatively discussed using the water lines and isotopic differences in water bodies. This will help to understand the eco-hydrological processes in the subalpine shrub zone that are weakly influenced by human activities, and provide a theoretical basis for the conservation of water resources and the restoration of vegetation ecosystems in the Qilian Mountains in the context of global warming.

During the growing season from May to October 2019, precipitation, soil water, plant water, river water, and groundwater were collected in sampling points (Figure 1). There are 681 samples in total, including 57 precipitation samples, 280 soil samples, 240 plant samples, 90 river water samples, and 14 groundwater samples.

Precipitation samples were collected using a standard rain gauge placed 1.5 m above the ground and with no obstacles around. After a precipitation event considered from 20:00 of the day to 20:00 of the next day, the precipitation amount was recorded and put into the sample bottles. The soil samples were sampled in 10 cm intervals, and the soil depths were 80 cm on the semi-sunny slope and 60 cm on the semi-shady slope with a sampling interval of about 15 days. Some soil samples were used to determine the stable isotopes of hydrogen and oxygen, and others packed in aluminum boxes were used to determine soil water content (SWC) by the classical drying method. Salix oritrepha Schneid., Salix cupularis, Salix sclerophylla Anderss., Rhododendron thymifolium Maxim., and Potentilla fruticosa L. were selected from the semi-sunny slope, and Salix oritrepha Schneid., Salix cupularis, Salix sclerophylla Anderss., Potentilla fruticosa L., Rhododendron przewalskii Maxim., Rhododendron anthopogonoides Maxim., and Caragana jubata (Pall.) Poir. were selected from the semi-shady slope. Every 15 days or so, 5–7 non-green bolting branches about 0.3 cm in diameter and 5 cm in length were cut from each plant with a relatively consistent growth condition, and the samples were immediately put into sample bottles after their epidermis and phloem were removed. River water samples were collected every 5 days at the Ningchan River and its tributary about 60 m away from the sampling points of soil water and plant water. Coming from a natural spring about 150 m away from the sampling points of soil water and plant water, groundwater samples were collected every 15 days or so, but no samples were collected in August. All samples were put into 10 mL polyethylene bottles and stored frozen at approximately −15 °C.

All experiments were completed at the Isotope Laboratory and the Soil Analysis Laboratory in the College of Geography and Environmental Science, Northwest Normal University.

One-way analysis of variance (ANOVA) and independent t-tests were used to analyze the differences between various types of water bodies in different slope directions and seasons.

The weighted average values of δD, δ¹⁸O, and d-excess in the study area from May to October were −45.62 ± 26.05, −7.78 ± 3.32, and 16.62 ± 6.16‰ (>10‰), the values ranged during −99.48 to 6.12‰, −15.59 to 0.02‰, and 3.91 to 27.67‰, respectively (Table 1). In the study area, the d-excess values of precipitation were mostly above 10‰, which was higher than those (<8‰) in the arid area of northwest China. This may be because the study area is located at high altitude mountainous and is mainly influenced by westerly water vapor and local recirculation water vapor (Ma et al. 2012).

The average values of δD, δ¹⁸O, and d-excess of plant water in the study area were −45.05 ± 9.61, −6.19 ± 1.87, and 4.48 ± 8.28‰, with values varying from −98.57 to −20.79‰, from −12.40 to −2.32‰, and from −39.33 to 22.16‰, respectively (Table 1). Compared to soil water, the mean and standard deviation of the stable isotope values of plant water were higher, which was consistent with the previous results (Che et al. 2019). However, Tetzlaff et al. (2021) showed that the isotope value of xylem water was lower than that of soil water. This may be related to the water absorption characteristics of different plant species in different study areas. In the study area, plant transpiration has a greater enrichment effect on stable isotopes, which leads to the stable isotope values of plant water higher than soil water.

As shown in Figure 3(c) and 3(d), the δD and δ¹⁸O values of plant water in the study area reached their maximum values on August 6 and May 21, respectively, and both of them decreased to the minimum values on October 13. In terms of different plant species, the δD and δ¹⁸O values of Salix oritrepha Schneid. were the largest (−41.44‰), but the δD value of Potentilla fruticosa L. (−50.76‰) and the δ¹⁸O value of Salix sclerophylla Anderss. were the smallest (−6.88‰). The range of δD values of Potentilla fruticosa L. was the largest (−98.57 to −36.21‰), while it was the smallest for Salix sclerophylla Anderss. (−55.36 to −43.46‰). The δ¹⁸O values of Salix oritrepha Schneid. varied in the largest range (−8.15 to 2.32‰), but those of Rhododendron anthopogonoides Maxim. varied in the smallest range (−7.14 to −5.22‰).

The average values of δD, δ¹⁸O, and d-excess were −53.10 ± 4.36, −8.88 ± 0.60, and 17.94 ± 2.92‰, and the variation ranged from −65.36 to −44.30‰, from −10.41 to −7.31‰, and from 7.01 to 24.89‰, respectively (Table 1). The stable isotope values of river water were the lowest among all types of water bodies in the study area, and also had the smallest range of variation. This indicated that the water source of the river was more stable and environmental factors minimally influenced its isotopic composition.

According to the results of one-way ANOVA, the δ¹⁸O values of precipitation in the study area were not significantly different in different seasons, the δD values of precipitation were significantly higher in the spring than in the summer(P < 0.05), and the d-excess values of precipitation were significantly higher in autumn than in other seasons (P < 0.05). However, the mean δ¹⁸O values were different in seasons, which showed summer (−7.35‰) > spring (−7.82‰) > autumn (−8.38‰) (Table 2). On the contrary, d-excess showed high values in autumn and low values in summer. This was due to the higher temperatures in the study area during the summer months despite higher precipitation, and the temperature effect of precipitation stable isotopes masked the precipitation effect. This was consistent with the results of the previous study (Ma et al. 2020).

The δ¹⁸O values of soil water in the study area were significantly different in spring and summer (P < 0.01), and δD values were significantly lower in spring than in summer and autumn (P < 0.01). The seasonal variation of δ¹⁸O of soil water in the study area showed that: summer (−7.34‰) > autumn (−7.60‰) > spring (−8.05‰) (Table 2), which may be related to the seasonal variation of soil water sources (Pu et al. 2020). In spring, seasonal permafrost started to melt, which increased the water content of surface soil and caused stable isotopes of soil water depletion. In addition, lower temperature resulted in their evaporative fractionation to weaken (Yong et al. 2020). In summer, soil water was mainly recharged by precipitation which had high stable isotope values, and high temperature led to isotope enrichment for evaporative fractionation. Benettin et al. (2018) also concluded that soil water exhibited a seasonal evaporation pattern, with higher temperatures and stronger evaporation in summer.

Plant type and growth stage can affect stable isotope abundance. Based on the growth characteristics of subalpine shrub plants, the growth period of shrub plants during the sampling period was divided into three stages: germination and leaf development stage (May–June), flowering and fruiting period stage (June–September), and leaf fall recession period stage (September–October). Judging from different growth periods, the δD and δ¹⁸O values of plant water in the study area were as follows (Table 3): germination and leaf development stage > flowering and fruiting stage > leaf fall recession stage. It could be seen that the δD and δ¹⁸O values of plant water showed a decreasing trend with the plant growth, reflecting that there were different characteristics of water absorption of subalpine shrubs in different growth periods. During the germination and leaf development stage, the shrubs had a strong absorption demand for water, but there was less precipitation, resulting in higher stable isotope values of plant water. During the flowering and fruiting stage, the shrubs also had a strong absorption demand for water, but there was more precipitation, resulting in a lower stable isotope value of plant water. During the leaf fall recession stage, the temperature and the precipitation were decreasing, resulting in a decrease in the stable isotope value of plant water.

The δ¹⁸O values of the river water in the study area did not show significant differences across different seasons, whereas the δD values showed significant variations between spring with summer and autumn (P < 0.05). However, when considering the seasonal mean values (Table 2), it showed that spring (δD: −56.47‰; δ¹⁸O: −9.28‰) < summer (δD: −52.92‰; δ¹⁸O: −8.82‰) < autumn (δD: −50.73‰; δ¹⁸O: −8.65‰), which may be related to the seasonal variation of river water sources (Sun et al. 2016). In spring, river water was mainly supplied by the meltwater of glaciers, snow, and permafrost (Li et al. 2016b), so the value of δ¹⁸O was lower. In summer and autumn, river water was mainly supplied by precipitation that heavy isotope was enriched, and its evaporation was also stronger, thus resulting in its stable isotope values higher correspondingly.

On the seasonal scale, the δ¹⁸O values of the mainstream river water were autumn (−8.96‰) > spring (−8.98‰) > summer (−9.10‰), and its seasonal variation was not obvious, while that of the tributary river water was autumn (−8.35‰) > summer (−8.53‰) > spring (−9.55‰). This may be related to the evaporative fractionation and the replenishment source of river water (Chen et al. 2019). In spring, the main recharged sources of the tributary river water were snow and permafrost meltwater, resulting in isotope values of river water lower. In summer and autumn, precipitation and soil water were the main recharge sources of the tributary river water, and evaporative fractionation results in the stable isotopic enrichment of river water. The mainstream river water had a large flow in summer, and the evaporative fractionation rate was relatively lower and the stable isotope value of the river water was also lower (Shi et al. 2019). In addition, the water source of the mainstream river water was relatively stable, so the seasonal variation of stable isotope values was smaller.

According to the results of one-way ANOVA, the δ¹⁸O and δD values of groundwater in the study area were not significantly different, but their mean values were different. In terms of seasonal mean values (Table 2), the δ¹⁸O values of groundwater showed summer (−7.15‰) > autumn (−7.80‰) > spring (−7.85‰), and the δD values of groundwater showed summer (−40.36‰) > spring (−48.07‰) > autumn (−48.60‰). This indicates the enrichment of stable isotopes by evaporative fractionation in summer. In addition, the seasonal variation of δ¹⁸O values of groundwater was the same as that of soil water, which further indicated that there was some recharge relationship between soil water and groundwater.

The δ¹⁸O values of soil water showed a consistent variation trend with soil depth on different aspects (Figure 6). Both the semi-sunny slope and the semi-shady slope showed a decreasing trend with increasing soil depth (except for soil water of 0–10 cm in autumn), which was consistent with the results studied by Qiu et al. (2019). In all seasons, the δ¹⁸O values of soil water on the semi-shady and semi-sunny slopes showed inflection points at the soil layer of 30–40 cm, and showed an overall variation trend of decreasing, increasing to stabilizing. The same was true for the trend in standard deviation (Figure 6(b) and 6(d)). The standard deviation of stable isotopes of soil water in the shallow layer was greater than in the deep layer, indicating that the deep layer was relatively stable. Since surface soil water is significantly affected by precipitation, evaporation, and plant transpiration, isotope values of soil water gradually decrease with increasing of soil depth. With the increase of soil depth, the mixing of surface soil water and deep soil water caused the isotope values of soil water to increase and gradually stabilize (Du et al. 2021).

The independent samples t-test showed that there was no significant difference in the δD and δ¹⁸O values of the same species (Salix cupularis, Potentilla fruticosa L., Rhododendron thymifolium Maxim., and Salix oritrepha Schneid.) in different slope orientations (P > 0.05). However, the mean values and ranges of variation of δD and δ¹⁸O of the same species in different slope orientations were different. For example, in terms of the mean values and ranges of variation of δD and δ¹⁸O (Figure 7), Salix oritrepha Schneid. had the largest difference between semi-sunny and semi-shady slopes, and Salix cupularis had the smallest difference. In addition, except for the δ¹⁸O of Potentilla fruticosa L. and Rhododendron thymifolium Maxim., the mean values of δD and δ¹⁸O of plant water showed the semi-sunny slope > semi-shady slope.

Based on the δD and δ¹⁸O values of soil water, plant water, river water, and groundwater, the water lines of different water bodies in the study area were established as follows: SWL: δD = 5.84δ¹⁸O − 3.96 (R² = 0.87); PWL: δD = 4.40δ¹⁸O − 17.80 (R² = 0.63); RWL: δD = 5.82δ¹⁸O − 1.39 (R² = 0.64); GWL: δD = 10.03δ¹⁸O + 30.33 (R² = 0.67) (Figure 8). The SWL has a much smaller slope (5.84 < 7.63) and intercept (−3.96 < 14.06) than the LMWL, which may be related to evaporation affecting soil water isotope enrichment. This is consistent with previous studies near the Qilian Mountains (Qiu et al. 2016; Liu et al. 2022). The slope and intercept of the PWL were the smallest among the five water bodies, indicating that plant transpiration was stronger than evaporation from the other water bodies. This suggests that plant water contributes more to the local recirculation of water vapor. However, Kong et al. (2013) concluded that evaporation from soil water and surface water dominated the local recirculation of water vapor in arid and semi-arid regions. The difference may be related to vegetation coverage because the study area has higher vegetation coverage and stronger transpiration. The slope and intercept of GWL were the highest among the five types of water bodies in the study area, which was little affected by evaporation.

The slope orientation is an important environmental factor that affects the water lines of some water bodies. Soil water and plant water samples collected in the study area from different slope orientations were used to establish soil water lines (the semi-sunny slope: δD = 5.46δ¹⁸O − 7.79, R² = 0.91; the semi-shady slope: δD = 6.21δ¹⁸O + 0.15, R² = 0.91) and plant water lines (the semi-sunny slope: δD = 4.40δ¹⁸O − 20.14, R² = 0.67; the semi-shady slope: δD = 4.96δ¹⁸O − 13.87, R² = 0.60). Both the SWL and the PWL exhibited lower slope in the semi-sunny slope than in the semi-shady slope, indicating that the soil evaporation and plant transpiration were stronger on the semi-sunny slope.

Currently, scholars mainly use the establishment of isotopic linear relationships among different water bodies to qualitatively determine their hydraulic relationships (Dutton et al. 2005; Yeh et al. 2014; Liu et al. 2019; Zhu et al. 2022). According to Figure 8, except for a small amount sample of plant water, samples of soil water, river water, and groundwater are mostly located in the vicinity of the LMWL, indicating a clear relationship between the replenishment of precipitation to other water bodies in the study area. This indicated that there was a fractionation effect of precipitation during infiltration to soil, and that SWL was deviated by the influence of evaporation. Plant water was scattered below the LMWL and in the vicinity of the SWL, indicating a hydraulic relationship among precipitation, plant water, and soil water.

The stable isotope differences among different water bodies played an important role in the qualitative analysis of their hydrodynamic relationships. The variation range of δ¹⁸O in soil water, plant water, river water, and groundwater all lay in the middle of precipitation (Table 1), indicating the replenishment effect of precipitation on other water bodies. This is consistent with the judgment of the method used to construct the water line of water bodies (Dutton et al. 2005). The isotopic composition of soil water is related to precipitation amount and intensity, runoff, and groundwater mixing (Koeniger et al. 2016; Sprenger et al. 2016; Pu et al. 2020). According to Table 1, the mean δ¹⁸O values were similar between precipitation (−7.78 ± 3.32‰), soil water (−7.78 ± 3.32‰), and groundwater (−7.78 ± 3.32‰) in the study area, which confirmed the existence of strong hydraulic relationships among precipitation, soil water, and groundwater. Among them, the δ¹⁸O value of soil water was the largest, indicating that soil water was replenished by both precipitation and groundwater. Soil water includes precipitation-converted soil water and groundwater-converted soil water, and its isotopic evaporation enrichment occurs during the infiltration of precipitation. The δ¹⁸O of plant water was the largest of the five water bodies (Table 1), indicating the most significant enrichment of stable isotope by plant transpiration (Sprenger et al. 2016). The stable isotope values of plant water are mainly influenced by precipitation, soil water, groundwater, and leaf transpiration (Sprenger et al. 2016). However, the main source of plant water was soil water. The variation range of isotope values of river water and groundwater was similar, indicating a hydraulic relationship between them (Yeh et al. 2014). Because groundwater was sampled at higher points than river water, it can be inferred that the river water in the study area was replenished by groundwater and soil water. In addition, the stable isotope values of river water were closest to those of precipitation in summer (Table 2), indicating that the river water was mainly replenished by summer precipitation. In five types of water bodies, the stable isotope values of groundwater were the most stable and their various ranges were the smallest. The groundwater was located deep and was little affected by surface temperature and water absorption of plant roots, so its stable isotope values were weakly fractionated by evapotranspiration.

Based on the stable isotope values of precipitation, soil water, plant water, river water, and groundwater in the subalpine shrub zone in the eastern Qilian Mountains from May to October 2019, this study analyzed the stable isotope characteristics and their differences of five water bodies, and qualitatively analyzed their replenishment relationships.

The stable isotopes of different water bodies are influenced by temperature, precipitation, season, slope orientation, plant growth stage, and water replenishment sources. Among five water bodies, the stable isotope values of precipitation exhibited the greatest variation, while those of groundwater were the most stable. Evaporative fractionation of soil water was significant, but plant transpiration was stronger than evaporation of other water bodies. The stable isotope enrichment of precipitation occurred in high temperature and low humidity environments. There were seasonal differences in the values of δD and δ¹⁸O of different water bodies. The isotope values of plant water and soil water were higher and more stable on the semi-sunny slope than on the semi-shady slope, which was related to the high clay content and the strong ability to retain water of the semi-sunny slope. There were differences in the stable isotope values of plant water at different growth stages (germination and leaf development stage > flowering and fruiting stage > leaf fall recession stage). Compared to the mainstream river water, the stable isotope values of the tributary river water were higher and its variation fluctuations were larger.

In terms of water replenishment relationships, there was a strong hydraulic relationship between the various water bodies in the study area, and precipitation was the ultimate source of other water bodies. Precipitation replenished soil water through infiltration, and part of soil water was absorbed by plants and part of it continuously infiltrated to replenish groundwater. The soil water isotopes were changed by mixing with old soil water and deep groundwater during the infiltration of precipitation. In addition, groundwater and precipitation replenished the river water.

This research was supported by the Oasis Scientific Research Achievements Breakthrough Action Plan Project of Northwest Normal University (NWNU-LZKX-202303). The authors thank their colleagues at the Northwest Normal University and the Chinese Academy of Sciences (CAREERI, CAS) for their help in fieldwork, laboratory analysis, and data processing.

Y.Z. analyzed the data and wrote the original draft. W.J. revised the original draft. L.Y. and H.X. collected the samples. F.Z., M.Z., and X.L. participated in the experiment. All authors reviewed the manuscript.

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

The authors declare there is no conflict.