Applying foliar stoichiometric traits of plants to determine fertilization for a mixed pine-oak stand in the Qinling Mountains, China

Site description

Experiments were conducted at the Qinling National Forest Ecosystem Research Station (QNFERS), located on the southern slope of the Qinling Mountains, Huoditang, Ningshan County, Shaanxi Province, China (32°18′N, 108°20′E). The altitude of the study area was from 1500 to 2500 m. The area has subtropical climate, with annual mean air temperatures around 8–10 °C, annual mean precipitation around 900–1,200 mm, and annual mean evaporation around 800–950 mm. The main soil type was mountain brown soil, developed from granite material, ranging from 30–50 cm depth. The total forest area in the station was 2,037 hectares. Natural forest occupied 93% of the total forest area in QNFERS, with various vegetation types distributed along an altitudinal gradient, such as evergreen deciduous mixed forest (mixed pine-oak forest), deciduous broad-leaved forest (oak, red birch), temperate coniferous forest (Chinese red pine, Armand pine), and cold temperate coniferous forest (spruce, fir). The most dominant forest type was mixed pine-oak forest restored after rotation felling between 1950 and 1998 with an average stand age of 42 years and an average height of 9.2 m. Common tree species included Pinus tabuliformis, P. armandii, and Quercus aliena var. acuteserrata, associated with Toxicodendron vernicifluum, Carpinus turczaninowii, Swida macrophylla, Acer mono, Tilia paucicostata, Carpinus cordata, Juglans cathayensis, Bothrocaryum controversum, Rhus potaninii, Dendrobenthamia japonica var. Chinensis, Pterocarya stenoptera, Acer davidii, Betula albo-sinensis, Betula albo-sinensis var. septentrionalis, Juglans mandshurica, Rhus punjabensis Stewart var. sinica, Prunus padus, and Sorbus folgneri, and understory species were abundant (Hou et al., 2017).

Field investigations and sampling were conducted from 10 to 15 September, 2014. Samples of litter and soil were collected from 13 long-term fixed plots in the mixed pine-oak stand. Each plot was 20 × 20 m. Tree species, number, Height (H), and diameter at breast height (DBH) were recorded and used to determine the dominant tree species and DBH classes. Each increase in DBH by 4 cm is a DBH class and mid-diameter is used to stand for it (DBH ranging from 6.1 to 10 cm, the corresponding mid-diameter is 8 cm). One hundred and fifty-six of the average standard trees (AST) in all DBH classes were determined in each of the 13 plots (File S1). Mature, sunlit leaves without disease or insect pests were sampled. Four leaves/needles from each AST were collected in each of the four directions (north, east, south and west) and at different stem heights (crown, intermediate section, and underpart), and mixed into one sample for each AST. Five subplots (1 × 1 m) within each plot were randomly established to collect litter and mineral soil (depth 0–30 cm, 10 cm each layer). Samples of litter were collected manually and soil samples were collected by an auger (internal diameter 38 mm). Soil volumetric rings (100 cm³) were also used to collect soil samples to measure soil bulk density using the cutting ring method. Five samples of litter and mineral soil from the same depth within each subplot were mixed into one sample. All samples (litter and mineral soil) were weighed in the field before being transported to the laboratory.

Chemical analysis

Samples of leaves/needles and litter were oven-dried at 60 °C to a constant weight, and then ground using a plant sample mill and sieved through a 1-mm mesh screen. Nitrogen concentrations in leaves/needles and litter were measured using a flow injection analyzer (FIA5000; FOSS, Hoganas, Sweden), while P concentrations were measured using the molybdenum blue colorimetric method after digestion in a H₂SO₄+H₂O₂ solution (Bao, 2000). Auger soil samples were air-dried under shade, then ground and sieved though a 0.149-mm mesh. Total soil N (TN) was measured using the Kjeldahl method (Kjeltec TM 8400; FOSS, Hoganas, Sweden), and total soil P (TP) was determined using acid digestion in a H₂SO₄+HClO₄ solution (Bao, 2000).

Data processing

Importance values (IV) of tree species were calculated following Busby, Vitousek & Dirzo (2010): (1)${\displaystyle \mathrm{RH}=\frac{\sum H_i}{H}\times 100\text{\%}}$ where, RH is the relative height, H_i, is the height of tree species i, and H is the height of all tree species in all plots. (2)${\displaystyle \mathrm{RF}=\frac{\sum F_i}{F}\times 100\text{\%}}$ where, RF is the relative frequency, F_i is the frequency of tree species i, and F is the frequency of all tree species in all plots. (3)${\displaystyle \mathrm{RD}=\frac{\frac{1}{4}\pi f\sum \underset{i}{\overset{2}{D}}}{D^2}}$ where, RD is the relative basal area, D_i is the DBH of tree species i, D is the basal area of all tree species in all plots, and f is the form factor of tree species (f_conifer = 0.40, f_broadleaf = 0.42). (4)${\displaystyle \mathrm{IV}=\frac{\mathrm{RH}+\mathrm{RF}+\mathrm{RD}}{3}\times 100\text{\%}.}$

Tree species with IV>10% were determined as dominant.

The big timber rate of dominant tree species was calculated as following. ${\displaystyle R_{bi}=\frac{N_{bi}}{N_i}\times 100\text{\%}}$ where R_bi is big timber rate (%) of tree species i, N_bi is stems of big timber of tree species i and N_i is total stems of tree species i in the plots.

Nitrogen and P stoichiometry (N and P concentration and N:P) of plants’ leaves was calculated following Du et al. (2011): (5)${\displaystyle S_t=\sum _{i=1}^n{C}_{ni}\times {\mathrm{IV}}_i}$ where, S_t is N and P stoichiometry of all tree species, C_ni is the concentration of N or P (mg g⁻¹) or N:P in leaves of tree i, IV_i is the importance value of tree i, and n is the number of dominant tree species.

Nutrient storage (t ha⁻¹) in litter was calculated as follows: (6)${\displaystyle S_l=C_l{B}_l}$ where, S_l is the litter nutrient storage (t ha⁻¹), C_l is element (N or P) concentration (mg g⁻¹), and B_l is the biomass of litter (t ha⁻¹).

Storage of nutrients (N and P) in soil was calculated as follows: (7)${\displaystyle S_T=\sum _{i=1}^n{E}_i{B}_i{D}_i}$ where, S_T is the soil nutrient storage (t ha⁻¹), E_i is the concentration of element i (mg g⁻¹), B_i is the bulk density (g cm⁻³) in layer i, and D_i is depth i (cm).

Differences of N, P concentrations and N:P between dominant tree species were analyzed using Statistical Analysis of the T-Test. The extremely significant difference (p < 0.01) , significant difference (0.01 < p < 0.05) and not significant difference (p > 0.05) between paired variables were demonstrated as different capital letters, different lowercase letters and the same lowercase letter respectively. The correlations between foliar N and P and N:P in the plants, foliar N and P, N:P in the plants and mineral soil, and litter were assessed using Spearman’s rank correlation. Significance levels were set at p = 0.05. All statistical analyses were performed using SPSS software (version 19.0 for Windows; SPSS Inc., Chicago, IL, USA). Figures were plotted using Origin 8.0 (OriginLab Corporation, Northampton, MA, USA).

Dominant tree species

There were 21 tree species in the mixed pine-oak stand. Aside from the photophilous species (P. tabuliformis, P. armandii, and Q. aliena var.acuteserrata), most of the species were shade-tolerant (Toxicodendron vernicifluum, Carpinus turczaninowii) or neutral with respective to light (Acer mono, Sorbus folgneri) (Table 1). The dominant tree species with importance values >10% were P. tabuliformis, P. armandii, and Q. aliena var.acuteserrata (Table 1).

The big timber rate for P. tabuliformis, P. armandii, and Q. aliena var. acuteserrata was 57.71%, 22.79% and 2.78% of current existing individuals respectively (Table 2).

Patterns of foliar N and P and N:P in the plants

Foliar N (11.84 ± 2.36 mg g⁻¹ to 21.72 ± 3.19 mg g⁻¹) and P (1.20 ± 0.14 mg g⁻¹ to 2.00 ± 0.31 mg g⁻¹) and N:P (9.62 to 10.87) exhibited large variation among tree species (Table 3). The general trend demonstrated that foliar nutrients and N:P of P. tabuliformis and P. armandii were less than that in Q. aliena var. acuteserrata (Table 3). Moreover, foliar stoichiometric variables of the plants were also less than those in individual tree species in the mixed pine-oak stand (Table 3).

In the plants, concentrations of foliar N and P were significantly, positively correlated (p = 0.000) with each other (Fig. 1A). The mathematical Log foliar N concentration was significantly (p = 0.02), positively correlated with the log N:P (Fig. 1B). Foliar P concentration and N:P were significantly, negatively correlated (p = 0.000) with each other (Fig. 1C).

Patterns of N and P and N:P in mineral soil and litter

Nitrogen concentrations and N:P varied markedly across mineral soil at 0–30 cm depth, ranging from 0.60 ± 0.05 mg g⁻¹ to 2.40 ± 0.10 mg g⁻¹ and 2.60 ± 0.41 to 6.81 ± 0.51, respectively (Table 4). However, P in mineral soil at 0–30 cm depth was significantly different, ranging from 0.23 ± 0.03 mg g⁻¹ to 0.36 ± 0. 03 mg g⁻¹ (Table 4). Bulk density of mineral soil increased with soil depth (Table 4). The indices, mean N:P and concentrations of N and P in litter were 2.49, 12.06 and 5.21 times of which in mineral soil respectively (Table 4).

Correlations between N and P concentrations in litter and mineral soil were complex and none was significant (Figs. 2A and 2B). At 0–10 cm depth, mineral soil N concentration increased with litter N concentration when litter N concentration was <17.9 mg g⁻¹, and vice versa (Fig. 2A). Soil P concentration increased with litter P concentration (Fig. 2B). At 11–20 cm depth, mineral soil N concentration decreased with litter N concentration when litter N concentration was <19.0 mg g⁻¹, and vice versa (Fig. 2A). Soil P concentration showed an exponential correlation with litter P concentration (Fig. 2B). At 21–30 cm depth, mineral soil N concentration increased with litter N concentration when litter N concentration was <17.9  mg g⁻¹, and vice versa (Fig. 2A). Phosphorus concentration of mineral soil increased linearly with litter P concentration (Fig. 2B).

Correlation between foliar N and P concentration in the plants and mineral soil

Foliar N concentration in the plants increased with N concentration of mineral soil at 0–10 cm depth (Fig. 3A). In contrast, foliar P concentration in the plants showed a significant, cubic correlation with P concentration of mineral soil (p = 0.048) (Fig. 3B). Although foliar N and P concentrations in the plants were not significantly (p = 0.09) correlated with mineral soil at 11–20 cm depth, foliar N concentration in the plants increased with N concentration of mineral soil (Figs. 3A and 3B). There was no correlation between foliar N concentration in the plants and N concentration of mineral soil at 21–30 cm depth (Fig. 3A). Foliar P concentration in the plants increased with P concentration of mineral soil at 21–30 cm depth (Fig. 3B).

Foliar N and P stoichiometric traits in the plants in mixed pine-oak stand

The positive correlation between foliar N and P concentrations (p = 0.000, Fig. 1A) in the plants is consistent with stoichiometric stability criteria of a fixed ratio of nutrient absorption in the plant (Koerselman & Meuleman, 1996). Previously, mean foliar N and P and N:P were reported as 18.6, 1.21 mg g⁻¹, and 14.4, respectively for 753 species of terrestrial plants across China (Han et al., 2005) and 18.3, 1.42 mg g⁻¹, and 11.8, respectively for 1,251 world terrestrial plants (Reich & Oleksyn, 2004). This study indicated that foliar N:P of dominant tree species (P. tabuliformis, P. armandii, and Quercus aliena var. acuteserrata.) and in the plants (10.30:1) were lower than both Chinese and global terrestrial plants. Similarly, foliar N concentration of P. tabuliformis (11.84 ± 1.40 mg g⁻¹) and P. armandii (12.98 ± 1.43 mg g⁻¹), and foliar P concentration of P. armandii (1.20 ± 0.14 mg g⁻¹) were also lower. The overall element composition of plants in an ecosystem is determined by the mix of species and by the physiological status of the dominant plants (Güsewell, 2004). Therefore, potential explanations for the observed patterns are that foliar element concentrations and ratios were strongly determined by genetic and physiological controls, and that these crucial factors prevented plants from responding to the natural availability of nutrients (Castle & Neff, 2009). Furthermore, the nutrient status of terrestrial plants has a strong local and regional signal due to acquiring nutrients via weathering and microbial decomposition in situ (Chadwick et al., 1999). In the study area, low air temperature (<10 °C) (Hou et al., 2017) may have hindered rock weathering and microbial activity. However, the concentration of foliar N (21.72 ± 3.19 mg g⁻¹) and P (2.00 ± 0.31 mg g⁻¹) in Q. aliena var. acuteserrata was higher than that reported for Chinese and global mean levels. Foliar P concentration of P. tabuliformis (1.23 ± 0.22 mg g⁻¹) was higher than the Chinese mean, but lower than the global mean. The growing period (from mid-May to late-September) in the study area was relatively short for plots on high elevation (more than 1,600 m) (Hou et al., 2017). The leaf life span of P. tabuliformis and P. armandii in the study area is 3 years. In the contrary, the leaf life span of Q. aliena var. acuteserrata is only 5 months. Therefore, shorter leaf life span and growing season may have caused higher N and P concentrations (Castle & Neff, 2009; He et al., 2006).

Response of foliar nutrients in the plants to soil nutrients

The N and P concentrations in mineral soil were generally low and the similar result was also found by Wu (2015). The N concentration in the top layer of mineral soil (0–10 cm) was above the standard of first class soil (>2.0 g kg⁻¹) (National Soil Survey Office of China, 1998); however, at 11–20 cm depth it only met the standard of third class soil (1–1.5 g kg⁻¹), and at 21–30 cm it met the standard of fifth class soil (0.5–0.75 g kg⁻¹) (National Soil Survey Office of China, 1998). The level of soil P concentration at 0–30 cm only reached the standard of third class soil (0.2–0.4 g kg⁻¹) (National Soil Survey Office of China, 1998). Litter decomposition and rock weathering are the main nutrient sources for mineral soil in natural forests. Plant element concentrations are largely determined by supplies of elements in soil and the chemical and physical characteristics of soil environments (Castle & Neff, 2009). However, our findings demonstrated that elements in litter did not significantly affect mineral soil. We found most of the correlations between foliar N and P concentrations in the plants and mineral soil at various depths were not significant, and soil nutrients did not explain more than 14% of foliar nutrients in the plants. Therefore, these results were not able to provide a better understanding and interpretation of the effects of litter and soil nutrients on mineral soil and foliar nutrients in the plants. Possible explanations for the observed patterns include high non-soluble chemical compounds in litter and low temperature (Hou et al., 2017), which slowed down litter decomposition. Furthermore, soil available N and P absorbed by plants have strong mobility and are easily leached (Thompson et al., 2010). The thickness of mineral soil (at 30–50 cm depth), pore ratio from 32.0 to 62.28%, high precipitation (800–1,200 mm), and low temperature (Hou et al., 2010; Hou et al., 2017) in the study area may have accelerated leaching of N and P in decomposed litter and mineral soil, leaving less N and P to be assimilated into leaves of the plants. In addition, forest harvesting in 1950–1998 was also key cause to decrease N and P concentrations in mineral soil.

Nutrient limitation of plants

It has been suggested that biomass N:P may be a better indicator of N or P deficiency than nutrient concentrations (Güsewell, Koerselman & Verhoeven, 2003). Previous studies have reported that plant growth is limited by concurrent N concentration <20 mg g⁻¹ and N:P < 14; while P limits plant growth at concentrations of <1 mg g⁻¹ and N:P >16. Therefore, co-limitation of N and P will occur when the concentrations and ratio of N and P meet these conditions (Aerts & Chapin III, 1999; Koerselman & Meuleman, 1996). The foliar N:P (10.3) and N concentration (9.08 mg g⁻¹) of the plants in the mixed pine-oak (Table 3) stand showed N limitation according to above mentioned criteria (Aerts & Chapin III, 1999; Koerselman & Meuleman, 1996).

The correlations between soil nutrient storage and foliar N:P of the plants indicated that foliar N:P of the plants was correlated with soil N storage (Fig. 4A). One possible reason may be that the growth of the plants was limited by N, and foliar N:P of the plants responses to soil N were more sensitive. However, N storage in mineral soil only explained 40% (Fig. 4A) of the variation in foliar N:P of the plants, and foliar N:P of the plants was not significantly correlated with soil P storage (Fig. 4B). This may have been a result of low P concentration in mineral soil, which also affected foliar N:P of the plants.

Our results showed that the total standing crop of elements in litter was estimated as 0.31 ± 0.07 t ha⁻¹ for N, and 0.02 ± 0.01 t ha⁻¹ for P; and 4.58 ± 0.15 t ha⁻¹ for N and 0.95 ± 0.09 t ha⁻¹ for P in mineral soil at 0–30 cm depth (Table 4). This suggested that trees might assimilate N and P entirely from that stored in mineral soil and litter. According to the criteria (Aerts & Chapin III, 1999; Koerselman & Meuleman, 1996), 1.49 t ha⁻¹ pure N should be used to fertilize forest land to relieve N limitation on tree growth.

Limitations of the analysis

We analyzed stoichiometric traits of leaves in the plants, mineral soil, and litter to explore the effects of N and P concentrations and N:P of litter and mineral soil on leaves in the plants. We also investigated the correlations between foliar N:P in the plants and N and P storage in mineral soil. The factors included in this study only explained 40% of the variation, which prevented us from fully understanding and interpreting total variation. Furthermore, besides criteria we used in the study, a more conservative estimate of N:P threshold is <10 for N limitation and >20 for P limitation (Güsewell, 2004; Sardans, Rivasubach & Peñuelas, 2012). Since there are multiple criteria for assessing nutrient limitation (Aerts & Chapin III, 1999; Güsewell, 2004; Koerselman & Meuleman, 1996; Sardans, Rivasubach & Peñuelas, 2012), we lack sufficient evidence to support an accurate estimation of fertilizer application rate. Finally, nutrient use efficiency and its influence on tree growth was not included in the estimation of pure N application rate, as it would have underestimated N. To verify these preliminary results, more study is required to detect the effects of fertilizer on stoichiometric traits of trees and mineral soil, and N and P interactions.