Seasonal and Depth Dynamics of Soil Moisture Affect Trees on the Tibetan Plateau

: The soil moisture (SM) influences tree growth with climate change. However, the spatial and temporal dynamics of tree water use strategies in climate-sensitive areas remain uncertain. Therefore, we collected the tree-ring oxygen isotope ( δ 18 O TR ) chronologies and divided the wet–dry gradients according to the precipitation on the Tibetan Plateau (TP). Further, the relationship between the δ 18 O TR and environmental factors was analyzed across different gradients. We found the following: (1) The SM during the growing season was the most important factor for δ 18 O TR . (2) The response of the δ 18 O TR to the SM had a lag in arid areas than in humid areas. (3) Trees absorbed the SM on the surface in humid areas ( r = − 0.49 to − 0.41, p < 0.01), while trees absorbed the SM from deep in the soil in arid areas ( r = − 0.48 to − 0.29, p < 0.01). The results demonstrated that trees were better able to cope with drought stress in arid regions because they used more stable deep soil water than in humid regions. Therefore, the findings will provide a scientific basis for water use of trees using the δ 18 O TR in complex environmental contexts. Trees with single water use strategies should be given more attention to keep ecosystems healthy.


Introduction
Climate change has profound impacts on forest ecosystems [1,2].Tree growth is mainly influenced by temperature, precipitation, and their interactions [3,4].It is well known that tree growth in arid regions is controlled by soil moisture (SM), while in humid regions, it is mainly influenced by temperature [5].Therefore, there is variability in the environmental sensitivity of trees in different climatic contexts.
The SM is the main source to sustain plant growth [6].It plays a vital role in maintaining ecosystem health.The SM varies with soil seasons and depths [7].There are also differences in the uptake strategies of SM by trees in different moisture environments [8].Some studies have found that the SM plays a role in tree growth on the Tibetan Plateau (TP) [9,10], especially during the growing season [11].Unfortunately, it is not practical to collect soil samples when conducting large-scale studies, making it exceptionally difficult to use measured soil data to study tree water use strategies.However, tree-ring oxygen isotope (δ 18 O TR ) is now used as an indicator to study long-term changes in the SM because of their stability [12].For example, the SM for nearly a century has been reconstructed using the δ 18 O TR in the Western Himalayas [13].
The δ 18 O TR is jointly controlled by the coupled processes of transpiration and water uptake in tree root systems [14].Therefore, the δ 18 O TR is an important indicator that can record the response of trees to seasonal climate change [15].A complex set of δ 18 O TR formation processes are influenced by external climatic factors [16,17].Trees absorb precipitation entering the soil and then preserve the precipitation signal in the δ 18 O TR through transpiration fractionation in the leaves [18,19].
The precipitation in monsoon areas has significant seasonal variations, and similar variations occur in the SM [20,21].Thus, the δ 18 O of SM varies at different depths and seasons.The δ 18 O TR stores SM signals through complex physiological mechanisms [22].For example, it has been found that the δ 18 O of Pinus tabulaeformis in the East Asian monsoon (EAM) zone is controlled by the SM [23].And it has been shown that trees used the shallow SM in the wet season and deep water in the dry season in the EAM region [24].
The TP has been found to have a trend of increasing temperatures [25] and decreasing precipitation in recent years due to shifts in the Asian summer monsoons (ASMs) [26,27].The δ 18 O TR was consequently affected in this region [28,29].At present, there are many articles on the use of δ 18 O TR to study climate reconstruction and the interpretation of physiological mechanism on the TP [30][31][32].There are few studies that have been conducted to use the δ 18 O TR to study water use strategies.Therefore, the purpose of this paper is to explore three main questions on the TP: (1) What are the seasonal dynamics of different gradients of SM use? (2) Is there a difference in the depth of water used by trees at different gradients?(3) Are the seasons as well as soil depths of different gradients of trees water use coupled?

Study Areas
As shown in Figure 1, we selected a total of six sampling points.The Delingha region (DE) is in the northeastern part of the TP.The Ganesh (GA) and Wache (WA) regions are in the southern of the TP.The Manali (MA) and Jageshwar (JA) regions are in the southwestern part of the TP.And the Karakorun region (KA) is in the west of the TP.The JA is the wettest region, and the DE is the driest region in our study (Figure 2).The difference in the annual total precipitation between these two regions is 1311 mm.On the basis of the annual total precipitation, the six regions are classified into humid areas (JA, WA, and GA regions) and arid areas (MA, KA, and DE regions).record the response of trees to seasonal climate change [15].A complex set of δ 18 OTR formation processes are influenced by external climatic factors [16,17].Trees absorb precipitation entering the soil and then preserve the precipitation signal in the δ 18 OTR through transpiration fractionation in the leaves [18,19].The precipitation in monsoon areas has significant seasonal variations, and similar variations occur in the SM [20,21].Thus, the δ 18 O of SM varies at different depths and seasons.The δ 18 OTR stores SM signals through complex physiological mechanisms [22].For example, it has been found that the δ 18 O of Pinus tabulaeformis in the East Asian monsoon (EAM) zone is controlled by the SM [23].And it has been shown that trees used the shallow SM in the wet season and deep water in the dry season in the EAM region [24].
The TP has been found to have a trend of increasing temperatures [25] and decreasing precipitation in recent years due to shifts in the Asian summer monsoons (ASMs) [26,27].The δ 18 OTR was consequently affected in this region [28,29].At present, there are many articles on the use of δ 18 OTR to study climate reconstruction and the interpretation of physiological mechanism on the TP [30][31][32].There are few studies that have been conducted to use the δ 18 OTR to study water use strategies.Therefore, the purpose of this paper is to explore three main questions on the TP: (1) What are the seasonal dynamics of different gradients of SM use? (2) Is there a difference in the depth of water used by trees at different gradients?(3) Are the seasons as well as soil depths of different gradients of trees water use coupled?

Study Areas
As shown in Figure 1, we selected a total of six sampling points.The Delingha region (DE) is in the northeastern part of the TP.The Ganesh (GA) and Wache (WA) regions are in the southern of the TP.The Manali (MA) and Jageshwar (JA) regions are in the southwestern part of the TP.And the Karakorun region (KA) is in the west of the TP.The JA is the wettest region, and the DE is the driest region in our study (Figure 2).The difference in the annual total precipitation between these two regions is 1311 mm.On the basis of the annual total precipitation, the six regions are classified into humid areas (JA, WA, and GA regions) and arid areas (MA, KA, and DE regions).

Tree-Ring Isotope Data
The δ 18 OTR data were downloaded from the NOAA Paleoclimatology Datasets (https://www.ncei.noaa.gov,accessed on 26 August 2023).Table 1 provides specific information on the six sampling sites.The experimental methods were checked when downloading data to ensure that the same experimental methods did not affect the accuracy of the study [33].Some scholars have already found that elevation and age do not affect the δ 18 OTR [34,35].

Tree-Ring Isotope Data
The δ 18 O TR data were downloaded from the NOAA Paleoclimatology Datasets (https://www.ncei.noaa.gov,accessed on 26 August 2023).Table 1 provides specific information on the six sampling sites.The experimental methods were checked when downloading data to ensure that the same experimental methods did not affect the accuracy of the study [33].Some scholars have already found that elevation and age do not affect the δ 18 O TR [34,35].

The SM Data
The SM data was obtained from the Global Land Data Assimilation System (http://Idas.gsfc.nasa.gov/gldas,accessed on 29 August 2023).It spans the period of 1948-2015, with a spatial resolution of 0.25 • × 0.25 • .Due to its good temporal and spatial resolutions, it has been used for a large number of studies in recent years.The SM data of three depths of 0−100 cm were selected.We transform the SM data (unit: kg/m 2 ) into volumetric water content (VWC, unit: m 3 /m 3 ) according to Equation ( 4).

Data Analysis
Pearson's correlation analysis as well as partial correlation analysis were used to study the main climatic factors affecting δ 18 O TR .The variability in the δ 18 O TR chronology was tested using one-way analysis of variance (ANOVA) in different regions.Meanwhile, the spatial correlation analysis of the δ 18 O TR and the SM was performed using MATLAB 2023.

The Spatial and Temporal Patterns of the δ 18 O TR
Figure 3 shows that the values of the δ 18 O TR in the arid areas as a whole (31.36‰) were larger than those in the humid areas (24.34‰) on the TP.But the dynamics of δ 18 O TR showed inconsistent spatial and temporal variation characteristics across wet and dry gradients.In addition, the values of δ 18 O TR showed an increasing trend in all regions except the DE region.And the increasing trend was overall more pronounced in the humid areas.The spatial dynamics of the δ 18 OTR showed that, as the climate becomes wetter, the values of the δ 18 OTR declined (Figure 4).The δ 18 OTR values are the highest (DE, 32.90‰) on the northern of the TP.On the contrary, the δ 18 OTR values are the lowest (WA, 19.40‰) on the southern of the TP.The spatial dynamics of the δ 18 O TR showed that, as the climate becomes wetter, the values of the δ 18 O TR declined (Figure 4).The δ 18 O TR values are the highest (DE, 32.90‰) on the northern of the TP.On the contrary, the δ 18 O TR values are the lowest (WA, 19.40‰) on the southern of the TP.The spatial dynamics of the δ 18 OTR showed that, as the climate becomes wetter, the values of the δ 18 OTR declined (Figure 4).The δ 18 OTR values are the highest (DE, 32.90‰) on the northern of the TP.On the contrary, the δ 18 OTR values are the lowest (WA, 19.40‰) on the southern of the TP.

The SM Dynamics at Different Depth Gradients
Figure 5 shows that the SM at different depths exhibits inconsistent variability characteristics on the TP (Figure 5).Specifically, the SM increased with a deeper soil depth (Figure 5a).And the SM in the DE region is much lower than in other regions.

The SM Dynamics at Different Depth Gradients
Figure 5 shows that the SM at different depths exhibits inconsistent variability characteristics on the TP (Figure 5).Specifically, the SM increased with a deeper soil depth (Figure 5a).And the SM in the DE region is much lower than in other regions.Figure 5b shows that the seasonal dynamics of SM are more pronounced in humid Figure 5b shows that the seasonal dynamics of SM are more pronounced in humid areas.The high values of SM occur from June to September in humid areas (VMC Jun-Sep > 45%).However, this phenomenon does not occur in arid areas.In the MA region, the high values of SM occur from June to September (VMC Jun-Sep > 35%).The high values of VMC occur from April to June (VMC Apr-Jun > 32%) in the KA region.But no seasonal dynamic was observed in the SM in the driest DE region.
The wettest JA region showed a decreasing trend for the SM, while the driest DE region showed an increasing trend for the SM (Figure 6).Overall, the surface SM increased in the wetter areas, while middle and deeper SM decreased.In arid areas, the SM decreased by 40-100 cm and increased by 0-40 cm.Meanwhile, it was found that the SM in humid areas fitted better with precipitation (Figure 7).

The Response of δ 18 O TR to Climatic and SM Factors
It is revealed that the overall δ 18 O TR responded negatively to PDSI, PRCP, and RH (Figure 8).The δ 18 O TR is positively correlated with T, Tmax, Tmin, and VPD in study areas (Figure 8).And δ 18 O TR in the humid areas are mainly influenced by climatic factors during the middle of the growing season in the humid areas and during the late growing season in the arid areas.

The Response of δ 18 OTR to Climatic and SM Factors
It is revealed that the overall δ 18 OTR responded negatively to PDSI, PRCP, and RH (Figure 8).The δ 18 OTR is positively correlated with T, Tmax, Tmin, and VPD in study areas (Figure 8).And δ 18 OTR in the humid areas are mainly influenced by climatic factors during the middle of the growing season in the humid areas and during the late growing season in the arid areas.The response of the δ 18 OTR to the SM in terms of season and depth were further investigated (Figure 9).The correlation between the δ 18 OTR and the SM was greater with SM0-10 cm and SM10-40 cm cm in July in humid areas.However, the δ 18 OTR was significantly negatively correlated with SM10-40 cm and SM40-100 cm in September in arid areas.The seasonal response of the δ 18 OTR to the SM lagged in arid areas.Specifically, the δ 18 OTR was significantly negatively correlated with SM at 0-40 cm (rJun-Aug = −0.49to −0.41, p < 0.05) in the JA and WA regions, while it was negatively correlated with the SM at all depths (rJun-Aug = −0.44 to −0.26, p < 0.05) in the WA region.However, the δ 18 OTR was significantly negatively correlated at all depths in the MA region (rJun-Sep = −0.47 to −0.31, p < 0.05), while the δ 18 OTR was more significantly negatively correlated with SM40-100 cm from July to September (rJun-Sep = −0.48 to −0.29, p < 0.05) in the KA and DE regions.The response of the δ 18 O TR to the SM in terms of season and depth were further investigated (Figure 9).The correlation between the δ 18 O TR and the SM was greater with SM 0-10 cm and SM 10-40 cm cm in July in humid areas.However, the δ 18 O TR was significantly negatively correlated with SM 10-40 cm and SM 40-100 cm in September in arid areas.The seasonal response of the δ 18 O TR to the SM lagged in arid areas.Specifically, the δ 18 O TR was significantly negatively correlated with SM at 0-40 cm (r Jun-Aug = −0.49to −0.41, p < 0.05) in the JA and WA regions, while it was negatively correlated with the SM at all depths (r Jun-Aug = −0.44 to −0.26, p < 0.05) in the WA region.However, the δ 18 O TR was significantly negatively correlated at all depths in the MA region (r Jun-Sep = −0.47 to −0.31, p < 0.05), while the δ 18 O TR was more significantly negatively correlated with SM 40-100 cm from July to September (r Jun-Sep = −0.48 to −0.29, p < 0.05) in the KA and DE regions.

The Correlation between δ 18 O TR and Key Factors during the Growing Season
The δ 18 O TR and the SM showed high correlations during the growing season (June-September).Therefore, the partial correlation analysis was further used to study the most important climatic factors affecting the δ 18 O TR (Table 2).The results found that δ 18 O TR was significantly negatively correlated with the SM at 0-10 cm in the JA region when VPD and RH were controlled.And the δ 18 O TR was significantly correlated with the SM at 0-10 and 10-40 cm in the WA and GA regions.The δ 18 O TR was highly significantly negatively correlated with SM at all depths in the MA region when VPD and RH were controlled.However, the δ 18 O TR was significantly negatively correlated with the SM at 40-100 cm in the KA and DE regions.

The Correlation between δ 18 OTR and Key Factors during the Growing Season
The δ 18 OTR and the SM showed high correlations during the growing season (June-September).Therefore, the partial correlation analysis was further used to study the most important climatic factors affecting the δ 18 OTR (Table 2).The results found that δ 18 OTR was significantly negatively correlated with the SM at 0-10 cm in the JA region when VPD and RH were controlled.And the δ 18 OTR was significantly correlated with the SM at 0-10 and 10-40 cm in the WA and GA regions.The δ 18 OTR was highly significantly negatively correlated with SM at all depths in the MA region when VPD and RH were controlled.However, the δ 18 OTR was significantly negatively correlated with the SM at 40-100 cm in the KA and DE regions.

The Spatial Patterns of δ 18 O TR Response to SM with Soil Depths
The response of the δ 18 O TR to the SM was generally negatively correlated in the study areas (Figure 10).But the correlation of the δ 18 O TR to the SM at 40-100 cm is stronger than the correlation to the SM at 0-40 cm (Figure 10).And the responses of the δ 18 O TR to the SM changed from negative to positive as the soil depth deepened in the southern and northeastern parts of the TP (Figure 10).The correlation decreases from the JA region to the DLH region with the SM at 0-10 cm.That means that, as the climate becomes drier, the correlation decreases between the δ 18 O TR and the SM at 0-40 cm.However, the positive responses of the δ 18 O TR to the SM at 40-100 cm increase in the DE region.
the correlation to the SM at 0-40 cm (Figure 10).And the responses of the δ 18 OTR to the SM changed from negative to positive as the soil depth deepened in the southern and northeastern parts of the TP (Figure 10).The correlation decreases from the JA region to the DLH region with the SM at 0-10 cm.That means that, as the climate becomes drier, the correlation decreases between the δ 18 OTR and the SM at 0-40 cm.However, the positive responses of the δ 18 OTR to the SM at 40-100 cm increase in the DE region.The temperature has shown a continuous warming trend [41], and precipitation has increased in the eastern and western parts [42,43] and decreased in the southwestern part of the TP since the 1960s [44].Trees will adopt multiple water use strategies to increase the chances of survival in this context.It has been shown that the δ 18 O TR could reflect the water uptake of seasons and the depths in trees [19].This means that the δ 18 O TR can record soil water δ 18 O information for different seasons and depths [31].
The δ 18 O TR is mainly controlled by the coupled processes of source water and external climate [15].In our results, the δ 18 O TR was more sensitive to the SM than to temperature during the growing season (Figures 8 and 9).The δ 18 O TR is derived from the SM, but the δ 18 O of SM varies with soil depths [45].The isotopic signals in the SM are preserved in the δ 18 O TR through the xylem during the growing season [46].Therefore, the δ 18 O TR can be used as one of the indicators of SM use in trees.
At the same time, the values of the δ 18 O TR were greater in arid areas (Figure 3).Lower SM leads to an increase in the ratio of internal leaf CO 2 (Ci) to atmospheric CO 2 (Ca) in leaves, which enhances the evaporative enrichment of leaf water in arid areas [47].Then, the tree root system absorbs heavy-source water, and the δ 18 O TR is increased further [15].During the conversion of sucrose produced by tree photosynthesis to cellulose, it undergoes a partial oxygen isotope signal exchange in the source water [48].In other words, the δ 18 O TR keeps the mixed signal of source water and leaf evaporation [49].These two processes result in a higher δ 18 O TR in arid areas.On the contrary, the decrease in the δ 18 O TR is due to the reduced evaporation and makes leaf water enrichment weaker in humid areas [50].

Seasonal Use Strategies of SM for Trees across Wet and Dry Gradients
In climate-sensitive areas, the SM is critical for tree growth.The life history stage [24], season [51], tree species [52], ability of plant access to water [53], and the amount of recent rainfall [14] all have an impact on tree water use strategies.In our study, the δ 18 O TR was significantly correlated with the SM during the growing season, but the δ 18 O TR responded to the SM later in arid areas (Figure 9).The δ 18 O TR is affected by the SM from June to August in humid areas (Figure 9).Because it is influenced by the ASM and receives adequate precipitation from it in these months, which further affects the δ 18 O TR .These results suggest that monsoon precipitation has a significant effect on tree growth (Figure 8).We believe that precipitation signals in the soil from the ASM are absorbed by tree root systems.Finally, the δ 18 O of precipitation is retained through the δ 18 O TR in humid areas.Our results have been confirmed in other studies on the TP [31,32].
The δ 18 O TR are affected by the SM from July to September in arid areas (Figure 9).It is worth noting that the response of the δ 18 O TR to the SM lagged in arid areas.We believe this is related to the extended growing season of trees on the TP [36].However, there are significant differences within the arid areas.The δ 18 O TR in the MA region was controlled by summer monsoon precipitation and RH [54].And trees mainly absorb winter snowfall in the KA region, which provides a stable water source [37].Therefore, the correlation between the δ 18 O TR and the summer SM is not high in the KA region.In the DE region, the δ 18 O TR was controlled by the SM from May to September [53], and this observation is consistent with that of our study.
The summer precipitation has a greater effect on the SM in the southeastern part of the TP [38].And the δ 18 O TR was significantly negatively correlated with precipitation during the growing season in Southern Mexico [55], Northwestern Tibetan Plateau [56], and Ordos Plateau [57].This is because summer precipitation replenishes plant water supplies [20].Therefore, the correlation between the SM and precipitation is higher in the southeastern part of the TP than in its northwestern part.

Depth Use Strategies of SM for Trees across Wet and Dry Gradients
As a typical climate-sensitive area, the TP has a very high spatial variability in precipitation [58].As a result, trees flexibly adjust their water use strategies to ensure growth.Trees use the deep SM preferentially in arid areas [44], whereas they use the shallow SM in humid areas [59].This is consistent with our results.However, the water use strategies for trees in complex environments deserve further research.
The δ 18 O TR was significantly negatively correlated with the SM at 0-10 cm in the JA region.The δ 18 O TR were significantly negatively correlated with the SM at 0-10 and 10-40 cm after the VPD or RH was controlled in the WA and GA regions (Table 2).This suggests that trees tended to take up the SM in the surface and middle layers in humid areas.Meanwhile, the surface soil can provide nutrients to the trees [4,60].In other words, trees preferentially absorb SM on the surface in humid areas.
The significant negative correlation between the δ 18 O TR and the SM at 40-100 cm in the KA and DE regions was revealed in arid areas (Table 2).In arid areas, capillary water induces trees to form deeper root systems to absorb deep soil water and groundwater to meet higher transpiration needs during the dry season [61,62].The highly significant negative correlation between the δ 18 O TR and the SM at all depths in the MA region can be explained by the fact that some SM isotope signals are masked by strong evaporation [63,64].On the other hand, the "bimodal" precipitation pattern may encourage trees to develop a well-developed root system in this region.
Trees have different coping strategies to climate extremes [60,65].Trees using a single water strategy will bear a greater risk of death [27,66].The chances of facing drought stress are increasing on the TP [67].Trees face greater challenges for survival in humid areas because of the single use of the shallow SM.We argue for the better monitoring of tree growth in the humid regions of this region.

Conclusions
In this study, we used the δ 18 O TR at six sites to investigate whether there are differences in water use strategies of trees across wet and dry gradients on the TP.The results indicated that the SM during the growing season was the main factor influencing tree growth.Trees mainly absorbed the shallow SM in the humid areas, whereas they mainly absorbed the deep SM in the arid areas.And the response of the δ 18 O TR to the SM in the arid areas lagged behind that in the humid areas.Our study provides a theoretical basis for the further use of δ 18 O TR to study how tree growth adapts to the SM changes.We argue that shallow soil water uptake is a threat to tree growth in the context of warming on the TP.Therefore, the monitoring of tree growth in the southeastern part of the region should be conducted further.

Figure 1 .
Figure 1.Geographical locations of study area and sampling sites.Figure 1. Geographical locations of study area and sampling sites.

Figure 1 .
Figure 1.Geographical locations of study area and sampling sites.Figure 1. Geographical locations of study area and sampling sites.

Figure 2 .
Figure 2. Meteorological elements of the study areas.

Figure 2 .
Figure 2. Meteorological elements of the study areas.

Forests 2024 , 16 Figure 3 .
Figure 3.The time series of the δ 18 OTR for each site.

Figure 3 .
Figure 3.The time series of the δ 18 O TR for each site.

Figure 3 .
Figure 3.The time series of the δ 18 OTR for each site.

Figure 4 .
Figure 4. Significance test for study areas.Inter-sample differences in δ 18 OTR values were assessed using one−way ANOVA, with different letters indicating significant differences (p < 0.05).

Figure 4 .
Figure 4. Significance test for study areas.Inter-sample differences in δ 18 O TR values were assessed using one−way ANOVA, with different letters indicating significant differences (p < 0.05).

Forests 2024 , 16 Figure 5 .
Figure 5. Dynamics of SM in different soil layers in different months.The squares in the boxplot diagram represent the median SM, different letters indicating significant differences (p < 0.05), the black dots represent the outliers and the range of the boxes represents the 25%-75% data interval (a).The blue line in the graph represents the 40−100 cm volumetric water content, the red line represents the 10−40 cm volumetric water content and the black line represents the 0−10 cm volumetric water content (b).

Figure 5 .
Figure 5. Dynamics of SM in different soil layers in different months.The squares in the boxplot diagram represent the median SM, different letters indicating significant differences (p < 0.05), the black

Forests 2024 , 16 Figure 7 .
Figure 7. Linear fi ing of SM and precipitation.The red area in the graph represents the 95% confidence interval and the red line represents the fi ed trend line.

Figure 7 .
Figure 7. Linear fitting of SM and precipitation.The red area in the graph represents the 95% confidence interval and the red line represents the fitted trend line.

Figure 10 .
Figure 10.The spatial correlation between SM and δ 18 OTR in June-September.The green point in the figure represents δ 18 OTR sampling points.The colors in the graph from red to blue to indicate that the correlation changes from negative to positive.

Figure 10 .
Figure 10.The spatial correlation between SM and δ 18 O TR in June-September.The green point in the figure represents δ 18 O TR sampling points.The colors in the graph from red to blue to indicate that the correlation changes from negative to positive.

1 .
The δ18 O TR Records Soil Water Use by Tree Growth on the TP

Table 1 .
The δ 18 O TR chronology information.

Table 2 .
Partial correlation analysis between δ18OTR and environmental factors from June-September.