Shao-an Pan, Guangyou Hao, Xuhua Li, Qiuhong Feng, Xingliang Liu,Osbert J. Sun,＊

a School of Ecology and Nature Conservation, Beijing Forestry University, Beijing, 100083, China

b Institute of Forestry and Climate Change Research, Beijing Forestry University, Beijing, 100083, China

c Institute of Applied Ecology, Chinese Academy of Sciences, Shenyang, 110016, China

d Sichuan Provincial Key Laboratory of Ecological Restoration and Conservation for Forest and Wetland,Sichuan Academy of Forestry Sciences,Chengdu,610081,China

e Sichuan Wolong Forest Ecosystem Research Station, Wenchuan, 623006, China

Keywords:Climate change Eco-safety Environmental vulnerability Faxon fir Hydraulic risk Habitat constraints Water stress Subalpine

ABSTRACT Global climate change has been seen to result in marked impacts on forest ecosystems such as accelerated tree mortality worldwide due to incidental hydraulic failure caused by intensified and more frequent occurrence of extreme drought and heat-waves. However, it is well understood how the tree hydrological strategies would adjust to environmental variability brough about by climate changes. Here we investigated the hydraulic adjustment as a mechanism of acclimation to different climate conditions along an altitudinal gradient in Faxon fir(Abies fargesii var.faxoniana)–a tree species that plays a key role in conservation of wildlife and maintenance of ecosystem services in subalpine forests. The hydraulic traits and selective morphological and physiological variables were measured seasonally along an altitudinal gradient from 2,800 to 3,600 m a.s.l. We found that the native percentage loss of conductivity (PLC) increased with altitude across the seasonal measurements. Both the native sapwood-specific hydraulic conductivity (Ks) and native leaf-specific hydraulic conductivity (Kl) significantly decreased with altitude for measurements in July and October,coinciding with the timing for peak growth and pre-dormancy, respectively. The morphological traits varied toward more conservative tree hydrological strategies with increases in altitude, exhibiting trade-offs with hydraulic traits. The total non-structural carbohydrates in both needle (NSCNeedle) and branch (NSCBranch) as well as photosynthetic capacity of current-year leaves played variable roles in maintaining the integrity of the hydraulic functioning and shaping the hydraulic adjustment under prevailing environmental conditions. Our findings indicate that Faxon fir possesses some degree of hydraulic adaptability to water limitation imposed by climate fluctuations in subalpine region through morphological and physiological modifications.

1. Introduction

The global patterns of vegetation are co-determined by geographical zonation of climate and hydrothermal adaptability of terrestrial plants(Kueppers et al., 2017; Berdugo et al., 2020). The plasticity in hydrological strategies is a critical determinant of the spectrum of hydrothermal conditions that a plant species can functionally cope with (Choat et al., 2012; He et al., 2019; Jiao et al., 2021); these strategies are developed over a long evolutionary history under prevailing climatic conditions(He et al.,2019).Climate change can lead to the risk of range shift or habitat constraints to plant species with poor hydrothermal adaptability(Chen et al.,2011;Elsen and Tingley,2015;Berdugo et al.,2020; Xu et al., 2020; Radcliffe et al., 2021), especially trees in wet forests that are not normally considered to expose to drought risks(Choat et al., 2012). There are increasing concerns on the occurrence of drought-induced xylem embolism as a climate change risk to terrestrial woody plants (Klein et al., 2018; Cardoso et al., 2020a, 2020b; Arend et al., 2021; Levionnois et al., 2021), but few studies have investigated the hydraulic dysfunction of trees associated with low temperature stresses in alpine environment. Freeze and frost may result in the occurrence of plant physiological water stress that induces embolism(Mayr et al., 2003; Willson and Jackson, 2006; Charrier et al., 2021).Trees grown at a high altitude site encounter difficulties in developing an efficient water transport system due to constraints on longitudinal growth by low temperatures (Petit et al., 2011). Charrier et al. (2014)found that subzero temperatures play an important role in the formation of freeze-thaw induced embolism in their studies of temperature effects on hydraulic conductivity in ten woody angiosperms. It has also been reported that the hydraulic architecture varies seasonally across different types of woody plants(Dai et al.,2020).

Fig. 1. Schematic illustration of hydraulic strategies of trees in response to temperature changes.

Embolism typically occurs due to the formation of cavitation in xylem vessel in angiosperms or blockage of tracheid in gymnosperms (Tyree and Spery,1989).The tracheid blockage in gymnosperms occur either by resin filling as a result of epithelial cell rupture under severe water stress,which clogs the conduits and prevents water flow(Partelli-Feltrin et al.,2021),or because of embolism induced by freeze-thaw events(Camarero et al.,2016;Jankowski et al.,2017;Zhang et al.,2019).The diameters of tracheid (or vessel) are commonly found to be related to freeze-thaw induced embolism resistance in trees, with greater the diameter less the embolism resistance (Davis et al., 1999; Mayr et al., 2002, 2003;Jankowski et al., 2017). In some cases, e.g. with co-occurrence of low temperature and water stress, a single freeze-thaw cycle may lead to embolism even in conifers with relatively small tracheid diameters(Willson and Jackson, 2006). Embolism refilling is one of the most important adaptive strategies (or repair method) in angiosperms for maintenance of hydraulic integrity, especially in winter (Niu et al.,2017).In gymnosperms,embolism refilling in tree branches occurs either through water absorption by needles or under negative pressure(McCulloh et al.,2011). Gymnosperms in cooler regions usually rely on small conduits to avoid freezing-induced cavitation (Hacke and Sperry,2001); whilst those under temperate climate are capable of generating positive xylem pressure in late winter or early spring to facilitate embolism refilling (Niu et al., 2017; Wang et al., 2018). The processes of positive xylem pressure generation in these tree species are often associated with relatively high sugar concentrations in the xylem sap during late winter or early spring, indicating the importance of non-structural carbohydrate (NSC) storage and remobilization for positive pressure generation through osmosis (Cirelli et al., 2008; Wang et al., 2018; see Fig. 1 for a schematic illustration of the processes). Findings from experimental studies suggest that NSC in the parenchyma of wood tissues provide the osmotic forces driving the refilling of embolized conduits(Trifilo et al.,2019).Long-term studies and micro-CT observations show that Norway spruce (Picea abies [L.] H. Karst.) grown at the upper timberlines can survive annual cycles of apparent winter embolism by mechanism of refilling (Mayr et al., 2020). The recovery of xylem transport in Norway spruce has been found to involve stored NSC as a strategy of tolerating low temperature-induced water supply limitations(Tomasella et al.,2017).

Faxon fir(Abies fargesii var.faxoniana[Rehder&E.H.Wilson]Tang S.Liu)is an ancient gymnosperm tree species originating in the Cretaceous(Florin, 1963). It naturally occurs in the subalpine region of Southwest China,between 27°and 37°N in latitude and 100°–105°E in longitude,and from 2,700 to 3,900 m above sea level (a.s.l.) in altitudinal range(Chen et al.,2021a).At altitudes above 3,000 m a.s.l.,the species forms pure forest stands and is the main tree species of the timberline in the upper reaches of the Minjiang catchment(Cheng et al.,2005).Given that Faxon fir-dominated forests support the wild populations of Chinese Giant Pandas by providing primary habitats and facilitating the growth of their dietary bamboo species (Fargesia denudate Yi), maintenance of the health and population stability of Faxon fir is vital for the regional wildlife conservation and provisioning of critical ecosystem services.

Many studies have demonstrated that global climate change imposes strong influences on natural distribution of plants,with likely occurrence of poleward(higher latitude)and upward(higher altitude)migrations of some tree species (Lenoir et al., 2008; Chen et al., 2011; Klinge et al.,2021;Beloiu et al.,2022).In mountainous regions,an upward shift in the distributional range of montane plants may occur due to thermal controls of the upper range limits (Pauli et al., 2012; Devi et al., 2020). Such upward migration from the downslope tree species may lead to further constrained habitats for tree species typically distributed in the upper slopes and toward the tree line(Elsen and Tingley,2015).Moreover,an intensified climate change in recent decades has been seen to result in increased occurrence and magnitude of extreme climatic events such as heat-waves and severe drought,leading to risks of tree mortality due to hydraulic failure caused by xylem embolism in alpine and subalpine regions (Nardini et al., 2014). For example, increased frequency and intensity of extreme climatic events have been observed to cause significant decreases in hydraulic conductivity in both a broad-leaved tree species

Schima superba and a coniferous tree species Cunninghamia lanceolata in the subtropical region of China(Qu et al.,2020).Given that Faxon fir has survived the glacial and interglacial periods during its evolutionary history (Florin, 1963), it is postulated that this species possesses some inherent ecophysiological characteristics to cope with the fluctuating environmental conditions on high altitudinal sites. While recent studies have revealed that Faxon fir can rely on adjustment of ectomycorrhizal strategies to adapt to varying environmental conditions(e.g.Chen et al.,2021a,2021b),intriguing questions remain that whether the species also displays hydrothermal plasticity or resilience to endure water supply limitations in subalpine environment,due to climatic drought or low soil temperature constraints at the timberline.

In this study, we investigated the hydraulic traits and related morphological and physiological characteristics in mature Faxon trees along an altitudinal gradient, and tested the occurrence and extent of hydraulic plasticity as a mechanism of adapting to a warmer and drier future climate in the region. We specifically address two questions: (1)how would the hydraulic traits in Faxon fir vary with changes in altitude,and (2) what would be the underlying controls on the elevational variations in the hydraulic traits in Faxon fir?We hypothesized that Faxon fir trees respond to increases in altitude by reducing tracheid diameters and increasing tissue NSC, and that the occurrence of such hydraulic plasticity would allow the species to cope with potential water limitations as a result of hotter and drier future climate predicted to occur in the region.

2. Materials and methods

2.1. Study site and experimental setup

Mature Faxon fir trees were sampled for measurements of hydraulic traits and related morphological and physiological variables along an altitudinal gradient from 2,800 to 3,600 m a.s.l.,at intervals of 200 m in elevation (i.e. at 2,800, 3,000, 3,200, 3,400, and 3,600 m a.s.l.,respectively).At each altitudinal site,we systematically selected up to 10 trees that reached the canopy for sample collections and measurements.Two healthy branches (approximately 1.5 m in length) from the upper canopy of each designated sampling tree were collected at pre-dawn(6:00 a.m.–7:00 a.m., Beijing Time) in July and October 2019, and January and April 2020, respectively. On each occasion, the branches were cut off from the sampling trees using an extended branch scissors and immediately placed in black plastic bags containing moist paper towels, after removing approximately 5-cm segment and wrapping the cut surface with a plastic bag filled with moist paper towels.They were then transported to laboratory for subsequent processing and measurements of leaf water potential, stem hydraulic conductivity, native embolism, tracheid diameter, and tissue NSC concentrations as described below.

Table 1 Changes in annual mean temperature(MAT),mean monthly soil temperature at 10-cm depth (TSoil) during measurement times and average diameter at breast height (DBH) of sampling trees along an altitudinal gradient in the Wolong Nature Reserves, Southwest China.

2.2. Extrapolation of climate data and soil temperature measurements

For lack of on-site meteorological data along the altitudinal gradient,we derived the values of MAT and mean monthly air temperature(MMT)for the five altitudinal sampling locations from a 0.25°×0.25°grid-cell dataset for the 20-year period 1997–2016. The dataset was originally developed by anomaly approach, with observations at more than 2,400 metrological stations across China (Xu et al., 2009; Chen et al., 2021a).Topographic corrections were made for specific locations using the lapse rate of air temperature at 0.65°C per 100 m rise in elevation(Zhao et al.,2008). Soil temperatures at 10-cm depth (TSoil) were measured at each location with two soil thermometers (iButton, DS1922L) during April 2017–April 2020(Table 1).

2.3. Leaf water potential and stem hydraulic conductivity

Pre-dawn leaf water potential(Ψ,MPa)was measured with a pressure chamber(SKPM1400-50,SKYE,UK),on current year twigs immediately upon transportation to laboratory.The measurements were conducted on two twigs from each of the detached branches as described above. The native hydraulic conductivity(Kh)was measured on 5-year-old branches from all sampling trees following the procedure of Wang et al.(2018).Kh(kg?m?s-1?MPa-1) was calculated as (Tyree and Dixon, 1986; Becker et al.,1999):

Kh= Jv/(ΔP/ΔL)

where Jv is the water flow rate through the stem segment (kg?s-1), and the ΔP/ΔL the pressure gradient across the stem segment(MPa?m-1).We also calculated the native sapwood-specific hydraulic conductivity (Ks,kg?m-1?s-1?MPa-1)by dividing Khwith the sapwood area(SA),and the native leaf-specific hydraulic conductivity (Kl, kg?m-1?s-1?MPa-1) by dividing Khwith total needle area distal to the branch segments.Fuchsindye was perfused into the branch segment under a hydraulic height of 50 cm in order to keep branch stem in their natural embolism state. The sapwood area was determined as the average value of the two ends of the segment.Images of the transverse cross-sections of the stained stem were obtained using a photo scanner,with the area of stained portion analyzed using ImageJ software (US National Institutes of Health, Bethesda, MD,USA). More than 100 fully expanded needles were collected from each branch sample to determine the projected leaf area by scanning and image analysis.All needles from each of the branch segments were then oven-dried to constant weight at 65°C for 48 h to measure leaf dry mass(LM).Leaf mass area(LMA)and the total area(LA)of the distal needles were then derived from the above measurements,and the LA to SA ratio(LA/SA)were calculated.

Table 2 Results of the two-way ANOVA showing the effects of season, altitude and their interaction.

2.4. Native embolism and tracheid lumen diameter (TLD)

The native xylem embolism was determined as an increase in hydraulic conductivity after embolism removal using a partial vacuum method(McCulloh et al.,2011). Briefly,following the measurements of Kh, the stem segments were submerged in perfusion solution (degassed 20 mmol?L-1KCl) under a partial vacuum overnight (approximately 12 h)to refill embolized tracheid.The native percentage loss of conductivity(PLC)was calculated as(Tyree and Dixon,1986):

The stem segments from six out of ten sampling trees at each altitudinal site location, which were previously used for hydraulic conductivity measurements, were examined for the xylem anatomy (Li et al.,2021).Cross-sections of～20 μm in thickness were processed by a sliding microtome (Model 2010–17, Shanghai Medical Instrument Corp.,Shanghai, China) and stained by 0.1% toluidine blue. They were examined by a light microscope under a magnification of 400×(Leica ICC50,Wetzlar,Germany)and photographed using an inbuilt digital camera for the growth ring of 2018 opposite the compression wood. The images were examined, and the diameters of individual tracheid lumens measured, using the Image J software. Three images for the outer two growth rings from each cross-section sample were analyzed, with measurements for a total of 2,500 tracheid for each location.

2.5. Measurements of non-structural carbohydrate (NSC)

NSC was measured in roots (NSCRoot), sapwood of trunk (NSCTrunk),branch (NSCBranch) and needles (NSCNeedle) on six out of ten sampling trees at each altitudinal site location.Small roots of the diameter 5–8 mm were collected from the top 20-cm soil layer. Increment cores of the sapwood were extracted at the breast height (1.37 m), on two opposite sides of the trunk parallel with the contour with a 5-mm inner-diameter borer. Branch samples were similar to those used for measurements of hydraulic conductivity, and needle samples were from the same branches.All samples for NSC measurements were terminated of growth activity within 2 h of collection by placing in a microwave oven at 600 W for 40 s and then dried at 70°C for more than 48 h to constant weight.Soluble sugar concentrations were measured with the anthrone method on oven-dry samples (Seifter et al., 1950; Wang et al., 2018), and the starch contents were determined following the procedure of Wang et al.(2018).

2.6. Measurements of leaf photosynthetic capacity

Photosynthetic gas exchange was measured on a clear day between 9:30 a.m. -11:30 a.m. in August of 2019, with a portable infrared gas exchange system (Li-6400, Li-Cor Inc., Lincoln, NE, USA). The measurements were made on detached branches from the upper-mid crown of three pre-designated sampling trees at each altitudinal site location.The branch samples were placed into a bucket,with the cuts submerged in water immediately upon detachment from sampling trees, and subjected to a 10-min induction at a saturating photosynthetically active photon flux density (PPFD) of 1,200 μmol?m-2?s-1and controlled leaf temperature at 25°C.CO2concentrations in the cuvette were set for 400 ppm using an injector system(Li-6400-01,Li-COR Inc.)with a CO2mixer and compressed CO2cartridges.The photosynthetically active radiation(PAR)was provided by a red/blue LED light source(Li-6400-02B,Li-Cor Inc.).After induction,measurements were made under conditions of with set for 1,000 μmol?m-2?s-1PPFD and 400 μmol?mol-1CO2concentrations, each repeating three times. The variables from the gas exchange measurements included net assimilation rate (A), stomatal conductance(gs),intercellular CO2concentration(Ci),and transpiration rate(Tr).We also computed the instantaneous light use efficiency(LUE=A/PAR)and water use efficiency(WUE=A/Tr) following Zhang et al.(2018).

2.7. Data analysis

Liner regression analysis was used to determine the relationships among various functional traits and to test changes in the functional traits with altitude. A two-way analysis of variance (ANOVA) was used to determine the effects of altitude, measurement time and their interactions using SPSS 18.0 (SPSS, Inc., Chicago, IL, USA). One-way ANOVA was performed to test the differences in the frequency of TLD distribution among locations. Multi-mean comparisons were made with Tukey's honest significant difference(HSD)test at P ＜0.05.

Fig.2. Seasonal variation in the altitudinal changes of(a)percentage loss of conductivity(PLC),(b)sapwood-specific hydraulic conductivity(Ks),and(c)leaf-specific hydraulic conductivity (Kl), in mature Faxon fir (Abies fargesii var. faxoniana).

Redundancy analysis (RDA) was used to determine the interrelationships of hydraulic traits(PLC,Ksand Kl)with all the climatic,morphological and physiological variables. A Monte Carlo permutation test(999 permutations)was used to examine the significance level of the RDA results. Hellinger or standardized transformation was used to normalize the data used in RDA.The significant factors identified by RDA were used to establish structural equation models(SEM)in the effects on the hydraulic trait variables PLC, Kland Ks, respectively. We also partitioned the variations of hydraulic traits(PLC,Ks,and Kl) attributable to climate(MAT,MMT and TSoil),tree organ NSC concentrations(NSCNeedle,NSCBranch,NSCTrunck,and NSCRoot),morphology(CA,LA,SA,LA/SA,and LMA),and physiology(Ψ,A,gs,Ci,Tr,LUE and WUE).All these analyses were performed with the R package(version 3.5.1)of RStudio.

3. Results

3.1. Variations of hydraulic traits along altitudinal gradient

PLC,Ks,and Klwere all significantly(P ＜0.05)or highly significantly(P ＜0.001) affected by altitude, season and an interaction between altitude and season (Table 2). PLC were significantly and posiviely correlated with altitude in the measurements of July and January 2020,and marginally in October 2019(Fig.2a).The differences in PLC among measurement times were greater at the two lower altitudes of 2,800 and 3,000 m a.s.l.,respectively,and then significantly reduced at the higher altitudinal sites(Figs. 2 and S1).

Fig. 3. Altitudinal changes of (a) total leaf area (LA), (b) sapwood area (SA), (c) cross-section area of branch (CA), (d) LA/SA, (e) leaf mass area (LMA), and (f)tracheid lumen diameter (TLD), in mature Faxon fir (Abies fargesii var. faxoniana).

Both Ksand Klwere significantly and negatively correlated with altitude (Fig. 2b and c), and greatly varied with measurement time(Fig. S1). There were significant and negative linear relationships between Ksand altitude in the measurements of July and October 2019(Fig. 2b1 & b2); whereas Ksdid not vary with altitude in the measurements of January and April 2020 (Fig. 2b3 and b4). Seasonally, Kswas significantly greater in July 2019 than at other measurement times at all altitudes except 3,600 m a.s.l.,with Ksbeing greater in July 2019 than in October 2019 (Fig. S1). Klhad significantly negative correlations with altitude in July and October 2019 and a marginally negative correlation in January 2020 (Fig. 2c1, c2 and c3), displaying similar seasonal variations as Ksat 2,800,3,400 and 3,600 m a.s.l.At 3,000 and 3,200 m a.s.l.,Kldid not vary significantly with measurement time (Fig.S1).

3.2. Morphological and physiological variations along altitudinal gradient

All morphological traits except TLD showed marked variations with seasonal measurements (Table 2). With exception of LMA, all other morphological traits examined, including LA, SA, CA, LA/SA and TLD,had significant and negative linear relationships with altitude regardless of the season (Fig. 3). LMA was significantly and positively related to altitude(Fig.3e).The frequency distribution of TLD was significantly(P＜0.01)affected by altitude,with a general trend of reduced TLD when moving higher in altitude and an anomaly at 3,400 m a.s.l.(Fig.S2).

NSC concentrations in both branch and roots also varied markedly with the seasonal measurements. The needle NSC concentration was affected both altitude and season, but not their interaction. NSC in the sapwood of trunk varied significantly with season and was affected by an interaction between altitude and season (Table 2). In branch, NSC concentration linearly increased with altitude(Fig.4a);whilst in needles and roots, NSC concentration varied with altitude in significant quadratic relationships (Fig. 4b and c). Seasonally, the NSC concentration in needles and branch were significantly(P ＜0.05)lower in July 2019 than at other measurement times except in needles at 3,400 and 3,600 m a.s.l.,while a clear pattern of changes in NSC concentration with measurement time were not found in roots and trunk (Fig. S3). In roots, the NSC concentrations were mostly lower in the two measurements in 2019 than in 2020 at 3,000,3,400,and 3,600 m a.s.l.(Fig.S3).Among the organs,NSC concentrations varied in the order of: needle ＞root ＞branch ＞trunk(Fig.S3).

The physiological variables of A,gs,Ci,Tr,LUE and WUE all tended to decrease with altitude(Fig.5).Water potential(Ψ)had maximum values(less negative) in measurements of July 2019 along the altitudinal gradient, and minimum values (more negative) close to -1.6 MPa,mostly in January and April 2020(Fig.S4).

3.3. Controlling factors of hydraulic trait variations

The axes 1 and 2 of RDA explained 70.7%and 19.6%of variations in the overall hydraulic traits along the altitudinal gradient, respectively(Fig.6).Dominant drivers of the hydraulic trait variables(PLC,Ksand Kl)were identified as the climatic factors of MAT(F=17.6,P ＜0.001)and MMT(F=48.9,P ＜0.001),the morphological traits of LA(F=3.3,P=0.044),SA(F=75.6,P ＜0.001),and LA/SA(F=48.4,P ＜0.001),the physiological variables of A (F = 6, P = 0.003), and LUE (F = 4, P =0.021), and NSC concentrations in needles (F = 19.6, P ＜0.001) and branch(F=7.1,P=0.004).Both Ksand Klwere significantly(P ＜0.05)and positively associated with MMT,MAT,SA,LA,and A,and negatively with NSC concentration in needles (NSCNeedle) and branch (NSCBranch).There was also a negative association between Kland LA/SA. PLC was found to be significantly (P ＜0.05) and positively associated with NSCNeedle,NSCBranch,and LUE,and negatively with MMT,MAT,SA,LA,A,and LA/SA.

The SEMs on the regulatory pathway of hydraulic traits revealed that MAT had a direct and highly significant negative control on PLC(Fig.7a),and direct and highly significantly positive controls on Ks(Fig.7b)and Kl(Fig.7c).In addition,there were indirect influences of MAT on PLC via its significant regulations of A, NSCNeedleand LUE (Fig. 7a), on Ksvia significant regulations of NSCNeedleand NSCBranch(Fig. 7b), and on Klvia significantly regulations of NSCNeedle, NSCBranch, and LA/SA (Fig. 7c).MMT only had indirect controls on all the three hydraulic trait variables,specifically on PLC via the pathways of A and NSCNeedle, on Ksvia the pathways of NSCNeedleand NSCBranch, and on Klvia the pathways of NSCNeedle, NSCBranch, and LA/SA (Fig. 7). Regression analyses further demonstrated significant negative relationships of the climatic factors MAT and MMT with PLC, and significant positive relationships with Ksand Kl(Fig. S5). Significant regressional relationships were also found between the three hydraulic trait variables and various morphological,physiological and biochemical(i.e.NSC)variables(Figs.S6 and S7).

Fig.5. Altitudinal variations in(a)maximum net photosynthetic rate(A),(b)stomatal conductance(gs);(c)intercellular CO2 concentration(Ci),(d)transpiration rate(Tr),(e)light use efficiency(LUE),and(f)water use efficiency(WUE)of mature Faxon fir(Abies fargesii var.faxoniana).Vertical bars illustrate±1 SE;different letters designate significant differences among altitudes at P ＜0.05 based on LSD test.

Variance partitioning shows that PLC was predominantly influenced by an individual effect of physiology with an explainable variance of 12.7%, followed by a joint effect between climate and organ NSC concentrations (9.5%), a four-way interactive effect among climate,morphology, physiology and organ NSC concentrations (8.7%), and an individual effect of organ NSC concentrations (6.5%; Fig. 8a). Kswas predominantly influenced by an individual effect of morphology with an explainable variance of 19.3%,as well as an three-way interactive effect among climate,morphology and organ NSC concentrations(8.3%)and a four-way interactive effect among climate,morphology,physiology and organ NSC concentrations(6.4%;Fig.8b).Morphology occurred to be the single most influential control of Klwith an explainable variance of 44.1%,with some minor influences from a three-way interaction among morphology, physiology and organ NSC concentrations (5.3%) and a three-way interaction among climate, morphology and physiology(4.2%; Fig.8c).

4. Discussion

Altitudinal changes in temperature in high mountains provide a practical solution to test the climate change responses and adaptability of plants (Alexander et al., 2015). In support to our hypothesis, Faxon fir responded to temperature decline along the altitudinal gradient with increased organ-specific NSC concentrations and reduced xylem tracheid diameter for avoiding hydraulic failure due to climate variability in subalpine environment.

4.1. Variations of hydraulic traits and related morphological variables along altitudinal gradient

In this study,the hydraulic traits in Faxon fir varied with site locations consistent with altitudinal changes in temperature.A significant increase in percentage loss of conductivity with altitude during the growing season suggests that water movement in trees grown at higher altitudes are constrained, possibly due to more severe winter embolism at high altitudes. Previous studies by others have demonstrated that low temperature may disrupt hydraulic conductivity through persistent embolism in conduits by freeze-thaw events and tracheid collapse induced by water-stress (Jankowski et al., 2017; Niu et al., 2017), resulting in low xylem water content that increases the risk of winter embolism formation(Lintunen et al., 2018). Our results are similar to the finding that coniferous trees at a high altitude site suffer heavy loss of conductivity in the winter(Mayr et al.,2014).In this study,PLC in Faxon fir trees at higher altitudinal site locations displayed relatively less seasonal fluctuations,consistent with the morphological traits of lower cross-section xylem area of branches and smaller tracheid lumen diameter(Figs.3 and S2).At a high altitude, the overwinter organs of trees could suffer from embolism due to frequently freeze-thaw cycles and frost drought events(Wang et al.,2018).In the present study,the xylem tracheid diameter(TLD)was found to generally decrease with altitude, albeit an anomaly at some distance below the timberline at 3,400 m a.s.l. (Figs. 3 and S2). It is generally accepted that trees usually maintain hydraulic integrity through phenotypic adjustments of the xylem structure when experiencing disadvantaged environmental conditions (He et al., 2019,Table 2).But this adjustment occurs at the expense of declining hydraulic efficiency,reflecting a trade-off between hydraulic safety and efficiency(Zhang et al.,2017; Wang et al.,2018).

Two measures of hydraulic conductivity, Ksand Kl, were found to have a maximum value during the active growth season in July and a minimum value during winter dormancy in January 2020, indicating that Faxon fir is capable of adjusting its hydraulic seasonally to maintain the balance between water supply and demand. An extensive loss of hydraulic conductivity in winter but almost complete recovery before leaf bud break in spring have been found to occur in other conifer species growing under harsh environmental conditions(Mayr et al.,2014;Wang et al.,2018).Our results differed with the findings of Wang et al.(2018)that the hydraulic conductivity was lower in summer and higher in early spring, due to lower water transport efficiency of newly formed xylem conduits in current year growth ring (Jacobsen et al.,2018).

Fig. 6. Redundancy analysis (RDA) ordination biplot of climate factors and selective traits of hydraulics, physiology and morphology. The percentages on the biplot respectively represent the variations of hydraulic traits explainable by RDA axes 1 and 2.The red arrows represent the hydraulic traits;the blue arrows represent the climate factors as well as selective physiological and morphological variables. PLC, percentage loss of conductivity; Ks, sapwood-specific hydraulic conductivity; Kl, leaf-specific hydraulic conductivity; LMA, leaf mass area; LA, total leaf area; SA, sapwood area; LA/SA, the ratio of LA to SA;NSCBranch,total non-structural carbohydrate concentration in branch;NSCNeedle,total non-structural carbohydrate concentration in needle; NSCTrunk, total nonstructural carbohydrate concentration in trunk; NSCRoot, total non-structural carbohydrate concentration in roots; MAT, mean annual air temperature;MMT, mean monthly air temperature; gs, stomatal conductance; LUE, light use efficiency; A, maximum net photosynthetic rate; Ψ, pre-dawn leaf water potential. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

The sapwood area(SA)in Faxon fir was found to significantly decline with altitude(Fig.3b).Trees in low temperature environment can benefit from small conduits to avoid embolism due to more frequent and severer freeze-thaw cycles (Mayr et al., 2002; Pittermann and Sperry, 2003),with smaller diameter of conduits usually associated with lower specific hydraulic conductivity(He et al.,2019).In thermally and hydrologically disadvantaged habitats,increased susceptibility to freezing injury or high carbon expenditure would lead to natural selection favoring the evolution of low hydraulic conductivity (He et al., 2019). Moreover,trees on more exposed sites would evolve with higher mechanical reinforcement for withstanding strong winds and greater snow loads,hence the need for more carbon investment in mechanical architecture (Taneda et al.,2019).This is supported by our findings of significantly decreased LA/SA but increased LMA with altitude (Fig. 3). Plants tend to allocate less carbon to aboveground tissues under conditions of thermal stress (Jin et al.,2019).Our findings together with those in literature conform to the freezing-tolerance hypothesis that the freezing temperature is an important climatic factor in determining the species of woody plants distribution (Wang et al., 2011; He et al., 2019). Faxon fir appears to adopt a“conservative”hydraulic strategy in response to low temperature environment at high altitude,to maintain an effective water transport.

4.2. Variations of non-structural carbohydrate (NSC) and controls on hydraulic conductivity

Fig.7. Structural equation models(SEMs)of dominant drivers of the variations in (a) percentage loss of conductivity (PLC), (b) sapwood-specific hydraulic conductivity (Ks), and (c) leaf-specific hydraulic conductivity (Kl), in mature Faxon fir (Abies fargesii var. faxoniana) along an altitudinal gradient. Standardized path coefficients are displayed on line arrows, with solid lines indicating positive effects, dash lines indicating negative effects, and thickness of the line designating the relative significance level of the effetcs. R2 value represents the proportion of total variance explainable for the specific dependent variable. ＊:significant at P ＜0.05; ＊＊: significant at P ＜0.01; “＊＊＊”: significant at P ＜0.001. MAT, mean annual air temperature; MMT, mean monthly air temperature; LUE, light use efficiency; NSCNeedle, total non-structural carbohydrate concentration in needle; NSCBranch, total non-structural carbohydrate concentration in branch; LA, total leaf area; SA, Sapwood area.

Funding

This study was supported by the National Key Research and Development Program of the Ministry of Science and Technology of China(Grant No.2016YFC0502104).

Data availability

The data set generated for the study area is available from the corresponding author on reasonable request.

Declaration of competing interest

The authors declare that they have no competing interests.

Acknowledgements

The authors are grateful to Dr. Yujie Wang for constructive suggestions on data analysis, and Yuxue Zhang, Chuanjing Wu, and Weixiang Cai for their technical assistance in sample processing.We would also like to thank Dr. Ruihua Pan for instrumental support and Shuyue Wang for technical assistance in laboratory and field work.Constructive comments from three anonymous reviewers helped improve the writing of the manuscript.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://do i.org/10.1016/j.fecs.2022.100040.

Abbreviations

PLC Percentage loss of hydraulic conductivity

KhNative hydraulic conductivity

KsSapwood-specific hydraulic conductivity

KlLeaf-specific hydraulic conductivity

LA Total leaf area of measurement branch

LMA Leaf mass area

CA Cross-section area of branch

SA Sapwood area of branch

LA/SA Ratio of total leaf area to sapwood area of measurement branch

TLD Tracheid lumen diameter

NSC Non-structural carbohydrate

A Maximum net photosynthetic rate

gsStomatal conductance

CiIntercellular CO2concentration

TrTranspiration rate

LUE Light use efficiency

WUE Water use efficiency

Ψ Pre-dawn leaf water potential

MAT Mean annual air temperature

MMT Mean monthly air temperature

MAP Mean annual precipitation

TSoilSoil temperature at 10-cm depth

RDA Redundancy analysis

SEM Structural equation model

ANOVA Analysis of variance

LSD Least significant difference

HSD Tukey's honest significant difference

GCMs General Circulation Models