Nutrient resorption and adaptation strategy characteristics of two plantations in subtropical China­—Ecological stoichiometry


 Background and aims

Leaf nutrients resorption is critical for considerations of how plants use and recycle nutrients in terrestrial ecosystems. However, information on nutrient resorption and adaptation strategies of the same plant species growing in areas with different geological backgrounds remain poorly understood.
Methods

We investigated one natural plantation of Pinus yunnanensis Franch. (PY) and one introduced plantation of Eucalyptus maideni F. Muell. (EM) growing under the same climatic conditions but different geological backgrounds (limestone, karst area vs clasolite, non-karst area) in Yunnan Province, China. The C, N, and P concentrations, nutrient restriction, nutrient resorption efficiency, stoichiometric homeostasis, and plant adaptation strategy indicators were investigated.
Results

The results showed that soil concentrations of C, N, and P were significantly higher in the karst areas compared to the non-karst areas both of the two plantations. Elemental composition of specific plant organs differed significantly between the two sites, while within sites, different organs showed different elemental compositions. In addition, leaf N: P ratios and leaf P resorption efficiencies indicated that plantations in subtropical China are mainly limited by P, which was more evident in the non-karst area. The PY plantation in both areas showed a “conservative consumption” nutrient use strategy, whereas the EM plantation in the two areas showed a “resource spending” nutrient use strategy.
Conclusions

Plants need to adapt physiologically and morphologically to the harsh conditions in karst areas, resulting in lower growth rates and biomass, more conservative nutrient use, and a high capacity to retain nutrients in the biomass. The findings of this study indicated that trees could synergistically accommodate leaf stoichiometry and nutrient resorption efficiencies in response to different soil types. Overall, our results provide support that the geological background should be considered during the process of vegetation restoration.


Introduction
Stoichiometry greatly affects ecosystem dynamics and functions due to their association with ecological processes including ecosystem species composition, species diversity, nutrient limitation, and the environmental adaptation of organisms Bai et al. 2019). Previous studies shown that Page 4/25 al. 2018a,b;Pang et al. 2019). However, information about the growth trees used for environmental restoration in such fragile regions is still scarce.
Karst ecosystems are highly vulnerable to climatic change not only due to the affect soil C dynamics (Hagedorn et al. 2010), but also because of the impact on ecosystem function. Therefore, studies on the biogeochemical cycling of C, N, and P in karst ecosystems may provide important information for the development and optimization of forest restoration programs in fragile environments. Therefore, one natural plantation composed of Pinus yunnanensis Franch. (PY) and one introduced plantation composed of Eucalyptus maideni F. Muell. (EM) were selected in a karst and non-karst areas of southern China to determine the differences in soil and plant ecological stoichiometry, nutrient restriction, NuRE, stoichiometric homeostasis, and plant adaptation strategy. The objectives were to (1) perform a quantitative research on the differences in soil and plant C, N, and P concentrations as well as their ratios; (2) provide references for the nutrient restriction and nutrient resorption, as well as different plant growth strategies; and (3) uncover the relationship of N and P resorption e ciency with leaf N and P concentrations and their ratio. These results provide insight into theoretical stoichiometric characteristics and support ecological restoration and conservation of plantations in fragile karst regions.

Materials And Methods
Experimental site and sampling area This research was conducted at the Xiaoguan Forest Farm, Jianshui County, Yunnan Province, southwest China (104°50′E, 25°40′N). The distance between the two sites (non-karst and karst) was less than 10 km. Because of the close distance, the two sites had a similar climate, although they had different geological backgrounds. The climate at the two sites is a typical subtropical monsoon climate, the mean annual temperature of approximately 19.8°C, an annual sunshine duration of 2,322 h, mean annual precipitation of 805 mm, rainfall is concentrated from May to October, and the frost-free period is approximately 307 d per year. Soils in the non-karst area have developed on clasolite, while soils in the karst area were limestone.
In both sites, the two plantations (PY and EM) were within the scope of the 'Grain to Green' project, meaning that prior to vegetation restoration, both sites were barren mountains. In June 2017, we established twelve sampling plots (20 × 20 m, three replicates) in each area; the plots were located more than 50 m from the forest edge to avoid edge effects. The two areas were similar in terms of vegetation, altitude, slope, and planting pattern. Vegetation surveys in both plantations were conducted in July 2017. The non-karst forests were dominated by trees, with a scattered distribution of shrubs and herbs and large areas of bare soil. The karst forests were dominated by trees and had a large number of shrubs and herbs, covering more than 75% of the surface. Compared to the karst areas, the two plantations in the non-karst areas were much taller and had a greater diameter at breast height ( Fig. 1 and Table 1). Based on the height (H) and diameter at breast height (DBH, 130cm) ≥ 5 cm in each plot, the biomass was estimated using allometric equations . The biomass of the two plantations were signi cantly higher in non-karst areas than in the karst areas (Table 2).

Sample collection and measurements
In each plot, soils were sampled in July 2017 using 4-cm diameter soil cores from a depth of 0-20 cm. The litter layer was removed, the ve cores were combined to form one composite soil sample of each plot. Plant samples were collected from twelve dominant trees within each plot. Green leaves were collected in July and senescent leaves were shaken off from branches towards the middle of the trees using a rod (Du et al. 2016); branches, stems, and roots were also collected at the same time. After collection, the plant materials were gently rinsed with deionized water and dried to constant mass (65°C for 72 h). The plant organs and the air-dried soil samples were ground and sieved through a 60-mesh sieve for analysis. To measure the concentration of soil organic carbon, 0.5 g soil samples were pretreated with HCL (10 mL, 1 mol L -1 ) for 24 h to remove carbonate and then analyzed using an elemental analyzer (vario MAX CN, Elementar in Germany) (Mu et al. 2016). Nitrogen (N) concentrations of soil and plant samples were determined after semi-micro Kjeldahl digestion using a ow injection autoanalyzer (Bao 2000). Phosphorus (P) concentrations were analyzed colorimetrically by the ammonium molybdate method (Luo et al. 2013).

Statistical analyses
The stoichiometric ratios of C, N, and P in plant organs and soil were calculated based on the mass of C, N, and P. To determine signi cant differences between the plant organs (green and senescent leaves, branches, stems, and roots) and soils in terms of C, N, P, we used one-way analysis of variance (ANOVA). The effects of different vegetation types and geological backgrounds; different vegetation types and plant organs on C, N, P, and stoichiometric ratios were analyzed by two-way ANOVA. The stoichiometric characteristics of soil and plant organs were analyzed via least square difference (LSD) tests, with a signi cance level of 0.05. All data analyses were performed using SPSS 19.0 (SPSS Inc., Chicago. IL, USA), and gures were generated via Origin 9.0 (Origin Lab., Hampton, MA, USA).
Nutrient resorption e ciency (NuRE, %) was calculated as the following equation: see equation 1 in the supplementary les.
Where Xgre and Xsen are the nutrient concentrations in green and senesced leaves, respectively, and MLCF is the mass loss modifying factor of 0.745 for conifer, and 0.784 for deciduous (Leonardus et al. 2012).
The homoeostatic coe cient (H) was calculated using the nutrient concentrations of individual leaves from the two plantations and the soil nutrients .

Results
Soil C, N, and P concentrations and stoichiometry Soil C concentrations were higher in the karst area than the non-karst area in both plantations (p 0.05). There was no signi cant difference between the two plantations in the non-karst area (Fig. 2a). Soil N concentrations have no signi cant difference between the two plantations both in the karst and non-karst areas (Fig. 2b). Soil P concentrations were higher in the karst area than the non-karst for both plantations (p 0.05), and there was no signi cant difference between the two plantations in the non-karst area (Fig.  2c). In general, the three elements concentrations were signi cantly higher in the karst area than the nonkarst areas for both plantations.
The C: N: P ratios in different vegetation types were signi cantly different. The soil C: N and C: P ratio in the karst area were signi cantly higher than the non-karst area for the two plantations (p 0.05), whereas there was no signi cant difference between the two plantations ( Fig. 2d,e). The soil N: P ratio in the nonkarst area were signi cantly higher than the karst area for both plantations (p 0.05), while there was no signi cant difference between the two tree species growing in different areas (Fig. 2f). The concentrations of soil C and P were signi cantly affected by vegetation types, geological backgrounds and their interaction (Table 3).

C, N, and P concentrations and stoichiometry of plant organs
The C concentrations in leaves, branches, stems, and roots were signi cantly higher in the karst area than in the non-karst area within the same plantation, except for the stems of EM (Fig. 3a). The N concentrations in roots were signi cantly higher in the karst area than in the non-karst area within the same plantation (Fig. 3b). The N concentrations in EM were not signi cantly different in leaves and stems. The P concentrations in leaves were not signi cantly different among the two plantations in the two different areas (Fig. 3c). In general, the C, N, and P concentrations differed signi cantly between plant organs and sites.
The C: N ratios in leaves and stems of PY showed opposite trends in the karst and non-karst areas, whereas no signi cant difference was observed in branches or roots (Fig. 3d). The C: N ratios in each organ of EM were not signi cantly different between the karst and non-karst areas. The pattern of C: P ratios in leaves were not signi cantly different among the two plantations in the two areas (Fig. 3e). The N: P ratios in leaves and roots were higher in the karst area than the non-karst area for the two plantations (p 0.05). In contrast, opposite trends were observed in branches and stems from the two plantations (Fig. 3f).
Tree nutrient concentrations and their ratios were signi cantly affected by vegetation types, plant organs and their interaction (p < 0.05). In the non-karst area, N concentrations and C: N ratios were not signi cantly affected by their interaction (Table 4). In the karst area, N concentrations were not signi cantly affected by vegetation types (Table 4).

Nutrient resorption characteristics
The NRE varied from 30.25 to 50.20%, and the PRE varied from 27.44 to 45.79% (Fig. 4 a,b). NRE was higher in the karst area than the non-karst area in both plantations (p 0.05); there was no signi cant difference between the two plantations in the non-karst and karst areas (p 0.05). PRE was higher in the non-karst area than the karst area of both plantations (p 0.05).
The NRE was positively associated with the N concentrations and N: P ratios of green leaves (Fig. 5, 6). The PRE was positively associated with the P concentrations of green leaves (Fig. 7). The PRE was positively associated with the N: P ratios of green leaves in the non-karst area of the two plantations (Fig.  8 a,c), whereas it was negatively correlated with the N: P ratios of green leaves in the karst area of PY. No signi cant relationships were found between PRE and N: P ratios of green leaves in the karst area of EM ( Fig. 8 b,d).

Stoichiometric homeostasis characteristics of different tree species
To test the stoichiometric homeostasis, we determined the associations between the leaf N and P concentrations, N: P ratios, and the concentrations and ratios for the soil resources (Table 5). For N concentrations, leaves of the two plantations in the two areas were 1.341, 0.444, 0.792, 0.961, respectively, indicating that PY in the non-karst area showed the highest homeostasis. For P concentrations, leaves from the two plantations in the two areas were 0.770, 0.543, 0.687, and 0.587, respectively. These ndings indicate that the same plantation tree species in the non-karst area had higher homeostasis than that in the karst area; PY showed higher homeostasis than EM in the non-karst area. For N: P ratios, leaves of the two plantations in the two areas were 1.701, 0.813, 0.561, and 1.019, respectively, indicating that PY in non-karst area showed the highest homeostasis. Overall, PY had higher homeostasis in the non-karst area than EM, while in the karst showed the opposite trends.

Discussion
Differences between soil stoichiometry characteristics of the two stands Soil C, N, and P are in uenced by climates, vegetation types, soil substrates, and parent materials Wang et al. 2018;Qiu et al. 2020). The karst area showed signi cantly higher C, N, and P concentrations than the non-karst area for both plantations. The reasons for these differences are unclear; however, we offer three different explanations: First, soil C, N, and P concentrations in the two eld sites were associated with vegetation growth; in the non-karst area, which had minimal understory vegetation and large parts of the soil surface exposed, this resulted in high nutrient leaching and soil erosion Ore ce et al. 2017). Second, due to the OM-Ca 2+ -mineral complex in soils rich in Ca, higher exchangeable Ca concentrations result in greater soil C and N pools in the karst area (Li et al. 2017). Third, with a high understory vegetation coverage in the karst area, plant litter and organic matter inputs increase, facilitating microbial decomposition processes and resulting in nutrient accumulation (Pan et al. 2015;McIntosh et al. 2016). In our study, we also observed that soil C, N, and P concentrations were signi cantly lower in the PY plantation than the EM plantation in both the areas, which may result from the acidity of the foliage and litter of coniferous trees (Cao et al. 2018).
Soil C: N: P ratio is an important indicator of soil quality. It also re ects the variability of ecosystem functions and nutrient cycles (Cao and Chen 2017). In general, the lower C: N ratio indicates N mineralization dominant (Bui and Henderson 2013;Zhao et al. 2015). In this study, all the treatments had lower soil C: N ratios (Fig.2d), indicating that the subtropical forests had a higher nitrogen mineralization rate. As soil P in the non-karst area was obviously lower than the karst area, soil N: P ratios in the nonkarst area was signi cantly greater compared with the karst area. This is most likely a result of higher precipitation and soil moisture in the study areas, which led to increased soil weathering and a more rapid release of P from karst areas than non-karst areas (Houlton et al. 2008;Yang et al. 2018).
Differences between plant C: N: P stoichiometry of the two stands Previous studies have shown that the nutrient concentrations differed between different plant organs, probably as a result of varying genetic expression (Han et al. 2005). In this study, C, N, and P concentrations differed between plant organs and study sites. The concentrations of C in the organs of each tree species in the karst and non-karst areas were not signi cantly different (Fig. 3a). This because C provides the structural basis and plant skeleton Qiu et al. 2020). Our research also showed that the C concentrations in PY were higher than that in EM in all the plant organs, which in line with previous studies. Overall, these ndings suggests that conifers are rich in C compounds compared to primary and secondary forests and broad-leaved plantations (Yuan and Chen 2009;Fan et al. 2015). Plants growing in subtropical areas under high temperatures, strong sunlight, and heavy precipitation exhibit a high capacity for cell division and therefore require large amounts of protein, such as Rubisco, for photosynthesis and plant growth. The large amounts of Rubisco, the pivotal enzyme of the Calvin cycle, may explain the relatively high N concentrations in leaves compared to other plant organs (Gorokhova and Kyle 2002;Reich and Oleksyn 2004;Bloom eld et al. 2014). The P forms in leaves have been shown to be relatively variable, as plants can store not only organic P, but also inorganic P in leaves (Agren and Weih 2012a;Mayor et al. 2014). In our study, leaf P concentrations were found to be higher than the average leaf P concentrations in terrestrial ecosystems (Elser et al. 2008). One explanation for this might be that plants under high soil P conditions store higher concentrations of inorganic P in the cytoplasm and vacuoles of leaves (Mayor et al. 2014). Previous studies showed that high N or P concentrations correspond to faster growth rates and stronger competitive ability for resources (Shipley et al. 2006;Huang et al. 2019).
Leaf stoichiometry is an important parameter for studying the nutrient restriction, cycling, and response of plants to the climate change. We found that the C: N ratio in leaves ranged from 9.81 to 15.12 with an average of 12.54, and the C: N ratio in other plant organs ranged from 12.20 to 28.52 with an average of 19.94, indicating that the N concentration in leaves was lower than that of other organs. These ndings are inconsistent with previous study conducted in subtropical China (Luo et al. 2020). This might be explained by the fact that element concentrations can be maintained at relatively stable concentrations to adapt to environmental changes (Jeyasingh et al. 2009). Differences in photosynthetic capacity, plant morphology, and adaptation strategies might explain why the same elements and stoichiometry of plant organs differed between the two eld sites.

Leaf N and P concentrations and nutrient restriction
Generally, leaf C, N, and P concentrations and N: P ratios are good indicators of the plant-soil interactions and the nutrient content of the environment (Aerts 1996). N and P have a close relationship to the production of proteins, amino acids, and nucleic acids as well as the biosynthesis associated with plant issues Pan et al. 2015). The leaf N: P ratio greatly re ects nutrient limitations in plants, affecting plant traits, community composition, and biodiversity. N: P ratio of less than 14 indicates N limitation, whereas a ratio over 16 re ects P limitation, while ratio from 14 to 16 suggests N and P colimitation during plant growth and development (Koerselman and Meuleman 1996). In this study, the leaf N: P ranged from 17.12-26.62, and the average leaf N: P across all treatments was 20.01; therefore, under our experimental conditions the growth of the trees was limited by P. The average leaf N: P in PY of both areas was 22.17, while the average leaf N: P in EM was 17.85 in both areas, indicating that the growth of PY was more limited by P than EM. Which consistent with the "growth rate hypothesis" that faster growing plants would have lower leaf N: P ratios due to a high requirement of P-rich ribosomal RNA relative to N-rich proteins.
Differences in plant species and diversity result in different nutrient use e ciencies and competition for nutrients (Bing et al. 2016). With the climate uctuations and greater extremes of climate change, it is essential to understand how plants that are dominant in a fragile ecosystem respond to stoichiometric changes. Previous studies have indicated that plants that exhibit higher growth potential require more nutrients to ensure biomass production; therefore, the must have a higher NuRE (Pan et al. 2015;. The averages of NRE and PRE in the two plantations were 39.91% and 34.61%, respectively ( Fig. 4 a,b), which were lower than those in global terrestrial forests (47.4% to 62.1% and 53.6% to 64.9%, respectively) (Vergutz et al. 2012;You et al. 2018). The unexpectedly low NuRE may be explained by the high variation nutrient use strategies in subtropical forests. In our research, the NRE in the karst area was greater than that in the non-karst area for both plantations. These ndings indicate higher N de ciencies in the karst area; PY had higher N de ciencies than the EM in both study sites. Our research also showed that NRE was positively correlated with the N concentrations and N: P ratios of green leaves (Fig. 5, 6). The PRE in the karst area was lower compared with the non-karst area for both plantations. Overall, the plantations had higher P de ciencies in the non-karst area, and PY had higher P de ciencies than EM in both eld sites (Fig. 4b). The higher PRE might be a strategy for adapting to the P de cient soils of plantations in subtropical China, which is in agreement with the ndings of previous studies (Qiu et al. 2020;Tong et al. 2020). The PRE was positively correlated with the P concentrations of green leaves (Fig.  7), which is also consistent with previous studies Zhang et al. 2018). The higher PRE than NRE in the non-karst area suggested that the plantations were generally limited by P availability, supporting the "relative resorption hypothesis" that plants would have higher NuRE when growing under N or P limitation (Gusewell 2005). Thus, we assume that the high N and P resorption rate contribute to the moderation of the N and P de ciencies in the two areas.
Element resorption as an adaptive strategy to nutrient limitation NuRE serves as a signi cant indicator of plants to sustain adequate nutrition concentrations, thereby affecting plant growth, nutrient uptake, inter-speci c competition, biomass production, and net productivity (Killingbeck 1996;Liu et al. 2016). Previous studies have shown that tree species from infertile soils usually have low leaf nutrient concentrations and higher NuRE values; thus, they will adopt a "conservative consumption" nutrient use strategy to ensure their survival and reproduction (Kobe et al. 2005;Yan et al. 2006). In contrast, tree species from nutrient-rich locations usually adopting a "resource spending" nutrient use strategy to grow (Wright and Cannon 2001;Zeng et al. 2017). In our research, the lower concentrations of nutrients in soil and leaves coupled with the high NRE and PRE (Fig. 2,3,4) indicate that PY in the two areas adopted a "conservative consumption" nutrient use strategy, EM in both areas adopted a "resource spending" nutrient use strategy.
In our study, both tree species in the karst area showed an increased NRE compared to the same tree species in the non-karst area; hence, it is expected that plants in karst areas would require less N from the soil for new growth. In our research, the PRE in the karst area was lower than that in the non-karst area for both plantations. High P resorption rates may contribute to the moderation of P de ciency in the nonkarst area. The reasons for this are still unclear; however, we propose that in the karst area, relatively high soil pH together with rich CaCO 3 make the total Ca 2+ content high in soil, which leads to a large exchangeable calcium reservoir, forming an insoluble Ca-P phase. As a result, the CaCO 3 /AP (available phosphorus) ratio increases and plant-available phosphorus decreases in soil (Carreira et al. 2006;Ma et al. 2009;Zhao et al. 2012;Hong et al. 2014). In non-karst areas, plants possessing a relatively high growth rate need high P allocation rates and high metabolic rates to meet their high energy demands for the synthesis of macromolecules (Vitousek et al. 2010;Sun et al. 2016). However, some authors have suggested that higher NuRE does not serve as a signi cant mechanism for plants to adapt to low-fertility habitats and that phylogenetic and genetic difference take effect (Yan et al. 2008).
The PRE was lower in the karst area compared with the non-karst area, whereas NRE showed the opposite behavior. Overall, poor growth was observed for the two tree species in the karst area (Fig. 1), which has also been observed for Arabidopsis thaliana (Yan et al. 2015). In karst areas, trees must physiologically and morphologically adapt to the harsh conditions. As a result, the growth rate decreases and the use of nutrients becomes more conservative, which might contribute to a competitive edge for plants (Du et al. 2011;Wang et al. 2015). The difference in stoichiometric homeostasis (H) may re ect the trade-off of nutrient investment strategy in plants (Yu et al. 2011). The H N , H P , and H N:P among different trees species varied in our research, suggesting that different trees exhibited different adaptation strategies. The results of our study led us to infer that in fragile ecosystems, the reuse of N and P from the senescent leaves might serve as a special nutrient preservation strategy to improve plant growth, and NuRE in trees are exible and can be adapted to different environments.

Conclusions
Our study showed that soil C, N, and P concentrations were signi cantly higher in the karst area than in the non-karst area. Both tree species had lower soil C: N ratios, indicating that subtropical forests had a greater nitrogen mineralization rate. Plant stoichiometry and stoichiometric homeostasis were different between the two plantations within different sites. In addition, the N: P ratio and PRE showed that both sites were P de cient, and the growth of PY was more limited by P than EM. Furthermore, dominant tree species in subtropical areas adopted different nutrient use and adaptation strategies. Under the different geological backgrounds, vegetation restoration leads to great differences in plant parameters such as physiological, morphological, element concentrations, and adaptation strategies. Our ndings provide a scienti c basis for the development and optimization of restoration projects in harsh, fragile ecosystems. Figure 1 Morphological, nutrient limitation, and resorption e ciency difference between the two stands growing on different soil types.  The stoichiometric characteristics of C, N, and P measured in different plant organs. (a) C concentration, (b) N concentration, (c) P concentration, (d) C:N ratio, (e) C:P ratio, and (f) N:P ratio. The lowercase letter denotes a signi cant difference among each plant organ (p 0.05). Error bars represent standard error.

Figure 4
Changes in NRE (a), PRE (b) and NRE: PRE (c) among different plantations. The capital letter denotes a signi cant difference between the four plantations (p 0.05), and the lowercase letter denotes a signi cant difference among the same plantation in different areas (p 0.05). Error bars represent standard error.