4.1 Potential driving forces of WUE, NUE, and PUE
The N:P ratio is considered a reliable indicator of nutrient limitation (Güsewell et al., 2004), and a leaf N:P ratio (mass basis) less than 14, between 14 and 16, and greater than 16 indicates N limitation, N and P colimitation, and P limitation, respectively (Koerselman and Meuleman, 1996). In our study, according to this criterion, in monoculture and mixed stands, the growth of R. pseudoacacia with two native species was limited by P availability (Cao and Chen, 2017), and the growth of A. sibirica was limited by N availability (Figure 3d). In addition, the N:P ratio of R. pseudoacacia was significantly lower in the RPAS stand, whereas that of A. davidiana was significantly higher in the RPAD stand. Therefore, mixed stands decreased the P limitation in R. pseudoacacia and the N limitation in A. sibirica. In addition, the N:P ratio of R. pseudoacacia and A. sibirica was significantly positively correlated with Nmass and soil AP (Figure 7a, b; Table 6), probably because the processing of N-fixing of rhizobia promotes soil TP transformation to soil AP (Su and Shangguan, 2020).
The results show that the dispersion in multifactorial PCA within the same species depending on the community was lower in R. pseudoacacia than in the other two native non-N-fixing species (Figure 2), indicating that the N-fixing capacity of R. pseudoacacia confers it with a partial independence from other biotic or abiotic factors, giving it some competitive advantages over non-N-fixing species. Generally, the mixed stand structure directly affects the morphological, physiological, and composition features of its species by influencing the available resources through canopy closure and root distribution (Forrester et al., 2006; Dawud et al., 2016). In addition, compared with monoculture, the mixed-species plantation structure may improve plantation productivity by increased resource acquisition capability while reducing competition because of some level of niche differences (Danescu et al., 2016; Coll et al., 2018). Therefore, Narea and Parea were higher and DBH, AH, and SLA were lower for R. pseudoacacia growing in mixed stands than they were for the same species growing in monoculture stands (Table 3).
Mixed species with different root depth and density can exploit various underground niches, leading to more nutrient consumption (Dimitrakopoulos and Schmid, 2004; Stubbs and Wilson, 2004). In our study, AN and AP were lower for R. pseudoacacia growing in the RPAS stand than for the same species growing in RP and AS stands, while AN and AP of R. pseudoacacia growing in the RPAD stand were between those for the same species growing in RP and AD stands (Table 4). Therefore, R. pseudoacacia and A. sibirica in the RPAS stand showed more nutrient consumption and better nutrient utilization capacity when growing in the mixed community than when growing in their respective stands. In contrast, this was not observed in mixed stands of R. pseudoacacia and A. davidiana in the RPAD stand. Thus, the results strongly suggest that mixing R. pseudoacacia and A. sibirica is an effective strategy to maximize soil nutrient resources on the CLP.
4.2 WUE and correlations with its driving forces
4.2.1 Effect of mixed stand on WUE
Under water-deficit conditions, a high WUE was an important trait for plant survival and growth (Gong et al., 2011). In this study, the WUE values of R. pseudoacacia in the RP stand and A. sibirica in the AS stand were higher than those of A. davidiana in the AD stand (Figure 3a, Figure 4). Studies showed that stomatal conductance decreases with increasing water stress, resulting in an increase in WUE (Liu et al., 2015). Thus, R. pseudoacacia and A. sibirica show stronger competition and ecological adaptability than A. davidiana for water stress growing in monoculture stands under these environmental conditions.
Generally, the success of mixed stands is related to functional complementarity and mutual promotion relationships among multiple tree species (Forrester et al., 2005). In this study, such relationships were not detected with unchanged WUE in species in mixed stands in some cases (Figure 3a, Figure 4), which is in agreement with previous studies and consistent with the observation that plant WUE was associated with functional traits linked to genetic conditions (Meier et al., 2022), such as leaf photosynthetic rate, stomatal conductance, and life cycle (Garrish et al., 2010, Su and Shangguan, 2020). However, mixed afforestation increased WUE of R. pseudoacacia when this species was grown along with A. davidiana in the RPAD stand by decreasing stomatal transpiration in photosynthesis (Figure 3a, Figure 4; Katul et al., 2010). Specifically, mixed species increase leaf area and decrease leaf N content per unit area as the canopy gradually closes (Table 3), leading to a decreased light-saturated net photosynthetic rate (Katahata et al., 2007).
In this study, WUE of A. sibirica growing in the RPAS stand was lower than that of A. sibirica growing in the AS stand (Figure 3a, Figure 4). This result strongly suggests that A. sibirica in the RPAS stand was under lower drought stress than the same species growing in the monoculture stand. Therefore, A. sibirica in the RPAS stand decreased its WUE and increased its use efficiency of other resources, such as light and nutrients (Wu et al., 2017). In contrast, there was no significant change in WUE of R. pseudoacacia, whether it was grown in the RPAS stand or in the RP stand (Figure 3a, Figure 4). The mixed-species plantation is divided by hydrological niches, which decrease the competition among species for limited water resources and strengthen regional drought resistance (Lebourgeois et al., 2013; Pretzsch et al., 2013). Therefore, tree species can adopt different water use strategies in mixed stands compared with monoculture stands.
4.2.2 Influence of potential driving forces on WUE
In arid or semi-arid regions, the significant increase in WUE can be attributed to the decrease in stomatal conductance as temperature increased or soil moisture decreased (Jia et al., 2022). When soil water stores gradually deplete (Lei et al., 2022), R. pseudoacacia can improve WUE with tight stomatal control and limited transpiration rates (Fu et al., 2020). Therefore, R. pseudoacacia monoculture with higher SD tends to expend more soil water stores relative to the same monoculture with lower SD (Tanaka-Oda et al., 2010). In addition, the larger the SD, the smaller the CA of individual trees. This explains the observation that WUE of R. pseudoacacia was positively correlated with SD but negatively correlated with CA (Figure 7a, Table 6; Brookshire et al., 2020). Moreover, mixed stands with higher SD and greater canopy closure but smaller tree growth experience lesser photosynthetically active radiation than taller trees, leading to the low temperature inside the plantation (Khanna, 2008; Nygren and Leblanc, 2015). However, WUE of A. davidiana and A. sibirica in mixed stands was positively correlated with SD and negatively correlated with CA (Figure 7b, c), strongly suggesting that the stand temperature is a dominant factor controlling WUE of understory plants.
With the increase in soil nutrients, plants tend to increase leaf nutrient concentration, reducing nutrient use efficiency, which in turn increase WUE (Ripullone et al., 2004). Furthermore, WUE of A. davidiana and A. sibirica was positively correlated with soil TN and TP, respectively (Figure 7b, c), which may be related to plant physiology. For example, sufficient soil nutrients can stimulate plants to improve leaf photosynthesis and accumulate biomass, while their effect on evapotranspiration was not significant (Patterson et al., 1997). Thus, the increase in leaf N and P concentrations is generally associated with an enhancement in mesophyll conductance and photosynthetic capacity (i.e., A) (Wright et al., 2004). Similar results were obtained in Quercus velutina (Jennings et al., 2016) and Larix principis-rupprechtii (Yan et al., 2022). In contrast, R. pseudoacacia exhibits an opposite trend than expected because higher soil nutrients imply a lower, rather than a higher, WUE (Figure 7a). One plausible explanation is that N-fixing bacteria increase the N source of R. pseudoacacia, causing its fine roots to absorb more water to transport N. As the body stored more water, WUE of R. pseudoacacia was decreased and then showed negative correlations with soil TN. In addition, along with absorbing N compounds, the fine roots absorb a large amount of soil P. Therefore, WUE of R. pseudoacacia was negatively correlated with soil TP. Further, the decreased rate of soil TP was greater than that of soil TN, which may be another reason for the increased P limitation of R. pseudoacacia. All in all, as an N-fixing plant, R. pseudoacacia was not dependent on soil N sources, and thus, WUE may be directly related with its capacity to fix N.
4.3 NUE, PUE, and their correlations with driving forces
In monoculture stands, NUE of R. pseudoacacia was lower than that of A. davidiana and A. sibirica (Figure 3b), but AN of R. pseudoacacia was not the largest of the three (Table 4), indicating that AN was not necessarily correlated with NUE (Dijkstra et al., 2015). The possible reason was that the growth of R. pseudoacacia was at least partially related to atmospheric N fixed by symbiotic N-fixing bacteria (Turner and Lambert, 2014; Nygren and Leblanc, 2015) and thus partially independent of soil N. In addition, litter decomposition of R. pseudoacacia can continuously increase AN (Table 4) (Voigtlaender et al., 2012). Furthermore, NUE of these tree species was significantly negatively correlated with soil TP and AP (Table 6), and NUE of R. pseudoacacia was lower than that of A. davidiana and A. sibirica (Figure 3b, Figure 4). Therefore, monoculture or mixed stands, including R. pseudoacacia, had higher TP and AP concentrations (Table 4). However, the fact that AP was the lowest in the RPAS stand (Table 4) was probably associated with the observed high water demand for R. pseudoacacia and A. sibirica in the RPAS stand, promoting the transport of water and nutrients in soil and plant root uptake (Du et al., 2011; Sardans and Penuelas, 2012). Moreover, PUE of R. pseudoacacia was the lowest (Figure 3c) and its AP was the largest (Table 4) in monoculture stands, which were in agreement with previous studies that found plants have lower nutrient use efficiency when they grow in nutrient-rich soil (Vitousek, 2010; Turner and Lambert, 2014).
In both monoculture and mixed stands, these tree species maintained similar NUE and PUE (Figure 3b, c). Therefore, their nutrient use strategies are mostly influenced by inherent but not external factors (Deng et al., 2019). However, the relative increase in NUE of R. pseudoacacia in the RPAS stand was higher than that of R. pseudoacacia growing in the RPAD stand (Figure 3b, Figure 4). In addition, TN and AN values of this species were larger in the RPAS stand than in the RPAD stand (Table 4). Therefore, mixing R. pseudoacacia and A. sibirica should better increase soil N content of the RPAS stand. The reason may be that root growth in A. sibirica, due to its characteristics, can increase soil aeration, which indirectly increases and facilitates the symbiotic N-fixation of rhizobia. Otherwise, mixed afforestation did not alter PUE of R. pseudoacacia (Figure 3c, Figure 4), which was consistent with previous studies that found mixtures of N-fixing species, including Albizia falcataria and Eucalyptus saligna, had no effect on the processes of soil P transformation compared with their monocultures (Zou et al., 1995). In our study, PUE of A. davidiana and A. sibirica decreased after mixed afforestation (Figure 3c, Figure 4), whereas their AP was lower in both RPAD and RPAS stands than in RP and AD stands (Table 4). Hence, PUE of these tree species was significantly negatively correlated with their AP (Table 6), which was in line with previous studies that found having mixed-species plantations with the same growth cycle or leaf phenology usually leads to increased competition for resources (Feller et al., 1999; Amazonas et al., 2018). Moreover, in the short–medium term when these two species grow with R. pseudoacacia, they could dispose of more P by the effect of R. pseudoacacia allocating resources to increase P mineralization and leaching; thus, as P availability increases, PUE of A. davidiana and A. sibirica decreases when these species grow along with R. pseudoacacia. For example, de-Dios-García et al. (2018) showed that mixed forest stands formed only by conifers may show higher competition for resources than conifer-broadleaved admixtures because of more similar plant characteristics. Therefore, these tree species in the RPAD or RPAS stand adopt different nutrient use strategies.
4.4 Trade-off between WUE and NUE, PUE, and N:P ratio
Generally, WUE is negatively correlated with NUE and PUE, because a high WUE is possible only if the photosynthetic machinery and energy transfer capacity is high, which requires high N and P concentrations or low NUE and PUE (Zhao et al., 2015; Dijkstra et al., 2016). Moreover, as an indirect effect, inefficient WUE can be associated with higher transpiration fluxes, favoring P and overall N (more soluble) uptake, as WUE was inversely related with NUE and PUE, that is, lower WUE was associated with higher NUE and PUE. In this study, we observed a negative relationship between WUE and NUE (Figure 6a), which is in line with previous studies (Gong et al., 2011; Dijkstra et al., 2016) that found the relationship was consistent with a direct trade-off effect, and the correlation coefficients (R2) in the increasing order were as follows: 0.55 for R. pseudoacacia, 0.80 for A. davidiana, and 0.95 for A. sibirica. This effect could be due to changes in physiological traits (such as leaf water potential and gas exchange parameter) and study sites with a heterogeneous environment (Yan et al., 2016). For example, air temperature and humidity can affect water depletion of plants, which in turn affects their evapotranspiration and ultimately leads to different WUE (Garrish et al., 2010). Furthermore, the processing of symbiotic N-fixation can alter root ability in absorbing water and utilizing N (Su and Shangguan, 2020).
N and P are two important limiting resources for plant growth, and both significantly affect WUE (Huang et al., 2015). However, few studies have confirmed a stable trade-off between WUE and PUE (Brown et al., 2011, Huang et al., 2015). This is because N and overall P uptake depend on variables other than soil water movement and transpiration, such as soil enzyme activities, root system form and size, and root exudates. However, for N-fixing species, water uptake and N and P uptake capacities are more disconnected, as these plants can fix N independently after soil uptake and can mobilize more soil P because of their higher capacity of promoting phosphatase synthesis in the plant–soil system by a better supply of N and thus better protein production capacity. This strongly explains why PUE of R. pseudoacacia was not correlated with its WUE (Figure 6b). A similar result was reported for Pinus pinaster in southwest Western Australia (Warren et al., 2005). However, WUE of A. davidiana and A. sibirica was significantly negatively correlated with PUE (Figure 6b), which is in agreement with a previous study (Zhou et al., 2016). This result was consistent with the high water demand of these species in mixed stands along with their higher capacity for P uptake. For example, these species have a greater capacity of phosphatase synthesis by a higher availability of N and thus of protein synthesis because of the N-fixing capacity, which altogether promote the available nutrients (e.g., AP) to the root by mass flow and reduced availability of soil nutrients in the rhizosphere (Cregger et al., 2014).
WUE of R. pseudoacacia was not positively correlated with the N:P ratio, as other studies found in the Qinghai–Tibet Plateau and the Xilingol grassland (Zhou et al., 2016). However, WUE of A. davidiana was negatively correlated with the N:P ratio (Figure 6c). Leaf P concentration was positively correlated with the photosynthetic rate (Güsewell, 2004). When the leaf N:P ratio was greater than 12, plant WUE negatively correlated with the photosynthetic rate (Garrish et al., 2010). Therefore, plant WUE is affected by the combined action of leaf N and P (Huang et al., 2015). In this study, WUE of A. sibirica was significantly positively correlated with the leaf N:P ratio (Figure 6c), which was consistent with a previous report in shrubs and trees (Yan et al., 2016), and it was also closely linked to the positive correlations between plant N:P stoichiometry and transpiration rate (Cernusak et al., 2007), thereby integrating the nutrient and hydrological cycles (Dijkstra et al., 2016; Su and Shangguan, 2020). Therefore, this relationship may be affected by multiple factors such as genetic traits and nutrient conditions.