Monthly variation of nutrients in the leaves of M. sieversii with different DBRs
The seasonal dynamics of leaf nutrient concentrations and REs, which are the key ecological indicators of the strategies of nutrient utilization in vegetation ecosystems, are one of the best diagnostic tools for determining the nutrient status of plants and can provide a good guide for proper fertilization (Chatzissavvidis et al. 2005; Liu et al. 2015). Plant nutrient stoichiometry and nutrient resorption during stand development are essential indicators for assessing forest degradation (Imaran and Gurmani 2011; Wang et al. 2015; Yan et al. 2017). The observed changes in leaf nutrient concentrations of deciduous plants associated with different developmental stages mostly present similar variation patterns. The leaf nutrient concentrations are relatively high at the early growth stage because the leaves are small. During summer, trees grow fast, the leaf area and weight increase rapidly, the rate of nutrient uptake by roots cannot keep up, and thereby the leaf nutrient concentrations decrease. During leaf senescence, part of the nutrient matter returns to the plant body, leading to nutrient reduction in leaves (Aerts 1996; Teklay 2004; Lu et al. 2018).
In this study, the similar seasonal change pattern of leaf nutrients was confirmed for wild apple trees with different DBRs, and leaf N:P, N:K, and P:K represented differences. These markedly monthly changes led to relatively high CV (Table S1). Previous studies have indicated that at the early developmental period (i.e., April for wild apple trees), small leaves and rapid cell division are observed. Thus, the protein and nucleic acid needed by the cell are high, leading to the high N and P contents in leaves (Smillie and Krotkov 2011). Moreover, the P and K contents in flowers are higher than that in leaves in the same period (Fig. 4), revealing the rapid growth rate of the reproductive organ and supporting the growth–rate–hypothesis (Elser 2000; Tian et al. 2018a). In addition, the seasonal changes of leaf nutrients vary with functional groups, reflecting the functional differentiation of nutrient utilization (Teklay 2004; Liu et al. 2015). The leaf N and P concentrations in evergreen plants present slightly seasonal fluctuations because of the relatively low intensity of physiological metabolism in these evergreen leaves (Tian et al. 2018a, 2019). By contrast, the physiological and metabolic activities and leaf nutrient parameters in the leaves of deciduous plants represent markedly seasonal change because of the differences in sunshine and temperature in different seasons in the temperate zone (He et al. 2019). In the present study, when the leaf nutrient data in April were removed, the variability in the leaf N, P, and N:P was evidently reduced (Table S1). Reports have shown that plants under a high-temperature environment have relatively low leaf N and P concentrations because a high temperature can enhance enzyme activity and photosynthetic efficiency in leaves, thereby accelerating the nutrient cycle and resulting in low N and P content (Reich and Oleksyn 2004; Han et al. 2005; He et al. 2008; Cornelissen et al. 2014). For wild apple trees, the temperature peaked in July and August, and the leaf N concentration reduced significantly, thereby resulting in the decline of leaf N:P.
Consequently, the leaf N, P, and K concentrations of M. sieversii decreased during growth, but their stoichiometric ratios showed an inconsistent change trend. The leaf K, N:K, and P:K of HD wild apples were different with those of LD and MD, but they showed the same change trend among the DBRs during the growth period.
Leaf nutrient status of M. sieversii with different DBRs
Most studies on nutrient concentrations and stoichiometric ratios are based on the mature leaves. Thus, the nutrient parameters in July were chosen for comparison. The mean leaf N concentration of wild apple trees in July (14.92 mg g− 1) was far lower than those of the global average (18.9 mg g− 1) in terrestrial vegetation, the global woody plants (18.22 mg g− 1) (Tian et al. 2018a), the Chinese terrestrial plants (20.24 mg g− 1) (Han et al. 2005), and the forest plants in NSTEC (18.30 mg g− 1) (Ren et al. 2007). The mean leaf P concentration of M. sieversii in July (1.90 mg g− 1) was higher than those of the global average leaf P concentration (1.77 and 1.20 mg g− 1) in the terrestrial vegetation, the global woody plants (1.10 mg g− 1) (Tian et al. 2018a), and the Chinese terrestrial plants (1.46 mg g− 1) (Han et al. 2005), but was slightly lower than that of forest plants in the NSTEC in China (2.00 mg g− 1) (Ren et al. 2007).
N and P are the major nutrients constraining plant growth worldwide. A leaf N:P of < 14 and > 16 indicate N and P limitations, respectively (Koerselman and Meuleman 1996). Hence, the leaf N:P threshold was considered as a useful index to determine the nutrient limitation in individual plants and ecosystems (Han et al. 2005; Ren et al. 2007; Sardans et al. 2012). The N:P threshold considerably varies with study area, ecosystem type, community type, functional group, and plant species (Yan et al. 2017), but low and high N:P correspond to N and P limitations, respectively (Wu et al. 2012; Yan et al. 2017). The leaf N:P ratios (5.48–9.67; 7.92 in July) of wild apple trees in all growth periods were far lower than 14, indicating N limitation. The leaf P concentrations were higher than 1.0 mg g− 1 (i.e., the lower limit indicating P limitation) (Wu et al. 2012; Tao et al. 2016; Yan et al. 2017). Thus, the low N:P ratio and relatively high P content revealed that the growth of M. sieversii was strongly limited by N.
Leaf K was the dominant factor influencing leaf N:K and P:K, and leaf K contents significantly varied among different DBRs. Therefore, this finding confirmed the close relationship among the K content, plant death, and population degeneration of wild apples. The leaf K content is closely related to the growth, drought resistance, and disease resistance of plants (Munson et al. 1985; Sangwan et al. 2008). The decrease in K content indicated that the resistance of the diseased plants was seriously inhibited; K was also related to the reproduction of plants (Liu et al. 2006). In this study, only 6/10 individuals of HD flowered, but no individual had a fruit in autumn, which might indicate that the lack of K had a strong limiting effect on the reproduction of wild apples. In general, plants with high soluble N content and low protein–N content are susceptible to disease (Duan and Fang 2017). The syntheses of protein and numerous carbohydrates (carbohydrate, cellulose, pectin, lignin, and starch) are completed by the enzymatic action of K. K promotes protein synthesis and reduces the content of soluble amino acids, thereby improving disease resistance (Munson et al. 1985; Liu et al. 2006). Therefore, the lack of K reduces the disease resistance of plants. The soil total and available N, P, and K contents of HD wild apples were higher than those of MD wild apples, indicating that the soils of HD did not lack K (Fig. S2). Our study indicated that the leaf K content of HD in April was significantly lower than those of LD and MD, and the flower K of HD in April and leaf K of MD (HD had no fruit and was not used here) in August were also significantly lower than LD (Fig. 4). Thus, the low leaf K of HD may be caused by the low absorption and utilization capacities of wild apples, that is, the high DBR inhibited the absorption of K and storage of wild apples. This phenomenon may lead to a positive feedback and further reduce disease resistance (Cui et al. 2019). In particular, the temperature in the study site dramatically changed in early and middle May, approaching 0 °C twice (Fig. 2). At the same time, the leaf K content of LD increased on May 20, indicating that wild apple trees may improve the K content to enhance the resistance of normal individuals (i.e., with few dead branches). The leaf K content of MD also increased slightly but that of HD did not. The coefficient of linear regression between leaf K and month of HD (R2 = 0.960) was higher than those of MD (R2 = 0.822) and LD (R2 = 0.772), proving that HD wild apples had almost no self-regulation ability to environmental change (Fig. S6). This result demonstrated again that an increase in dead branches can reduce K absorption and resistance. In addition, Zhang et al. (2020) reported that the long-term water deficit caused by A. mali infestation may be the key factor leading to the decline of wild apple forests. Consequently, the death of branches of M. sieversii can reduce both K and water use efficiencies at the same time.
Stoichiometric relations of M. sieversii with different DBRs
The present study indicated that leaf N:P was sensitive to the changes in P concentration (rather than N). Similarly, leaf N:K and P:K were more sensitive to the changes in K and P, respectively. Thus, the leaf stoichiometric ratios were influenced by P and K. Reports have indicated that the allometric scaling exponent between leaf N and P was not correlated with leaf N content (Tian et al. 2018a, b). However, this exponent showed a significantly negative relationship with leaf P, that is, the change in leaf P drove the allometric and stoichiometric relationships between leaf N and P, thereby confirming that leaf P was the main factor influencing leaf N:P in this study (Tian et al. 2018a, b). However, the significant relationship between N:P and P did not suggest P limitation for M. sieversii but suggested N limitation because of low N:P.
In the present study, the difference among leaf N–P, N–K, or P–K scaling exponents for three DBRs was not significant, and a common slope (0.846, 0.884, or 0.952) existed. However, the intercepts for the three DBRs showed considerable difference, reflecting differences in actual elemental concentrations. For N and P, the leaf N–P scaling exponents for three DBRs (0.835, 0.913, and 0.775, respectively) were higher than the 2/3-power law of major plant groups and biomes (including 2500 species across dramatic biogeographic gradients) reported by Reich et al. (2010) and that (0.676) of global green leaves (Yuan et al. 2012), which posit that plants with high growth rates require high allocation of P-rich ribonucleic acid and a high metabolic rate to support the energy demands of macromolecular synthesis (Reich et al. 2010). In fact, the allometric scaling slopes generally differed with functional groups, families, study sites, or regions in the range of 0.366–1.928 (Tian et al. 2018a). In the present study, the slopes of LD and MD were relatively higher than HD (P > 0.05), indicating that the P allocation or accumulation rate of HD was relatively higher than that of individuals with low DBR, showing a relatively high metabolic rate of HD individuals. Reports showed that there was no significant genetic divergence between the dead and living individuals, which indicated that the death of M. sieversii was random and out of genetic correlation (Li and Zhang 2018). And the death of branches of wild apple trees also did not change the salicylic acid content in twigs, which can inhibit pathogenic microorganisms (Sangwan et al. 2008). Therefore, the same scaling slopes between different DBRs and the consistent interrelations between three elements and the stoichiometric ratios may reveal the inherent nutrient allocation strategy of M. sieversii.
Leaf nutrient resorption of M. sieversii with different DBRs
The nutrient resorption from senescing leaves is directly available for further plant growth and a major nutrient conservation mechanism, which makes plants less dependent on soil nutrients (Vergutz et al. 2012; Yan et al. 2015; Lu et al. 2018). Therefore, leaf nutrient resorption has important implications for nutrient cycling. Most studies have found that compared with the species in nutrient-rich soils, the species in nutrient-poor soils have higher nutrient RE, which is important to plant growth in nutrient-poor environments (Killingbeck 1996; Lu et al. 2018; Zhang et al. 2019).
Our findings suggested that the order of nutrient RE of leaf N, P, and K in M. sieversii is K > N ≥ P, showing relatively low NRE and PRE. In general, the RE of trees is higher than that of shrubs, and the RE of angiosperms is higher than that of gymnosperms (Lu et al. 2018; Drenovsky et al. 2019; Zhang et al. 2019). A previous study has reported that the mean NRE and PRE from senescing leaves in global deciduous shrubs and trees are 54.0% and 50.4%, respectively (Aerts 1996). Thus, approximately half of N and P pools in leaves can return to other tissues through nutrient resorption. Another global study on terrestrial plants has shown that the NRE, PRE, and KRE are 62.1%, 64.9%, and 70.1%, respectively (Vergutz et al. 2012). Thus, more than half of the nutrients in leaves can return to the plant body. The NRE, PRE, and KRE of woody plants in northern China are 56.8%, 63.6%, and 50.7%, respectively (Zhang et al. 2019). The PNE of Vaccinium uliginosum that grows at the tree line on Changbai Mountain in China has reached 74.84% (Liu et al.2015). However, in the present study, the mean NNE and PNE of M. sieversii of three DBRs were 30.3% and 27.9%, respectively, and were far below the mean RE in numerous studies (Lu et al. 2018; Zhang et al. 2019). Thus, the N and P conservation capacity and utilization efficiency of M. sieversii were low, indicating that wild apple trees were more limited by soil N and P. On the contrary, the high KRE (48.87–57.47%) of M. sieversii indicated that wild apple trees were less limited by soil K. KRE increased with increasing DBR (P > 0.05), revealing that a high DBR resulted in high need for K. In general, the soil nutrient availability, temperature, precipitation, aridity index, and leaf life can influence the RE of leaf nutrient (Lu et al. 2018; Drenovsky et al. 2019). However, our findings showed that the death of branches and population degradation can affect the nutrient resorption of wild apple trees to a certain extent.
Influencing factors on leaf nutrient stoichiometry of M. sieversii with different DBRs
The nutrient content of plants can be affected by many factors (Vergutz et al. 2012). In the present study, the study site (with an area of 0.24 hm2) was unique. Although the soil background can be considered homogeneous, most soil variables were different among the three DBRs (Fig. S4 and S5). Despite this condition, most soil variables did not represent significant correlations with leaf nutrient parameters. Leaf K showed significant negative correlations with most soil variables. This phenomenon was probably related to poisonous and harmful weeds (Yan and Xu 2010). The wild fruit forests in Xinyuan County in the Western Tianshan Mountains were seriously degraded due to diseases, insect pests, and overgrazing (Zhang et al. 2020). As such, grazing prohibition and medicine sprinkling (to eliminate A. mali) were implemented. However, the main negative effect of grazing prohibition is the overgrowth of the poisonous and harmful weeds, such as Urtica cannabina and Urtica dioica (Fig. S7). The life cycles of these poisonous and harmful weeds almost coincide with that of wild apple trees. Thus, these weeds compete with wild apple trees for nutrients, which can increase the limitation of nutrient absorption of diseased plants especially the K absorption of HD. The growth of weeds can result in abundant litters and increase soil organic matters. However, it is likely to inhibit the K absorption of wild apple trees (Table 4), thereby reducing the plant resistance and further accelerating the decline of wild apple trees. Proper grazing or weeding during the growing season of weeds reduce their abundance and biomass and may slow down the limitation of K absorption and utilization by severely degraded M. sieversii.
The difference in temperature and precipitation in each month may also be an important reason for the difference in leaf nutrients especially in the arid area, where the release and transfer of soil available nutrients are considerably affected by soil moisture (Niu et al. 2013). At large spatial scales, leaf N and P concentrations decrease with increasing mean annual temperature and mean annual precipitation (Elser et al. 2007; Ren et al. 2007). This study partially verified the abovementioned conclusion, that is, leaf N and P decreased with increasing temperature, whereas leaf N and P increased with increasing precipitation and RH. According to the temperature–plant physiology hypothesis, the increase in N and P content in leaves can compensate for the decrease in metabolic rate at low temperature. Therefore, with increasing temperature, the N and P contents in leaves tended to decrease (Reich and Oleksyn 2004; Sun et al. 2019). In the tropics and subtropics, the high precipitation can enhance N and P leaching and reduce the availability of soil nutrients, thereby resulting in the decrease in N and P contents in the leaves of evergreen woody plants (Reich and Oleksyn 2004).
However, this study is in contrast to the large-scale results of precipitation, indicating that the plants in the arid region of Central Asia are still subject to drought stress (Tao et al. 2016). Precipitation is concentrated in spring in wild fruit forests (Fig. 2), which is the early stage of the germination and flowering of wild apple trees, and the nutrient concentrations in the leaves are high. In the late stage of plant growth, the nutrient concentration in the leaves decreases, and precipitation is also reduced. This phenomenon may cause the trend of increasing leaf N and P with increasing MTP. In the arid areas of Central Asia and in the Tianshan Mountains, the lack of precipitation (less than 600 mm) is the primary limiting factor of vegetation growth (Zhang 1973; Yan and Xu 2010). Thus, the increase in leaf N and P contents in leaves with increasing precipitation is the actual demand of plant growth. It has been confirmed that in a drought year (2016), the wild apple trees with high proportion of damaged branches (51–75%) showed significant lower leaf K content than that with low proportion of damaged branches (11–25%); however, in a wet year (2017), no significant difference existed (Zhang et al. 2020). Meanwhile, the RH and MTP had the same trend, showing positive correlation with leaf N, P, and K. This result indicated that relatively humid air was conducive to the absorption and accumulation of plant nutrients. Therefore, artificial rainfall should be used in the dry summer to increase precipitation, air humidity, and soil moisture and is conducive to the growth, nutrient absorption, and accumulation of wild apple trees.