Leaf nutrient resorption of two life-form tree species in urban gardens and their response to soil nutrient availability

Study site

The research was performed in Taiyuan city, Shanxi Province, China, at 37.27°N and 38.25°N. Taiyuan is located in central Shanxi Province and is surrounded by mountains on three sides(west, north and east) and the Jinzhong Basin on one side (south). The overall topography is high in the north and low in the south, and the Fenhe River flows through the city. With an average annual temperature of 9.5 °C and 456 mm of precipitation, the city has a warm temperate continental monsoon climate. The zonal soil type is mainly brown soil. Site condition differences among urban areas are mainly reflected in the limited differences in groundwater levels and soil characteristics caused by distance to a river.

Experimental design

Based on the total number and spacial frequency of the tree species (evergreen: deciduous = 1: 5.83) used in Taiyuan city, we selected 40 most generally used tree species belonging to two life forms of six evergreen trees and 34 deciduous trees.

To ensure representativeness and facilitate observation and sampling, the selected sample site was along the East Riverside Road. At the same time, the sample sites and management conditions were consistent. Among the three selected tree species, each tree species met the criteria of good growth, similar age, thorax/ground diameter, plant height and crown width. Each sample site contained three sample trees belonging to the same tree species. For each tree species, nine sample trees were selected from three sample sites. In total, 120 sample points (Fig. 1) and 360 sample trees were selected for 40 tree species (Table 1).

Sampling and measurement

Green leaf and soil samples were collected during mid-July 2021. Senesced leaf samples were collected in November 2021. Green leaf samples were collected from the upper, middle, and lower parts of the crown to create a mixed sample. The 0–30 cm soil samples were taken by a soil drill after the litter removed firstly. For shrubs, soil samples were collected from three sites at 120° angles away from the crown radius. For trees, they were collected from three sites at 120° angles 0.5 m away from the trunk. Senesced leaf samples were collected from the leaves remained on the stalk but fell off naturally when the trees were gently shaken. The green leaf samples, the soil samples and the senesced leaf samples came from three sample trees in each sample site were respectively mixed to one composite sample. Totally, 120 green leaf samples, 120 senesced leaf samples and 120 soil samples were collected.

The leaf samples were first dried at 105 °C before being dried at 65 °C to a constant weight. The dried leaves and soil were then crushed and passed through a 100-mesh sieve. The total nitrogen (TN) and total phosphorus (TP) concentrations (Table S1) were measured with the colorimetric method. The process involved adding 5 ml H₂SO₄ and 6 ml H₂O₂ to the plants and mixing 1.85 g catalyst and 5 ml H₂SO₄ with the soil. Then, the samples were digested in a deboiling furnace, filtered at a constant volume, and finally analyzed by an automatic discontinuous chemical analyzer (Smartchem 450) (Hu et al., 2022). The soil concentrations of NH${}_4^{+}$-N, NO${}_3^{-}$-N, NO${}_2^{-}$-N and available P (AP) (Table S2) were also tested with the colorimetric method.

Data analysis

Accounting for the leaf mass loss when leaves senesce, we recalculated the senesced leaf nutrient concentrations to compensate for the underestimation of NuRE using the mass loss correction factor (MLCF) (Van Heerwaarden, Toet & Aerts, 2003). The MLCF values varied according to life forms: 0.745 for conifers, 0.780 for evergreen broadleaved trees and 0.784 for deciduous broadleaved trees (Vergutz et al., 2012). In this research, all the data on nutrients in senesced leaves (Nu_sen) underwent quality correction. The calculated Nu_sen was also considered the nutrient utilization efficiency (Tang et al., 2013). The NuRE was quantified as follows: ${\displaystyle \text{NuRE}=\frac{{\mathrm{Nu}}_{\mathrm{gr}}-{\mathrm{Nu}}_{\mathrm{sen}}}{{\mathrm{Nu}}_{\mathrm{gr}}}\times 100\text{\%}}$ where Nu_gr represents the amount of nutrients in green leaves, while Nu_sen represents the amount of nutrients in senesced leaves.

The N in green leaves (N_gr), P in green leaves (P_gr), N in senesced leaves (N_sen), P in senesced leaves (P_sen), NRE, PRE and soil nutrient data were log10-transformed to meet the assumption of normality. One-way ANOVA was employed to analyse the N and P concentrations and N:P in green and senesced leaves between the two life forms. Additionally, the T test with Bonferroni adjustments was used to examine differences in the characteristics of nutrient absorption between the two life forms. A linear fitting equation was employed to analyse the correlations between soil parameters and the leaf index for the two life forms. Redundancy analysis (RDA) was used to identify the soil factors that explained the differences in N and P utilization between the two life-form tree species. We calculated the sorting axis lengths, and the lengths were all less than 3. We also calculated the variance inflation factor (VIF) for all soil factors using variance inflation factor analysis, and all VIF values were less than 10. All data are the mean ± SD. All data analyses were conducted in R 4.2.1 (R Core Team, 2022) and IBM SPSS Statistics 20 (SPSS, Inc., Chicago, IL, USA).

N and P concentrations and N:P in green and senesced leaves between the two life forms

The N (F = 46.17, p = 0.000) and P (F = 29.94, p = 0.000) concentrations and N:P (F = 5.20, p = 0.023) were significantly higher in the green leaves than in the senesced leaves of the 40 typical garden tree species (Table 2). There were significant differences between the N (F = 10.219, p = 0.002) and P (F = 5.455, p = 0.021) concentrations and N:P (F = 12.611, p = 0.001) in green leaves between the two life forms (Table 2). Senesced leaves between the two life forms exhibited significant differences in N (F = 12.802, p = 0.001) concentrations and N:P (F = 10.595, p = 0.001) and nonsignificant differences in P concentration (Table 2). Additionally, except for the P concentration in the senesced leaves, the N and P concentrations and N:P in the evergreen trees were significantly lower than those in the deciduous trees (Table 2).

NRE, PRE and their correlated relationship between the two life forms

The NRE and PRE of the evergreen and deciduous trees were 56.1 ± 13.1% and 41.1 ± 17.0%, respectively (Fig. 2). The PRE was significantly lower than the NRE (F = 58.733, p = 1.17e − 11). The NRE (F = 10.114, p = 0.002) and PRE (F = 4.991, p = 0.015) were significantly different between the evergreen and deciduous trees, and those of the evergreen trees were significantly lower than those of the deciduous trees (Fig. 2). The NRE was significantly positively correlated with the PRE among the deciduous trees (F = 10.68, p = 0.002) (Fig. 3). However, the NRE had no obvious relationship with the PRE among the evergreen trees.

Soil nutrient concentrations

According to the results, there were no significant differences in the soil nutrients between the evergreen trees and deciduous trees (Table 3).

Relationship between leaf indexes and soil nutrients in the two life forms

The NRE and PRE in the two life forms were distinctly correlated with soil nutrient concentrations and N:P. For the evergreen trees, the NRE (F = 5.628, p = 0.031) and N_gr (F = 6.393, p = 0.022) had a positive relationship with soil N (Figs. 4A–4B). The NRE (F = 4.529, p = 0.049), N_gr(F = 6.831, p = 0.019), P_gr (F = 5.231, p = 0.036) and P_sen (F = 5.673, p = 0.030) had a significantly positive relationship with soil N:P (Figs. 4J–4L and 4N). For deciduous trees, the NRE had a greatly negative relationship with soil N (F = 6.307, p = 0.014) (Fig. 4A), and the PRE had a similar relationship with soil P (F = 5.905, p = 0.017) (Fig. 4D). N_sen had a positive relationship with soil N, NH${}_4^{+}$-N, NO${}_3^{-}$-N, NO${}_2^{-}$-N and N:P (F = 14.05, p = 0.0003; F = 5.151, p = 0.025; F = 43.76, p = 1.86e − 09; F = 4.116, p = 0.045; F = 8.075, p = 0.005) (Figs. 4C, 4F, 4H, 4I and 4M). N_gr had a distinctly positive relationship with soil NO${}_3^{-}$-N (F = 26.87, p = 1.14e − 06) (Fig. 4G). P_sen had a significantly positive relationship with soil P (F = 5.186, p = 0.025) (Fig. 4E).

For the evergreen trees, the amount of variation in N_gr, P_gr, N_sen and P_sen in the first and second axes explained by the RDA was 39.36% and 7.634%, respectively. The cumulative amount of variation explained by the RDA was 46.994% (Fig. 5A). N_gr, P_gr, N_sen and P_sen were significantly correlated with soil TN (F = 5.5482, p = 0.002), and the RDA explained 21.4% of their variation (Fig. 5A). The variation in NRE and PRE explained by the first and second axes of the RDA was 32.89% and 8.814%, respectively. The cumulative amount of variation was 41.704% (Fig. 5B).

For the deciduous trees, the first and second axes of the RDA explained 22.76% and 1.058% of the variation in N_gr, P_gr, N_sen and P_sen, respectively. The cumulative amount of variation was 23.818% (Fig. 5C). N_gr, P_gr, N_sen and P_sen were substantially associated with soil TN (F = 6.9873, p = 0.004), NO${}_3^{-}$-N(F = 17.8502, p = 0.001) and AP (F = 3.0538, p = 0.040), and the RDA explained 1.3%, 20.6% and 0.09% of their variation, respectively (Fig. 5C). The first and second axes of RDA explained 8.423% and 1.165% of the variation in the NRE and PRE, respectively. The cumulative amount of variation was 9.588% (Fig. 5D). The PRE and NRE were distinctly related to the soil TP ( F = 4.7381, p = 0.025), and their variation was 5.6% (Fig. 5D). All the data accurately reflected the relationship between leaf N_gr, P_gr, N_sen, P_sen, NuRE and soil nutrients.

N and P concentrations and N:P in leaves

Leaf nutrient concentration is a key indicator of plant nutritional status (Chen et al., 2021a; Chen et al., 2021b). In our research, the green leaf N concentration (13.38 g kg⁻¹) was lower than the average N concentration (17.57 g kg⁻¹) in woody plants in the north-south transect of eastern China (NSTEC) (Ren et al., 2007). Plants mainly obtain P and N from the soil (Zhu et al., 2016; Zhang et al., 2018), and the leaves in urban green spaces are removed regularly every year, resulting in serious soil N loss (Hu et al., 2022). The soil N concentration in this area was low (TN = 0.61 g kg⁻¹), so the leaves obtained less N from the soil. Since plants in southern China are generally limited by P and the soil P concentration in northern China is relatively sufficient, the leaf P concentration (2.90 g kg⁻¹) in this area was relatively higher than that in woody plants (1.39 g kg⁻¹) in the NSTEC (Ren et al., 2007).

In this research, the green leaf P and N concentrations in the deciduous trees were significantly higher than those in the evergreen trees. This result might be attributed to the fact that deciduous trees need to accumulate more N and P in a short time to complete growth-related activities. Due to the longer leaf life cycle of evergreen trees, they take much more time to conserve nutrients (Yan, Zhu & Yang, 2018). Compared with evergreen trees, deciduous trees had higher levels of litter decomposition and nutrient absorption (Wu et al., 2012). The calculated Nu_sen was also considered the nutrient utilization efficiency (Tang et al., 2013). In this research, the senesced leaf N concentration in the evergreen trees was significantly lower than that in the deciduous trees, indicating that the N utilization efficiency was higher than the P utilization efficiency (Tang et al., 2013) in this area. They were consistent with hypothesis 1.

In most terrestrial ecosystems, the availability of N and P limits the growth of plants. The concentrations of N and P in green leaves and the N:P can be used as a standard of plant nutritional limitation (Tian, Yan & Fang, 2021). Plants are limited by N when N:P <14 and when the N concentration is <20.0 mg g⁻¹ and the P concentration is >1 mg g⁻¹ (Wu et al., 2012). In this study, the leaf N:P in the two life forms was less than 14. Additionally, the leaf N concentrations were less than 20.0 mg g⁻¹ in most trees, and the P concentrations were more than 1 mg g⁻¹ in 40 trees. The N concentrations were lower than those of the same tree species in other areas (Wang et al., 2013), and the N:P was lower than that of 753 species in China (Dong et al., 2023). Combined with the field investigation, according to the growth and physiological characteristics of the trees, we had reason to believe that the garden trees in this area were limited by N. In addition, the N:P in the deciduous trees was higher than that in the evergreen trees, indicating that the evergreen trees were more sensitive to being limited by N than the deciduous trees. The results also showed that under the same urban garden background, N limitation was different between the evergreen and deciduous trees and which was consistent with hypothesis 2. The reason for the difference in N limitation between the two life forms in the same area may be the genetic characteristics of the tree species themselves or the variation in suitable survival to the external environment formed in the long-term evolution process.

Differences in NRE and PRE in the two life forms

Senesced leaves fall to the ground and gradually breakdown into inorganic nutrients due to physical and microbiological processes, and these remineralized nutrients will eventually be absorbed by plants(biogeochemical cycle) (Han et al., 2013). The NRE (56.13%) was higher than the PRE (41.14%), indicating that plant absorption of N was higher than that of P. However, these NRE and PRE values are lower than those (62.1% and 64.9%, respectively) of global terrestrial forest ecosystems. This result might be due to the differences in the trees or habitats selected for this research and how urbanization associated with N and P inputs has altered the soil N and P status and plant N and P absorption (Hu et al., 2011). Additionally, when plants are restricted by N, they tend to accomplish total N absorption but inadequate P absorption (Chen & Chen, 2021).

Compared with the evergreen trees, the deciduous trees had much higher NRE and PRE, indicating that nutrient utilization in the evergreen trees was lower than that in the deciduous trees (Chen & Chen, 2021). P resorption variation is more affected by climate and soils, while N resorption variance relates more to plant functional type (Tang et al., 2013). Evergreen trees, a type of highly specialized plant, have developed an adaptation technique to thrive in nutrient-poor soil (Reich, Walters & Ellsworth, 1992). In barren soil, extending the leaf lifespans of evergreen trees can allow the growth rate to reach its maximum (Reich, Walters & Ellsworth, 1997) and nutrient losses to reach their minimum during leaf senescence and shedding (Aerts, 1995). During the growth season, deciduous trees produce more leaves, and they shed their leaves annually (Reich et al., 1997). Thus, deciduous trees often invest more nutrients to support rapid growth during shorter growing seasons (Brant & Chen, 2015). Thus, in this study, in comparison to evergreen trees, deciduous trees had much higher NRE and PRE. These findings indicated that the two life-form garden tree species have different N and P utilization strategies, which is consistent with hypothesis 1. Additionally, the variation in the decomposability-related characteristics of the litter (such as lignin: N and N:P) and the habitat climate (temperature/precipitation) may work in conjunction to affect the availability of soil nutrients and, consequently, the levels of NRE and PRE in evergreen and deciduous trees (McGroddy, Daufresne & Hedin, 2004).

In addition, in the deciduous trees, the NRE was positively associated with the PRE, indicating that N absorption increases concurrently with P absorption. This result was a crucial premise for stoichiometry control (Chen & Chen, 2021), further showing that the deciduous trees had a synergistic effect on N and P. A meta-analysis by Aerts(Aerts, 1995) also showed that NRE and PRE have a strong association with one another. Our result is consistent with this finding.

Relationship between leaf N_gr, P_gr, N_sen, P_sen, NuRE and soil nutrients in the two life forms

Soil nutrient availability and leaf nutrient status are considered important control factors for nutrient absorption processes (Xu et al., 2020; Chen & Chen, 2021). Our findings demonstrated that soil nutrients had a considerable impact on leaf N and P concentrations as well as on NuRE in the two life forms. The availability of N and P had a significant impact on how soil nutrients were utilized, which had an impact on NuRE (Liu et al., 2020).

For the evergreen trees, the NRE and N_gr were positively correlated with soil TN, indicating that the evergreen trees did not reduce their resorption levels to adapt to the increase in soil N concentration. The increase in the NRE was mainly due to the decrease in the N_gr. We predicted that the evergreen trees would store a large amount of N in their green leaves when the soil N supply was sufficient or possibly N-limited, while the N in senesced leaves was almost unaffected, thus improving the NRE. Soil N:P can be used as a measure of N saturation, showing the availability of nutrients for plant growth (Stark, Ylanne & Tolvanen, 2018). The P_gr and P_sen were positively related to soil N:P, indicating that the evergreen trees adopted a nutrient utilization strategy with high concentration and high return to maintain their normal physiological activities when soil P was relatively scarce compared with N. The NRE and N_gr in the evergreen trees were positively associated with soil N:P, and the RDA (Fig. 5A) showed that soil TN had the highest variation, which indicated that soil TN supply was the key factor affecting the N and P concentrations in the evergreen tree leaves.

For the deciduous trees, the NRE was negatively correlated with soil TN, while the PRE was negatively associated with soil TP, indicating that the NRE and PRE decreased with increasing soil TN and TP. We concluded that the deciduous trees did not need higher resorption efficiency to maintain N and P concentrations when the N and P supply in the soil was sufficient. The result is consistent with that of Kobe, Lepczyk & Iyer (2005). N_sen was positively associated with soil N in various forms and N:P, indicating that N_sen increased with increasing soil N concentration and availability. The deciduous trees mainly affected the N_sen through the soil N concentration, which further affected the NRE. The dominant soil factors affecting the N and P concentrations in the deciduous trees were soil TN and NO${}_3^{-}$-N (Fig. 5C), and soil TP was the main soil factor affecting the NuRE (Fig. 5D), indicating that soil N mainly influenced the leaf N and P concentrations; however, soil P mainly affected the leaf NuRE in the deciduous trees. The evergreen trees experienced more variation in soil accumulation than the deciduous trees, demonstrating that evergreen trees were more sensitive to soil N and P concentrations than deciduous trees, which is consistent with hypothesis 3.