Soil water sources in permafrost active layer of Three-River

9 Water in permafrost soil is an important factor affecting the ecology of 10 cold environments, climate change, hydrological cycle, engineering, and 11 construction. To explore the variations in soil water in the active layer due 12 to permafrost degradation, the soil water sources in the Three-River 13 Headwater Region were quantified based on the stable isotope data (δH 14 and δO) of 1140 samples. The results showed that the evaporation 15 equation was δH = 7.46 δO 0.37 for entire soil water. The stable 16 isotope data exhibited a spatial pattern, which varied over the soil profile 17 under the influence of altitude, soil moisture, soil temperature, vegetation, 18 precipitation infiltration, soil water movement, ground ice, and 19 evaporation. Based on the stable isotope tracer model, precipitation and 20 ground ice accounted for approximately 88% and 12% of soil water, 21 respectively. High precipitation contributed to the soil water in the 3900– 22 4100 m, 4300–4500 m, and 4700–4900 m zones, whereas ground ice 23 contributed to the soil water in the 4500–4700 m and 4900–5100 m zones. 24 Precipitation contributed approximately 84% and 80% to the soil water in 25 grasslands and meadows, respectively, whereas ground ice contributed 26 approximately 16% and 20%, respectively. Precipitation; 27 evapotranspiration; physical and chemical properties of soil; and the 28 1 https://doi.org/10.5194/hess-2021-558 Preprint. Discussion started: 8 November 2021 c © Author(s) 2021. CC BY 4.0 License.


39
Soil water is the critical element of the water cycle and is closely where Q t is the total runoff discharge, Q m is the discharge of 291 component m, and C m j is the tracer j incorporated in the component m.

292
For isotope hydrograph separation, one of the tracers should be an isotope.

293
If there are more than four endmembers, calculation software, such as 294 IsoSource, must be used (Phillips and Gregg, 2003  River > source region of the Yellow River (Fig. 3b).

330
In addition, a negative correlation was observed between the δ 18 O and 331 d-excess values of soil water in the study area (Table 1). However, the 332 correlation coefficients were not significant, indicating the multiplicity 333 and complexity of the evolution of stable isotopes in soil water.  (Table 1). These  (Table 1).      Accordingly, it can be inferred that soil water in the study area is 597 recharged by multiple sources.

598
The relationship between soil water and the LWML varied respectively. These results confirmed that the degree of influence of evapotranspiration on stable isotopes decreased with increasing altitude.

606
The stable isotope values of soil water were generally clustered and 607 distributed with precipitation at different altitudes, indicating that 608 precipitation was the major soil water source in the study area.

609
The relationship between soil water and the LWML also varied  concentrations of ground ice, precipitation, and soil water in the study 630 area (Fig. 8). Accordingly, these δ 18 O and δ 2 H data were selected for 631 analysis because they could effectively characterize the sources. There suggesting that soil water was a mixture of them (Fig. 9). Therefore, 635 precipitation was considered as the first endmember and ground ice as the 636 second endmember. Soil water was also characterized during the 637 sampling period (Fig. 9). In the study region, precipitation and ground ice 638 accounted for approximately 88% and 12% of soil water, respectively, in 639 July 2019; hence, precipitation was the major source of soil water in July, 640 which is the rainy season. However, a large amount of ground ice likely 641 melted before July as soil temperatures increased, particularly in shallow 642 soils.

643
On sunlit slopes, the estimated contributions of precipitation and 644 ground ice to soil water were approximately 90% and 10%, respectively, 645 whereas those on shady slopes were approximately 86% and 14%, 646 respectively (Fig. 10). This difference can be explained by the following 647 reasons.
(1) The effect of solar radiation was stronger on sunlit slopes, as 648 the soil temperature was higher, ground ice melted faster, and melting 649 period started earlier than on shaded slopes; hence, the ground ice content 650 was higher on shady slopes. (2) Evapotranspiration was higher on sunlit 651 slopes than on shady slopes, whereas the soil water content was lower on 652 sunlit slopes than on shady slopes. In addition, piston flow development 653 was more favorable on sunlit slopes when precipitation occurred, as the 654 soil acted as a "dry sponge" with a strong capacity to absorb water.

655
The soil water sources also varied significantly at different 656 elevations (Fig. 10). The area below 3700 m was characterized by 657 seasonally frozen soil, where the major soil water source was 658 precipitation because the seasonally frozen soil melted before July; 659 therefore, ground ice did not contribute to soil water in this region. For whereas the contribution of ground ice gradually increased (Fig. 10). The   isotopes of soil water were affected by altitude (Fig. 11), no evident effect   (Table 2).

781
These findings indicate that (1) the stable isotopes of surface soil water  (2) rapid precipitation infiltration is more favorable with increased root 875 system, thereby mixing soil water and leading to more negative stable        Fig.1 The location of Three-River Headwater Region and distribution of 1281 sampling sites for soils and waters