**Leaf Age Compared to Tree Age Plays a Dominant Role in Leaf** δ**13C and** δ**15N of Qinghai Spruce (***Picea crassifolia* **Kom.)**

**Caijuan Li 1,2, Bo Wang 1,2, Tuo Chen 1,\*, Guobao Xu 1, Minghui Wu 1,2, Guoju Wu <sup>1</sup> and Jinxiu Wang 1,2**


Received: 17 January 2019; Accepted: 31 March 2019; Published: 4 April 2019

**Abstract:** Leaf stable isotope compositions (δ13C and δ15N) are influenced by various abiotic and biotic factors. Qinghai spruce (*Picea crassifolia* Kom.) as one of the dominant tree species in Qilian Mountains plays a key role in the ecological stability of arid region in the northwest of China. However, our knowledge of the relative importance of multiple factors on leaf δ13C and δ15N remains incomplete. In this work, we investigated the relationships of δ13C and δ15N to leaf age, tree age and leaf nutrients to examine the patterns and controls of leaf δ13C and δ15N variation of *Picea crassifolia*. Results showed that 13C and 15N of current-year leaves were more enriched than older ones at each tree age level. There was no significant difference in leaf δ13C values among trees of different ages, while juvenile trees (<50 years old) were 15N depleted compared to middle-aged trees (50–100 years old) at each leaf age level except for 1-year-old leaves. Meanwhile, relative importance analysis has demonstrated that leaf age was one of the most important indicators for leaf δ13C and δ15N. Moreover, leaf N concentrations played a dominant role in the variations of δ13C and δ15N. Above all, these results provide valuable information on the eco-physiological responses of *P. crassifolia* in arid and semi-arid regions.

**Keywords:** Leaf δ13C; Leaf δ15N; Growth stage; Environmental factors; Relative importance

#### **1. Introduction**

As one of the most powerful tools for studying plant eco-physiology, stable isotope techniques provide fundamental insights into how plants interact with and respond to biotic or abiotic environmental factors, helping us to better understand the relationship between plants and their environment [1–3]. In particular, leaf carbon isotope composition (δ13C), which reflects the balance between leaf conductance and photosynthetic rate [4], is widely used to analyze intraspecific or interspecific differences in photosynthetic and physiological characteristics [5,6], to measure the long-term water use efficiency under different environmental conditions and to reveal significant functional changes in plant metabolism and adaptation to various environmental stresses [7–9]. In addition, the natural abundance of 15N in leaves or roots has been proposed as an important tracer to reflect the outcome of different processes affecting δ15N compositions, thereby providing an integrative measure of terrestrial N processes [10–12].

However, to our knowledge, multiple previous studies have been conducted on large global or regional-scale variations in plant δ13C or δ15N values, while the understanding of δ13C and δ15N

patterns on intermediate spatial or temporal scales is rather limited [11,12]. More importantly, variations of δ13C and δ15N values during different plant development and growth stages (for example, stand age class, tree age class) have been neglected. Currently, increasing attention has been paid to investigating the significant variations in stable carbon isotopes among different plant organs, such as leaves, stems, shoots, roots, or different plant species including C3 and C4 plants [3,13–15]. However, research about the variations in the natural abundance of 13C and 15N with leaf habit, phenological leaf traits or leaf age class on intermediate scales remains incomplete [10,13]. For example, a previous study demonstrated that leaf age was of special interest when exploring isotope fractionation, because younger leaves show different physiological properties and mechanisms of carbon and nitrogen assimilation compared to older leaves [13,16]. Likewise, significant variations in the δ15N values with stand age were discussed. Li et al. [10] reported that leaf δ15N variations at the community, plant growth form and species levels were significantly reduced with increasing stand age over shorter times and at smaller spatial scales. In addition, Vitoria et al. [17] demonstrated that there was no significant difference in leaf δ13C and δ15N values of evergreen and deciduous species within a site. In addition, despite tremendous progress over the past few decades in investigating what causes variation in plants δ13C and δ15N values, limited studies have allowed a comprehensive explanation for the relative importance of study variables on carbon and nitrogen compositions [16,17]. Therefore, it is necessary to conduct further investigation on the natural abundance of 13C and 15N during different plant growth stages over intermediate spatial or temporal scales.

Moreover, there are a variety of other abiotic and biotic factors that control leaf δ13C and δ15N values during plant development and growth [18,19]. For example, leaf δ13C values changed with leaf habit, morphology, genetics and irradiance [1,16], which may reflect differences in photosynthetic water use efficiency. Current work has demonstrated that plant δ13C is also influenced by various environmental factors such as precipitation, humidity, soil moisture, and air temperature [18,19]. Furthermore, leaf functional elements such as nitrogen (N) and phosphorus (P) also play a key role in δ13C values through their indirect effects on photosynthetic capacity and the synthesis of proteins, DNA and RNA [20,21]. However, our knowledge on the relative role of these parameters in leaf δ13C values remains incomplete. Whether and how these established relationships could hold true in specific biomes remains largely uncertain. This uncertainty is especially true for forests of different study regions that might display contrasting responses to climate [16]. Additionally, compared to plant δ13C, the relationship between leaf δ15N and those mentioned factors has received less attention. Earlier studies have focused on the spatial or seasonal variation in plant δ15N values along a specific factor gradient, but did not consider the relative importance of those variables in the variations of δ15N values [11,12]. Thus, for detailed knowledge, additional empirical studies are required to address the relative effect of biotic or abiotic factors in the variations of δ13C and δ15N values at specific biome levels.

Qinghai spruce (*Picea crassifolia* Kom.) as a common coniferous evergreen species is widely distributed at altitudes ranging from 2300 to 3300 m in the subalpine and alpine environments of Qilian Mountains in the Northern China, in which the availability of water, nutrients and temperature is crucial for determining plant performance, abundance and distribution. It exhibits a wide tolerance to different environmental conditions and has significant ecological function in northwest China. However, limited studies have been conducted on temporal and spatial variations in the stable carbon and nitrogen isotope compositions in different aged leaves and trees of *P. crassifolia.* Therefore, it is necessary to fully understand the effects of the variables mentioned above on the δ13C and δ15N values of *P. crassifolia*.

We hypothesized that the δ13C and δ15N values of *P. crassifolia* would change with leaf age and tree age to adapt to the growth stage needs. In addition, environmental variables could contribute to the growth and eco-physiology of *P. crassifolia*. The main objectives of this study were (i) to quantify the variation in *P. crassifolia* leaf δ13C and δ15N values along the leaf age and tree age gradients and provide evidence of the physiological mechanisms underlying the variations in the δ13C and δ15N

values; (ii) to investigate the relationship between leaf δ13C, leaf δ15N and leaf nutrients; and (iii) to explain the relative role of leaf physiological properties and leaf nutrients in the variations of δ13C and δ15N.

#### **2. Materials and methods**

#### *2.1. Site Description*

The research was conducted in the Shuang Longgou region (longitude 102◦17 18"–102◦33 42" E; latitude 37◦18 12"–37◦25 18" N) located northwest of Tianzhu County at the eastern margin of the Qilian Mountains (Figure 1). The climate is generally characterized as a semiarid continental climate with water availability being the major abiotic factor-limiting plant growth. Mean annual precipitation is less than 400 mm and mean annual air temperature in this temperate location is approximately 1.5 ◦C, respectively. Moreover, the dominant forest species in this study area is *P. crassifolia*, which grows naturally with no manual management, and *Juniperus przewalskii* Kom., *Betula albo-sinensis* Burk. and *Populus davidiana* Dode are the second minor contributors to the shady understory. Additionally, the shrub species are mainly dominated by *Salix cupularis*, *Rhododendron simsii* Planch., *Caragana jubata* (Pall.) Poir. and *Spiraea alpine* Pall.

**Figure 1.** Geographic location of the collection sites (slg-2, slg-3, slg-4) of *Picea crassifolia* located in Shuanglonggou region.

#### *2.2. Sample Collection*

In 2015, *P. crassifolia* forests were selected from sites with similar conditions, such as topography, the composition of the undergrowth vegetation and stand age. Three plots (slg-2, slg-3, and slg-4 are shown in Figure 1) ranging from 2824 to 2914 m were established for study. Table 1 provides the general information related to the sampling sites. For each sampling plot, trees were divided into four/five groups according to the classes of diameter at breast height. Subsequently, there were five breast-height diameter classes determined and three to five trees with similar diameter in each group were chosen as the samples. The sampled *P. crassifolia* leaves developed in full light of an open canopy and were carried out from the upper third portion of the tree crown.

Moreover, considering the variation caused by differences in leaf nutrient contents at different orientations of the shoots, only leaves with a healthy appearance (avoiding damaged leaves) were cut with a pole pruner, and sampling was carried out from different orientations as much as possible. Overall, 4 years' leaves (from current year and up to 3 years-old) of each sampled tree were collected in our study. There was a clear joint on the branch between growing seasons, which aided in the determination of leaf age. We defined leaf age as 0 for the needles from the current, and then the next group was year 1 (the previous year+1) and so on. Thus, leaves aged 0–3 years were detached from the twigs before sending the samples to the laboratory for analysis. Additionally, for the estimation of tree age, tree-ring cores of selected *P. crassifolia* trees were collected at tree breast height (approximately

1.3 m above the surface) using an increment borer. Actual tree ages of sample cores were determined using dendrochronological methods [22].


**Table 1.** Site information for four sampling quadrats.

MH is mean tree height and M-dbh is mean diameter at breast height.

#### *2.3. Stable Carbon and Nitrogen Isotope Analyses*

The samples were washed in distilled water to remove dust particles, air dried before oven drying at 65 ◦C for 12 h and at 110 ◦C for 10 min to deactivate the enzymes, ground into a homogeneous fine powder, and sieved in the laboratory. A stable carbon and nitrogen analysis was performed in the Environmental Stable Isotope Laboratory, Institute of Environment and Sustainable Development in Agriculture, No.12, Zhongguancun South Street, Haidian District, Beijing 100081, China.

The isotopic compositions of the leaf samples were measured on an Isoprime100-EA mass spectrometer (Germany). The carbon or nitrogen isotope ratios are expressed relative to an international standard using the delta notation:

$$
\delta\_{\text{sample}} = (R\_{\text{sample}} - R\_{\text{standard}}) / R\_{\text{standard}} \tag{1}
$$

where δsample was defined by this relationship, *R*sample indicated the 13C/ 12C or 15N/ 14N ratio of the sample, and *R*standard indicated the 13C/ 12C or 15N/ 14N ratio of the standard. The international standard reference for carbon was PDB (Pee Dee Belemnite), and for nitrogen, it was an average of 15N/ 14N from atmospheric air [23].

#### *2.4. Statistical Analysis*

We analyzed the dataset by subdividing them into four groups based on leaf age (current year leaves, 1-year-old leaves, 2-year-old leaves, and 3-year-old leaves) and tree age (<50-year-old, 51 to 100-year-old, 101 to 150-year-old, and >150-year old), respectively. For leaf δ13C and δ15N values of different leaf ages and tree ages, the mean, median, standard error, and coefficient of variation (CV) were calculated, respectively. Here, the analysis was the leaf age–tree age combination. We first used two-way analysis of variance and Tukey's post hoc test to compare differences of leaf δ13C and δ15N values between leaf ages and tree ages. Next, the regression analysis was applied to investigate the relationship between leaf δ13C, leaf δ15N and leaf nutrients (leaf N, P concentrations and the C:N ratios). Furthermore, we calculated the relative importance (refers to the quantification of an individual regressor's contribution to a multiple regression model) of each predictor on leaf δ13C and δ15N with the R package relaimpo [24].

The R package relaimpo demonstrates six different metrics for assessing the relative importance of regressors (all regressors are uncorrelated) in the model [24]. Each predictor's contribution is just the R<sup>2</sup> from univariate regression, and all univariate *R*2-values add up to the full model *R*2. *R*<sup>2</sup> represents the proportion of variation in y that is explained by the p regressors in the model. Correlation analysis was conducted using the SPSS 22.0 [25] and R 3.2.4 [26].

#### **3. Results**
