Introduction

Shrublands account for a considerable part of terrestrial vegetation cover across the globe (Peng et al. 2021). Shrublands are widely distributed in harsh treeless sites as at high latitudes and high elevations, in particular, they are the unique woody species above the treeline (Liang et al. 2012). There is growing evidence that shrublands play a key role in the carbon sink, nutrient cycling, hydrological regulation and biodiversity maintenance in treeless biomes (Knapp et al. 2008; Myers-Smith et al. 2011; Pearson et al. 2013). Growth dynamics of shrubs under climate change are closely linked to the ecological services provided by arctic and alpine shrublands (Ding et al. 2021). Thus, understanding how shrub growth responds to climate change is highly relevant for ecological monitoring and conservation as well as bush management in the arctic and alpine regions (Gazol and Camarero 2012).

Climate warming had a significant impact on high-latitude biomes and already triggered substantial changes in the growth of arctic shrubs (Au and Tardif 2007; Post et al. 2009; Martin et al. 2017). Based on the analyses of annual ring-width series of shrubs, it was found that summer temperature was the main growth-limiting factor for deciduous (e.g., Betula nana and Salix pulchra) and evergreen shrubs (e.g., Juniperus communis subsp. nana) across large parts of the circumpolar north tundra (Hallinger et al. 2010; Blok et al. 2011). In some arid areas, shrub growth appeared to be constraint by soil moisture through either direct moisture limitation of photosynthesis or indirect effects on active layer depth (Elmendorf et al. 2012; Ackerman et al. 2017; Gamm et al. 2018). Besides, snow condition was found to exert opposing influences on the growth of deciduous shrubs in northeastern Greenland and the growth of evergreen shrubs in Canadian low arctic tundra (Schmidt et al. 2010; Christiansen et al. 2018). There has also been a report of diverging growth of coexisting trees and shrubs near the treeline in Russian Polar Urals (Pellizzari et al. 2017). Given this background, shrub-ring based studies should explicitly consider how regional warming and non-thermal variables (e.g., moisture) drive shrub growth dynamics, particularly in alpine regions where microclimate conditions rapidly change depending on elevation and topography (Barry 2008).

Alpine regions such as the Tibetan Plateau host some of the highest treelines and shrublines in the world representing the uppermost limits of the existence of woody plants (Miehe et al. 2007; Wang et al. 2015; Lu et al. 2021). Therefore, they provide an excellent opportunity for studying growth dynamics of woody plant communities along extreme altitudinal gradients. Dendroecological studies have shown that radial growth of alpine deciduous shrubs (e.g., Hippophae rhamnoides) sampled near the treeline was temperature sensitive on the northeastern Tibetan Plateau (Xiao et al. 2007). By contrast, due to the intense evaporation, pre-monsoon moisture condition was the main climatic variables constraining the radial growth of evergreen shrubs (Juniperus pingii var. wilsonii) above the treeline on the central Tibetan Plateau (Liang et al. 2012). The radial growth of evergreen shrubs (Rhododendron campanulatum) was mainly controlled by winter minimum temperature at the treeline in the central Himalaya (Panthi et al. 2021). Another study indicated that the radial growth of evergreen shrubs (Rhododendron nivale) and trees reaching the treeline (Abies georgei var. smithii) reflected the signal of mean July temperature on the southeastern Tibetan Plateau, where humid conditions dominate (Liang and Eckstein 2009; Liang et al. 2009). However, it remains unclear whether the climate-growth relationships for deciduous shrubs along the altitudinal gradients from subalpine shrublands to alpine shrubline would be consistent on the eastern Tibetan Plateau. The alpine willow Salix oritrepha Schneid. is a widespread deciduous shrub forming the upper shrubline across the eastern Tibetan Plateau and could be a model species for investigating variations of radial growth of deciduous shrubs along the altitudinal gradients up to the shrubline.

In this study, we aimed to fill this knowledge gap by comparing climate-growth relationships of S. oritrepha forming the shrubline along two elevational transects (4200–4600 m) located in the dry and wet regions on the eastern Tibetan Plateau. Our objectives were: (1) to characterize the growth patterns of the alpine willow shrubs in recent decades; (2) to compare the growth-limiting factor for willow shrubs along two elevational transects. Due to similarly cold climates on the eastern Tibetan Plateau, we hypothesize that the growth of shrubs along the altitudinal gradients will be temperature limited when seasonal moisture is sufficient, but will be moisture limited under warmer and drier climate conditions.

Materials and methods

Study area and climate

The study area is located on the eastern Tibetan Plateau (Fig. 1). Due to its complexity of geological structure and special atmospheric circulation, this area is highly sensitive to climate change (Zhao et al. 2015). The study area includes the wet forest region of Nangqian County (BZ hereafter) and the dry forest region of Leiwuqi County (LWQ hereafter) on the eastern Tibetan Plateau (Fig. 1). Based on the meteorological data in Nangqian County (96.28°E, 32.12°N, 3644 m), the mean annual precipitation is 535.8 mm and the mean annual temperature is 4.4 °C, July and January were the warmest (mean temperature of 13.2 °C) and coldest months (− 5.9 °C), respectively. According to the meteorological record in Leiwuqi County (96.36°E, 31.13°N, 3810 m), the mean annual precipitation is 606.7 mm and the mean annual temperature is 3.3 °C, the warmest and coldest months were July (mean temperature of 12.4 °C) and January (− 6.8 °C), respectively. The mean temperature during summer showed significant positive trends in these two regions during the past decades (p < 0.01). The warming rate of summer during the past decades in Leiwuqi County was slightly higher than that of Nangqian County (Fig. 2). The average summer and annual 3-month SPEI (Standardized Precipitation Evapotranspiration Index; cf. Vicente-Serrano et al. 2010) values in Leiwuqi County (mean ± SD; summer: − 0.01 ± 0.85; annual: 0.013 ± 0.63) were lower than in Nangqian County (summer: 0.043 ± 0.81; annual: 0.025 ± 0.61) between 1960 and 2019 (Fig. S1).

Fig. 1
figure 1

Locations of the shrub study sites (yellow squares) along two altitudinal transects (from 4200 to 4600 m) situated on the eastern Tibetan Plateau. Abbreviations: BZ (wet Nangqian County), LWQ (dry Leiwuqi County)

Full size image
Fig. 2
figure 2

Changes in mean summer temperature and precipitation in Nangqian County (a BZ region) and Leiwuqi County (b LWQ region)

Full size image

Study species

In BZ and LWQ, Balfour spruce (P. likiangensis var. balfouriana) is the dominant tree species between 3800 and 4300 m on the north-facing slopes, whereas S. oritrepha grows from 4200 to 4600 m and dominates the woody plant communities above the treeline (here located at 4300 m). S. oritrepha is one of the most widespread alpine shrub species on the eastern Tibetan Plateau (Fang et al. 2011). Previous studies confirmed that S. oritrepha can be used to conduct the dendrochronological studies in light of its clearly visible ring boundaries (Lu et al. 2016).

Field sampling and chronology development

In the field, 123 stem discs of S. oritrepha were randomly selected along the elevational gradients in the BZ and LWQ regions (Table 1). Shrub individuals were cut from the major stems at the root collar close to the soil to acquire older samples (Lu et al. 2019). Previous studies via a serial-sectioning method corroborated that S. oritrepha basal stems contain the utmost of rings (Lu et al. 2016). Wood samples were air-dried and were sanded with gradually finer sandpapers to distinguish annual rings. Rings were measured with TSAP-LINTAB 6 system (accuracy: 0.001 mm). The wood samples cross-dating was then checked using the COFECHA software (Holmes 1983).

Table 1 Descriptive information of BAI chronologies for the alpine shrub (S. oritrepha)
Full size table

To retain potentially important low-frequency signal, we converted ring widths into basal area increment (BAI) given the concentric radial growth of S. oritrepha shrub. We then removed the biological growth trend in individual BAI series with a negative exponential curve and calculated the mean, detrended BAI site series using bi-weight robust means in ARSTAN programme (Cook 1985). In this way, standard BAI chronologies were developed at each study site. Then, we calculated the commonly used dendrochronological statistics (see details in Table 1) on standard BAI series to measure the internal coherence of BAI series within each site. An EPS threshold above 0.85 was defined to assess the reliability of the established chronologies for the common period (Cook and Kairiukstis 2013). Meanwhile, the standard chronologies of ring-width indices (RWI) were also built in the same way for further comparison.

Data analyses

To calculate climate-growth correlations, monthly climate data (temperature, precipitation) from local meteorological station were used for succeeding analyses. First, standard BAI site chronologies of shrubs were related to monthly variables (mean temperature, mean maximum and minimum temperatures, total precipitation) considering regional climate data. Relationships were assessed from September of the previous year to September of the current year using Pearson correlations. We considered the periods when EPS was above the 0.85 threshold within each site (Table 1). Among the monthly temperature variables, only the most significant temperature variable was shown in the results. Second, linear mixed-effects models (LMEs) were applied to quantify the best climatic predictor of shrub growth at each site. Climatic factors were set as fixed effects in the LMEs, whilst years and individuals were set as random effects to explain the nonindependence of data within years and individuals (Crawley 2007). The models with the fewest explanatory variables and the minimum Akaike Information Criterion (AIC) were selected (Wagenmakers 2003). We ran the LMEs using the nlme package and R software (Pinheiro et al. 2020).

Results

Chronology development and growth pattern

Six shrub standard BAI and RWI chronologies were established along two elevational gradients encompassing the shrubline and the treeline. No missing rings were found (Table 1, Fig. 3). The longest (41 years) and shortest (26 years) S. oritrepha chronologies were built in sites LWQ4600 and LWQ4200, respectively. Along the altitudinal transect in BZ, the mean ring-width of study shrub varied from 0.26 to 0.33 mm, whilst the average BAI ranged from 4.63 mm2 to 5.47 mm2 (Table 1). Similarly, along the altitudinal transect in LWQ, the mean shrub-ring width ranged between 0.26 and 0.36 mm, whilst the BAI varied from 4.58 mm2 to 5.95 mm2 (Table 1). In general, all shrub BAI chronologies showed significant increasing trend in the past decades (p < 0.01), with higher mean values in the wet BZ region (0.35 mm2 yr−1) than in the dry LWQ region (0.28 mm2 yr−1). However, shrub growth declined since 2011 for both regions with only three sites reaching the significant level (p < 0.05) (Table 2).

Fig. 3
figure 3

Standard shrub BAI chronologies along two altitudinal transects for the wet BZ (a) and dry LWQ (b) regions. Solid and dashed lines correspond to the BAI indices and the sample depth, respectively

Full size image
Table 2 The growth trends of shrub BAI chronologies since 2011 and their correlations with mean July SPEI (*p < 0.05)
Full size table

Climate-growth relationships

Both BAI and RWI chronologies were significantly and positively correlated with mean June or July temperature in the two study regions (Fig. 4; Fig. S2). The BAI chronologies were more responsive to growing-season temperatures than the RWI chronologies. Therefore, we used shrub BAI chronologies for further analyses. In BZ, all S. oritrepha BAI chronologies showed significantly positive responses to mean July temperature (p < 0.01) regardless of the elevation (Fig. 4a). However, no significant precipitation-growth correlations were detected in this region (Fig. 4c). In LWQ, shrub BAI chronologies were significantly and positively associated with mean July temperature in two elevational sites (LWQ4200, p < 0.01; LWQ4400, p < 0.05), whereas another elevational site (LWQ4600) responded positively and significantly to mean June temperature (p < 0.05) (Fig. 4b). At site LWQ4600, shrub growth was significantly and negatively associated with precipitation of previous December (p < 0.05). At site LWQ4400, it responded positively and significantly to precipitation of previous October and negatively with March precipitation (p < 0.05) (Fig. 4d). Specifically, five out of six sites showed significant positive responses to mean summer temperature before 2011 (Table 3). In the period 2011–2018, correlations between growth and summer temperature were not significant observed (Table 3).

Fig. 4
figure 4

Correlation coefficients calculated between the shrub BAI chronologies and climatic variables (mean monthly temperature and monthly precipitation from September of the prior year (lowercase letters: s, o, n, d) to September of the current year (uppercase letters: J, F, M, etc.)) at different elevations in the wet BZ (a, c) and dry LWQ regions (b, d). Tm mean monthly temperature, Pr monthly precipitation. The dashed lines indicate significant levels of p < 0.05 (black, red and blue dashed lines correspond to the sites at elevations of 4200, 4400 and 4600 m, respectively). The highest correlation for each site is indicated by an asterisk

Full size image
Table 3 Correlation coefficients calculated between the shrub BAI chronologies and climatic variables (mean summer temperature and summer total precipitation) during the two sub-periods (pre-2011 and 2011–2018)
Full size table

As shown by the LMEs, mean July temperature was the best predictor of shrub growth at three elevational sites (BZ4200, BZ4400, BZ4600) in BZ (Table 4). In LWQ, the best predictor of shrub growth was July (sites LWQ4200 and LWQ4400) or June temperature (site LWQ4600) (Table 4).

Table 4 Summary of the best-fitted linear mixed-effects models (LMEs) fitted to growth data at six shrub sites (*p < 0.05; **p < 0.01)
Full size table

Discussion

We showed that suitable shrub-ring width and BAI series are sensitive proxies of climatic impacts on mountainous woody plant communities encompassing the shrubline. Specifically, the series of the deciduous alpine willow shrub S. oritrepha can be used to quantify climate-growth relationships in treeless mountain regions.

Along two elevational gradients, the radial growth of S. oritrepha decreased with increasing elevation as temperature decreased. This result was in line with previous studies on evergreen shrubs of eastern Tibetan Plateau and deciduous shrubs in central Alps (Rixen et al. 2010; Lu et al. 2015; Boscutti et al. 2018). The result was also similar to the findings on tree growth along the altitudinal gradients from wet regions in Asia, Europe and South America (Di Filippo et al. 2007; Massaccesi et al. 2008; Liang et al. 2010), confirming the dominant role of thermal factors in those sites with abundant moisture. Generally, altitudinal gradients in high-elevation forest regions are mainly characterized by colder climates upwards, which limit the physiological activities (e.g., photosynthesis and respiration rates as well as the cambial activity) of alpine woody plants (Körner 2007; Gaire et al. 2020). Besides, decreased soil nutrient availability due to the reduction of soil temperature (Körner 2003) may also result in the decline of radial growth observed in the uppermost shrubs, which were not the youngest populations.

The results of correlation analyses indicated that mean June or July temperature was the main climatic factor constraining the radial growth of S. oritrepha along two altitudinal transects up to the shrubline on the eastern Tibetan Plateau. Despite the distinct regional climatic conditions, the shrubs of the two forest regions showed similar growth responses. These results could be explained by the fact that June and July are the warmest months of the year in these regions and drive photosynthesis rates, cambial activity and the radial growth of alpine woody plants (Liang and Eckstein 2009; Li et al. 2013, 2016). All the shrub BAI series showed the significant increasing trends during the past decades, implying that climatic warming (particularly summer warming) would enhance shrub growth when moisture regime is sufficient. Interestingly, the correlation between July temperature and shrub growth decreased upwards in the two regions, suggesting that moisture regime could also be a relevant driver of shrub growth there. However, all precipitation-growth correlations were not significant in six shrub sites during the growing season. Given that the climatic sensitivity of woody plant growth is largely age-dependent (Rozas et al. 2009), the mean length of shrub chronology increased with increasing elevation in the two study regions suggesting that shrub age could mediate the growth-temperature relationships. In addition, shrub growth might be influenced by local factors such as microenvironmental conditions. For instance, microtopography, soil nutrition availability and biotic interactions could create site-specific microhabitats which impact on growth-temperature associations (HilleRisLambers et al. 2013; Ellison et al. 2019; Mu et al. 2021).

Shrub BAI chronologies in the two regions declined over the last decade. They were positively associated with main growing-season SPEI suggesting that post-2010 shrub growth could be moisture limited. The shrub growth responses to summer temperature showed a non-significant trend in recent decade. The growth decline under warm-dry climatic conditions was also observed in Tibetan juniper (Juniperus tibetica Kom) on the eastern Tibetan Plateau (Mou et al. 2019). However, we did not detect significant shrub growth-SPEI correlations, which could be attributed to the short time span of their ring-width series. Significant decreasing growth trends since 2011 only occurred at three shrub sites, possibly indicating the impacts of local-site conditions (e.g., microtopography, soil nutrients and microclimate) (Mu et al. 2021). In particular, shrub meristems are more coupled to soil microclimate conditions than tall tree meristems which usually react quickly to changes in atmospheric climate conditions (Körner 2003). Further monitoring of shrub growth is necessary to test if the hotter and drier climate in the near future can lead to the shift in limiting factors for alpine shrubs (Wu et al. 2019; Buchwal et al. 2020). Collectively, shrub growth would benefit from climate warming when the moisture content is high enough, whereas shrub growth could be suppressed if the warmer and drier climate conditions prevail, thus supporting our hypothesis.

Conclusions

This study is one of the first comparing climate-growth relationships of shrubs along the altitudinal gradients encompassing the uppermost shrubline on the eastern Tibetan Plateau. The radial growth of shrubs was enhanced by warmer June or July conditions in the wet BZ region and the relatively dry LWQ region regardless of site elevation. Overall, the growth of shrubs will benefit from climate warming when seasonal moisture is relatively sufficient, but the increasing trend in shrub growth would be reversed due to the warming-induced drought stress. This study provides useful information for understanding climate-growth relationships of shrubs along wide altitudinal gradients in mountain treeless ecosystems. Further measurements of in situ microclimate conditions (soil moisture and temperature) combined with key plant functional traits are needed to clarify the mechanisms for temperature-growth couplings of high-elevation shrubs.

Author contribution statement

YW and JJC designed the research; YW and YH did the field work; YH did the shrub-ring experiment; YH, YW, BL, RH and JJC analysed the data and wrote the paper.