The research area is located in Gongbu Nature Reserve, Gongbu Jiangda County, Nyingchi City, Tibet Autonomous Region (E93°13′–94°50′, N28°40′–30°20′). The annual average temperature is above 7°C, and the frost-free period is more than 150 days. The annual average precipitation and evaporation are 500–700 mm and 1,300–1,700 mm, respectively. The annual average relative humidity is 50–75%, the annual sunshine hours are more than 1,900 h. The area has a large altitudinal difference (2,300–6,060 m), with an average altitude of more than 3,500 m, and the vertical distribution of plants are obvious. The main trees distributed in the region include dark coniferous forest dominated by Abies georgei var. smithii (Viguie et Gaussen) Cheng et L. and Picea likiangensis var. linzhiensis Cheng et L.F. Fu, bright coniferous forest dominated by Pinus densata, evergreen broadleaf hard forest dominated by Quercus aquifolioides Rehd. et Wils., and deciduous broad-leaved forest dominated by Betula platyphylla Suk. and Populus davidiana Dode.

Samples were collected according to the standardized measurement manual of plant functional traits (Cornelissen et al., 2003; Pérez-Harguindeguy et al., 2013; Moretti et al., 2016). At the end of the growing season for alpine plants on the Tibetan plateau in September 2020, we established sample plots in mature forests at altitudes of 3,186, 3,291, 3,417, 3,519, and 3,620 m above sea level. The general information of each sample plot along altitude gradient is shown in Table 1. Three 20 m × 20 m quadrates were built at each altitude, 15 quadrates in total. The altitude, location, slope (SL), slope aspect (SP), vegetation coverage, and soil temperature, soil moisture, soil bulk density of each quadrates were measured. In each quadrates, three healthy mature individuals of Quercus aquifolioides with similar DBH and stand age were selected. Four well-grown, unshaded branches from the middle of the tree were collected, followed by five mature leaves and three current year twigs from each branch. A total of 180 mature leaves and 108 current year twigs were collected at a single altitude. Samples were marked, placed into plastic bags, and stored in a 4°C incubator, and taken to the laboratory to complete the determination of functional traits within 24 h.

All the statistics were calculated using SPSS software (v.25.0, IBM Corp., Armonk, NY, United States) and plotting was performed on Origin 2021b (Origin lab, Northampton, Massachusetts, United States). A p-value < 0.05 was considered for assessing significant statistical effects. One-way ANOVA and least significant difference method (LSD) were used to analyze the significance of the differences in functional traits among different altitude gradients. Only the functional indicators with significant differences among altitudes were selected, the relationship between them and altitudes were analyzed by regression, and the optimal regression model was selected in accordance with the regression coefficient. Pearson correlation analysis was performed to explore the relationships between functional traits. The functional traits of each leaf and current-year twig were selected as data of species. Three topographical factors [altitude (AL), slope (SL), and slope aspect (SP)] and 11 soil factors (pH, SBD, ST, SW, TN, TP, TK, AP, AK, TOC, DOC) were used as environmental data. Redundancy analysis (RDA) was performed using Canoco 5.0 (Microcomputer Power, United States) to screen the main functional traits of the leaves and current-year twigs of Quercus aquifolioides and evaluate the effects of environmental factors on the functional traits of Quercus aquifolioides populations. Prior to RDA, the significance of the effect of each variable was assessed using a Monte Carlo permutation test.

The variation characteristics of leaf functional traits (LL/LW, LA, SLA, LV, LD, and RWC) along altitude gradient are shown in Figure 1. All the functional traits were significantly different (p < 0.001, Table 3) among altitudes to some extent, and they showed linear variation with the altitude. LL/LW and LD linearly increased with the altitude (p < 0.001), whereas LA, SLA, LV, and RWC linearly decreased (p < 0.001). The variation characteristics of the functional traits of current-year twigs (TL, TV, TD, ILG, DGT, and HV) along altitude gradients are shown in Figure 2. All the functional traits were significantly different (p < 0.05), and they showed linear variation with altitude. TL (p < 0.05), TV (p < 0.001), and LEI (p < 0.05) linearly increased with the increase in altitude, whereas TD, TLD, and HV linearly decreased (p < 0.05).

Pearson correlation analysis results showed that among the indicators of leaf functional traits, LA had a very significant positive correlation with SLA (r = 0.361, p < 0.001) (Figure 3), but this significantly positive relationship just existed in 3,186 and 3,620 m (Figure 4 and Supplementary Table 2). LA was significantly negatively correlated with LDMC (r = −0.362, p < 0.001) and LD (r = −0.901, p < 0.001) (Figure 3), which were only existed in low altitude (Figure 4 and Supplementary Table 2). SLA showed a significant positive correlation with LA (r = 0.361, p < 0.001), LW (r = 0.391, p < 0.001) and LPM (r = 0.320, p < 0.001) (Figure 3), meanwhile, showed significantly negatively relationship with LDMC (r = −0.437, p < 0.001) and LD (r = −0.518, p < 0.001). These significantly negatively relationship existed at five altitudinal gradients except for 3,519 m (Figure 4 and Supplementary Table 2). LDMC showed significant negative correlation with LA (r = −0.275, p < 0.001) and LV (r = −0.160, p = 0.016) and significant positive correlation with LD (r = 0.291, p < 0.001) and RWC (r = 0.383, p < 0.001).

TLN was positively correlated with TLA (r = 0.767, p < 0.001) at five altitudinal gradients (Figures 5, 6 and Supplementary Table 2), but negatively correlated with TLD (r = −0.310, p < 0.001). It was found in the Pearson correlation analysis between functional traits of all current-year twigs that there was no relationship between TLA and HV (Figure 5), but TLA was significantly positively related with HV at each altitude (Figure 6 and Supplementary Table 2). LA-single only showed positive correlation with HV (r = 0.560, p < 0.001). LSE was positively correlated with TLA (r = 0.743, p < 0.001), LEI (r = 0.658, p < 0.001) (Figure 5). TLD was negatively correlated with TL (r = −0.625, p < 0.001), TBD (r = −0.629, p < 0.001), TV (r = −0.559, p < 0.001), TDW (r = −0.518, p < 0.001) and TLA (r = −0.343, p < 0.001), respectively (Figure 5).

The cumulative percentage of species-environment relationship between axis 1 and axis 2 was 98.29 and 99.9%, respectively, indicating that there was a high correlation between functional traits and environmental factors, which could clearly reflect the influence degree of dominant factors on leaf and twig functional traits (Table 4). The total contribution value of SL, TN, SBD, pH and DOC to 14 leaf functional traits reached 99.9% (Table 5). SL and TN contributed the most to leaf functional traits and had significant correlation (p < 0.05). The total contribution value of AL, SBD, SL, ST and pH to 14 functional traits of current-year twigs reached 99.9% (Table 5), and AL had the highest contribution value to functional traits of current-year twigs and had significant correlation (p < 0.05). Figure 7 clearly showed the correlation between leading factors and functional traits (factors with contribution rate < 0.1 and p value > 0.9 were not shown in Figure 7). In Figure 7A, LT and LDMC are located in the upper left corner of the RDA ranking diagram, and are positively correlated with DOC, but negatively correlated with pH. LD and LL/LW were located at the left end of axis 1, which were positively correlated with SL and negatively correlated with TN. The other leaf functional traits were located in the upper right corner of the sequence diagram, and were positively correlated with TN and SBD, but negatively correlated with SL. In Figure 7B, LA-single and TLD are located in the upper left corner of the RDA ranking diagram, and are negatively correlated with AL and SL. TD, LSE and HV at the left end of axis 1 were positively correlated with ST and SBD, but negatively correlated with pH. Other functional traits of current year twigs were located in the lower right corner of the sequence diagram, and were positively correlated with AL and SL.

The present study showed that LL/LW and LD increased with altitude, whereas LA, SLA, and RWC decreased (Figure 1). The increase in LL/LW indicated that leaves gradually became longer, and the photosynthates of Quercus aquifolioides were used for the longitudinal extension of LA. The reason for the decrease in LA with the increase in altitude may be that the efficiency of light energy acquisition ability of leaves decreased under low-temperature environment, and leaves needed to reduce the rate of respiration and transpiration by reducing LAs (Craine and Lee, 2003). Meanwhile, high altitude and low temperature make the growth rate of trees slow down, LA decreases, photosynthetic rate decreases, net growth decreases, plant cell volume decreases, cell wall thickens, plant leaves become harder, and more photosynthates accumulate to prepare for the winter and the next year’s growth (Hölscher et al., 2002; Wright et al., 2002). SLA decreased with increasing altitude, indicating that in high-altitude areas, Quercus aquifolioides has a better ability to retain the acquired resources and could better adapt to resource-poor high-altitude habitats, which is an adaptation to limited resources (Niu et al., 2020). Combined with higher speed winds at high altitude, it is likely to cause damage to the plants, and the reduced LA makes SLA smaller. In low-altitude areas with relatively abundant resources, high SLA and small LD lead to a large light capture area of the leaves. Thus, plants have a higher production capacity. RWC is not only closely related to photosynthesis, but it also reflects the water and heat conditions of the plant’s living environment to a certain extent (Siefert et al., 2015; Wang et al., 2019). In the present study, RWC decreased linearly with the altitude gradient, which is closely related to the unique and drastic change trend of the altitude gradient of the Qinghai–Tibet Plateau. The water stress caused by the high altitude and low temperature produced lower soil water content, while the low RWC was one of the strategies of adaptation of Quercus aquifolioides to external water and heat conditions in Tibetan plateau.

TL (p < 0.05), TV (p < 0.001), and LEI (p < 0.05) increased linearly with the altitude, whereas TD, TLD, and HV decreased linearly (p < 0.05) (Figure 2). With the increase in altitude, the increase in the length of the twigs is needed for the upper competition resources. The spatial position of the plant branches in the community determines its competitiveness and the ability to obtain resources. A longer TL helps increase light acquisition and reproduction effectiveness and further improves the ability to withstand the harsh environment of high altitude (Sun et al., 2019; Wang et al., 2019). The present study revealed that as the altitude increased, TL increased, which led to an increase in TV. With the increase in TL, TDW and TV also increased. The LSE, TLD, and TD of the current-year twigs of Quercus aquifolioides at high altitude were lower than those in low-altitude areas. Studies have shown that while branchlets grow longitudinally, they also need radial growth to enhance their mechanical support, thereby enhancing the transport of water and nutrients (Pérez-Harguindeguy et al., 2013; Sun et al., 2019; Zhu et al., 2019). Thus, along altitude gradient, TSA increased slowly than TLA (Supplementary Table 1 and Supplementary Figure 1), resulting in a decrease in HV in high-altitude areas. Although the altitude distribution of twig functional traits of woody plants in mountainous areas of China have been widely discussed, no consistent conclusion is available. Zhu et al. (2019) found that the biomass of stems in mixed forests at mid-altitude levels was smaller than that of evergreen and broad-leaved forests at other altitudes in Wuyi Mountain. In the study on the response of stem and leaf traits of Rhododendron przealskii to altitude in Southeastern Qinghai–Tibet Plateau, no significant differences were found in the stem length, stem thickness, specific stem length, and stem biomass of each branch at different altitudes (p > 0.05) (He et al., 2021). In the present study, all twig functional traits demonstrated significant differences between altitudes. The reason is that the mountains in Southeast Tibet have obvious characteristics of mountain vertical belts, which are directly facing the Yarlung Zangbo River water vapor channel and are extremely sensitive to the Indian monsoon response. The correlation of temperature and precipitation between altitudes is significant (Spicer et al., 2003). Thus, the gradient changes in environmental factors lead to the heterogeneity of plant population development and distribution patterns.

Given that multiple traits are often coordinated with one another, plant functional traits could not be viewed and analyzed in isolation, and the trade-off and combination of plant functional traits determines the plant’s life history strategy (Siefert et al., 2015; Funk et al., 2017; Wang et al., 2019; Aiba et al., 2020). The present study found that LA was significantly positively correlated with LV and SLA but significantly negatively correlated with LDMC and LD. Many previous studies have confirmed the co-variation relationship between leaf size and other plant functional traits, and that the formation of leaf size was related to leaf energy balance (Wright et al., 2002; Craine and Lee, 2003; Read et al., 2014; Wang et al., 2019). The possible reason is that in high-altitude and low-temperature habitats, the time for leaves to receive light energy and assimilate carbon dioxide is shortened. Plants may increase the nitrogen content of leaves to stabilize carbon capture capacity, while the cost of respiration and transpiration of leaflets is lower; thus, a low metabolic rate could be better maintained (Zhang et al., 2012). To a certain extent, LDMC represents the plant’s ability to use a certain environmental resource and indicates how much leaves are built (Read et al., 2014). Species with high construction cost per unit LA usually have thicker leaves and higher leaf tissue density (Li and Bao, 2014), especially in high-altitude and low-temperature environments, due to the increase in mesophyll cells and the thickening of cell walls, especially epidermal cell walls and thick and small leaf formats that reduce the internal water loss of plants. Thus, SLA decreases, and the time required for the leaves to repay the initial construction cost through the carbohydrates formed by their own photosynthesis production is longer (Hikosaka et al., 2002).

In the present study, the supported LA per twig cross-sectional sapwood area (HV) was positively correlated with LSE (supported LA per branch length). The results showed that with the increase in the length of the current-year twig, the cross-section diameter of the twig and the number of supporting leaves and LA on the twig increased. Increasing studies have shown that the trade-off between leaf size and leaf emergence is a trade-off shared by woody plants, herbaceous plants, and even a single angiosperm (Wright et al., 2002). Studies have shown that under low temperature, low nutrient soil, and high-altitude stress environment, the biomass allocated to stems at the level of twigs is usually higher, and the leaves have higher LA per unit mass (Wright et al., 2002; Sun et al., 2006). The present study also found a very significant positive correlation between TL of and TD of the current-year twigs, and this finding was related to the trade-off between the longitudinal growth and the radial growth of the plant. The longitudinal growth of the shoots is related to the efficiency of light capture and the spread of seeds; meanwhile, the shoots need to grow radially while growing longitudinally to increase their mechanical support and increase the transportation of water and nutrients (Xiang et al., 2008; Sun et al., 2019). Most studies have shown that current-year twigs, as a metabolically active component, has much greater longitudinal growth rate than radial growth (Xiang et al., 2008; Pérez-Harguindeguy et al., 2013; Zhu et al., 2019). The present study found no significant correlation between TD and other functional traits of the current-year twigs. However, previous studies have shown that low wood density is usually related to the rapid growth of stem diameter and volume, and species with low wood density usually have high LA per branch mass (Wright et al., 2006). In the research on wood density and environmental factors, wood density is strongly related to leaf water potential. Species with high xylem density are usually easier to survive in drier soil and have a lower risk of duct occlusion (Hacke et al., 2001). The twig tissue density of the current-year twigs of Quercus aquifolioides decreased with altitude. In addition, the twigs showed decreased twig tissue density in harsher external habitat, indicating the plant’s increased leaf water potential and deepened roots that could resist high altitude, high winds, and low temperatures in high-altitude areas.

The present study found SL (slope gradient) and TN contributed the most to leaf functional traits and had significant correlation (p < 0.05), which was consistent with Wang’s study on leaf functional traits of apricot (Prunus armeniaca L.) (Wang et al., 2019), they found slope exerted significant effects on SLA (F = 5.578, p = 0.004), LDMC (F = 7.017, p = 0.001), LWC (F = 9.144, p = 0.000), LA (F = 4.797, p = 0.008). As important topographic factors, slope aspect and slope affect plant growth and leaf functional traits by regulating the combination of water and heat. However, we did not find evidence of slope aspect was the main topographic determinant of leaf functional traits. A study in a tropical dry forest landscape in the Balsas river basin in S-central Mexico showed evidence of filtration processes of plant functional strategies to deal with drought and heat stress, operating in the dry forest mostly along topographic positions but not much along the gradient generated by slope aspect (Méndez-Toribio et al., 2017). A number of studies has showed soil nitrogen content was response for the changes in plant functional traits and functional diversity (Jia et al., 2011; Roa-Fuentes et al., 2015; Shovon et al., 2019). In nitrogen-poor sites, changes in functional diversity were mainly attributable to the presence of fast-growing species in sites with higher soil moisture (Poorter and Evans, 1998) and in dry but nitrogen-rich sites, the NUE (nitrogen use effectiveness) of fast-growing species is compromised by moisture stress (Gong et al., 2011). Meanwhile, the present study also found AL was the main determinant of functional traits in current-year twigs of Quercus aquifolioides in southeast Tibet. Previous studies had showed a lot of evidence of altitude is the main topographical factor affecting the functional traits of dominant species (He et al., 2020; Zhang et al., 2021), specially in the Himalayas, stem and branch traits that integrate multiple plant organs and functions are better predictors of species’ altitudinal distributions than leaf traits along the extreme environmental gradient (Maharjan et al., 2021).