The study area is located in Yanbian County, Panzhihua City, Sichuan Province, China (101.52°∼101.53°E, 26.95°∼26.96°N), within the dry-hot river valley at the junction of Sichuan and Yunnan provinces ( Figure 1 ). This region experiences a typical South Asian tropical dry-hot valley climate, with an annual average temperature was 19.2°C. The three main soil types are bauxite, red soil, and yellow-red soil.

The predominant variety of M. alba cultivated in this region is Yunsang No. 2, which is used for both fruit and leaf production. In the study area, M. alba is primarily found at altitudes between 1,000 and 2,000 m asl, with most trees concentrated in the mulberry forest at 1,200–1,500 m asl. At altitudes of 1,000∼1,200 and 1,500∼2,000 m asl, M. alba was observed to be scattered along roadsides, ditches, and near houses.

All samples were collected in June 2023 from eight sampling sites in mulberry orchards that had been established for more than 3 years, at altitudes ranging from 1,200 to 1,500 m asl. Sampling was conducted on both sunny and shady slopes ( Figure 1 ). The mulberry trees had trunk diameters of 6–10 cm, with row spacing of 1.5–2 and plant spacing of 0.6–1 m.

At each sampling site, three 10 m² × 10 m² sampling plots were randomly selected. Withon each plot, three 1 m² × 1 m² quadrats were placed along the diagonal (at both ends and the midpoint). In each quadrat, five soil samples were collected from the center and corners at a depth of 0 to 25 cm using a soil drill, then combined into a single composite sample. Similarly, healthy M. alba leaf samples were collected from each quadrat.

M. alba leaves were oven-dried at 70°C for 48 h and then powdered. Soil samples were air-dried and subsequently ground. The ground samples were used to determine organic carbon content, total nitrogen content, and total phosphorus content using the potassium dichromate oxidation external heating method, the micro-Kjeldahl method, and the ammonium molybdate method, respectively. Measurement results are expressed as nutrient content per unit mass (g kg⁻¹).

Statistical analyses were performed using R (version 4.4.1) and SPSS 22.0 (SPSS Inc, Chicago, USA). Variance decomposition was conducted to assess the relative effects of altitude, slope, and their interactions using the “vegan” package in R. Nutrient levels and stoichiometric ratios between sunny and shady slopes were compared using a t-test. To identify trends in nutrient contents and stoichiometric ratios at different altitudes, linear curve fitting was applied to examine the link between C, N, and P contents, stoichiometric ratios, and altitude. Spearman correlation analysis was used to assess the relationships between the stoichiometric ratios and C, N, and P contents.

The distribution ranges of difference values (D-values) for nutrient contents and stoichiometric ratios between M. alba leaves and soil were analyzed to determine if M. alba leaves maintained synchrony with soil. Firstly, the nutrient concentrations and stoichiometric ratios of mulberry leaves and soil were standardized as follows:

Subsequently, the difference between the two standardized values was calculated (e.g., leaf C − soil C). If the distribution of D-values remained within the 95% confidence interval of its mean, it indicated that changes in leaves were synchronized with those in the soil. Conversely, if the distribution exceeded the 95% confidence interval, it was assumed that leaf changes were not synchronized with soil changes.

The strength of plant stoichiometric homeostasis was analyzed on a log–log scale using the model: log(y) = log(c) + (1/H)log(x), where y represents the content of C, N, or P, or the ratios of C:N, C:P, or N:P in leaves; x represents the content of C, N, or P, or the ratios of C:N, C:P, or N:P in soil; and c is a constant. Values for H and c are determined through regression analysis of the relationship between y and x. The value of 1/H was derived from the regression slope between log x and log y, ranging from 0.00 to 1.00. To assess stoichiometric homeostasis, one-tailed tests with α = 0.10 were conducted. If the regression relationship was not significant (p > 0.10), the plant was classified as “strictly homeostatic”. However, if the regression relationship was significant (p < 0.10), stoichiometric homeostasis was categorized into four levels: homeostatic (0 < 1/H < 0.25), weakly homeostatic (0.25 < 1/H < 0.5), weakly plastic (0.5 < 1/H < 0.75), and plastic (1/H > 0.75) ( Figure 2 ).

All figures in this paper were created using Origin 2024 (OriginLab Corporation, Northampton, MA, USA).

The soil had significantly lower C, N, and P contents, as well as C:N, C:P, and N:P ratios, compared to M. alba leaves (p < 0.05). Moreover, the coefficients of variation for soil C, N, and P contents, as well as C:N, C:P, and N:P ratios, were higher than those for M. alba leaves ( Table 1 ).

Slope significantly influenced soil C, N, C:N, C:P, and N:P (p < 0.01), whereas altitude had a significant effect on soil P (p < 0.01). Additionally, soil C:P was influenced by both slope and altitude. Variance decomposition analysis indicated that slope was the primary factor influencing soil C, N, C:N, C:P, and N:P. For soil C, N, and N:P, the order of influence magnitude was slope > interaction effects > altitude. For soil C:N and C:P, the order of influence magnitude was slope > altitude > interaction effects. The dominant factor affecting soil P was slope ( Figure 3 ).

The nutrient content and stoichiometric ratios of the soil differed significantly between the shady and sunny slopes (p < 0.05). The shady slope had a higher concentration of carbon (15.47 g·kg⁻¹) than the sunny slope (5.93 g·kg⁻¹). Similarly, nitrogen content was significantly higher on the shady slope (2.10 g·kg⁻¹) compared to the sunny slope (0.69 g·kg⁻¹). In contrast, the difference in phosphorus content between the two slopes was not significant (p > 0.05), with the shady slope having a slightly higher concentration (0.69 g·kg⁻¹) than the sunny slope (0.68 g·kg⁻¹). The C:P and N:P ratios were significantly higher on the shady slope than on the sunny slope (p < 0.05), whereas the C:N ratio was lower (p < 0.05) ( Figure 4a ).

However, no significant difference was found in the stoichiometric ratios or the C, N, and P contents of M. alba leaves between the shady and sunny slopes (p > 0.05) ( Figure 4b ).

The response of nutrient contents and stoichiometric ratios in soil and M. alba leaves to altitude varied significantly between sunny and shady slopes ( Table 2 ).

On the sunny slope, only soil C showed a significant linear increase with altitude (p < 0.05). In contrast, on the shady slope, soil P decreased linearly with altitude, while soil C:P and N:P increased (p < 0.05).

For M. alba, leaf C scaled linearly and positively with altitude on the sunny slope (slope = 0.0186, R ² = 0.66, p < 0.05) but negatively on the shady slope (slope = − 0.0146, R ² = 0.69, p < 0.01). Leaf N, P, and N:P decreased linearly with increasing altitude on the shady slope (p < 0.05) but showed no significant change on the sunny slope (p > 0.05). Similarly, leaf C:N increased linearly with altitude on the shady slope (p < 0.05) but remained relatively constant on the sunny slope (p > 0.05). In contrast, leaf C:P showed a positive linear relationship with altitude on the sunny slope (slope = 0.0599, R ² = 00.57, p < 0.01), but no significant linear relationship was found on the shady slope (p > 0.05).

A positive correlation between was found between soil C and N (r = 0.98, p < 0.05). Additionally, both soil C and N were positively correlated with soil C:P and N:P but negatively correlated with soil C:N (p < 0.05). However, the correlations among leaf C, N, and P and leaf stoichiometric ratios were all found to be insignificant (p > 0.05). Leaf N was negatively correlated with leaf C:N, but positively correlated with leaf N:P (p < 0.05). Similarly, leaf P showed a negative correlation with leaf C:P but a positive correlation with N:P (p < 0.05). Furthermore, no significant correlation was found between soil and leaf (p > 0.05) ( Figure 5 ).

On both shady and sunny slopes, the D-values in nutrient contents and stoichiometric ratios between M. alba leaves and soil did not stabilize within a specific range (95% confidence interval of the means). These results suggested that the nutrient contents and stoichiometric ratios of M. alba leaves did not correspond to changes in soil ( Figure 6 ).

In this study, the regression relationships between M. alba leaves and soil on both shady and sunny slopes were not significant (p > 0.10). Therefore, M. alba should be classified as strictly homeostatic ( Table 3 ).

The dry-hot valley has traditionally been regarded as a nutrient-poor and water-scarce environment, unsuitable for plant growth (Ya et al., 2004; Duan et al., 2017). However, in this study area, the average soil organic carbon, total soil nitrogen, and total phosphorus were 10.7, 1.4, and 0.68g·kg⁻¹, respectively. These values indicate that the soil nutrient level in the study area is at a medium level compared to the national average (Fangyan et al., 2007; Tian et al., 2010; Zi-qi et al., 2022). Further analysis revealed that the soil nutrient level on the sunny slope was lower, consistent with most studies on dry-hot valleys (Guo, 2014; Guan, 2022; Hang et al., 2022; Yiting et al., 2023), whereas the nutrient level on the shady slope was at a medium level. These findings highlight the importance of distinguishing between the two slopes in future research.

Soil stoichiometry is a useful tool for understanding the cycling of elements in soil (Klemmedson and Wienhold, 1992; Zechmeister-Boltenstern et al., 2015). Soil C:N and C:P ratios reflect the rates of soil organic matter (SOM) decomposition, nutrient mineralization or immobilization, and plant nutrient limitation. A lower C:N ratio indicates a faster SOM mineralization rate, with the cumulative SOM rate being lower than the decomposition rate (Enwezor, 1976; Bui and Henderson, 2013; Pan et al., 2024). Although less frequently used, the soil C:P ratio also serves as a useful indicator of the source/nature of organic matter, with a lower C:P suggesting a higher SOM mineralization rate (Bui and Henderson, 2013). In this study, the C:N ratio in the dry-hot valley (8.33) was lower than the Chinese national average (11.9) (Tian et al., 2010; Pan et al., 2024) and findings from other studies on dry-hot valleys (Changming et al., 2022; Yiting et al., 2023; Zhifeng et al., 2024). Similarly, the C:P ratio in the study area (16.21) was significantly lower than the Chinese national average (61) (Tian et al., 2010; Pan et al., 2024) and findings from other studies on dry-hot valleys (Changming et al., 2022; Yiting et al., 2023; Zhifeng et al., 2024). These results suggest that SOM mineralization in the study area was higher than the Chinese national average and other reported dry-hot valley studies (Tian et al., 2010; Bui and Henderson, 2013).

Nitrogen and phosphorus are essential nutrients for plant growth and often serve as limiting factors in terrestrial ecosystems (Reich and Oleksyn, 2004; Yanan et al., 2014; Zhang et al., 2018). The N:P ratio in plant tissues is a widely used indicator of nutrient limitation, with values > 16 suggesting phosphorus limitation and values < 14 indicating nitrogen limitation (Güsewell, 2004; Han et al., 2005; Bertrand et al., 2019). In this study, no significant difference was observed in the N:P ratio of M. alba leaves between the sunny and shady slopes in the dry-hot valley, with an average of 10.28 on the sunny slope and 11.10 on the shady slope—both below 14. These findings suggest that M. alba growth is limited by nitrogen. The C:N and C:P ratios of plant leaves reflect nitrogen and phosphorus use efficiency as well as the plant’s capacity for carbon fixation (Sun et al., 2019; Pan et al., 2024). In this study, the C:N and C:P ratios of M. alba leaves in the dry-hot valley were 13.65 and 146.04, respectively, which were lower than the value reported for global (37.1, 469.2) and Chinese forest ecosystems (28.5, 513.0) (Han et al., 2005). This indicates that the plant’s nitrogen and phosphorus utilization efficiency, as well as its carbon fixation capacity, are lower in this study area compared to broader global and Chinese ecosystems (Sun et al., 2019). Additionally, reducing the C:N and C:P ratios is a known adaptive strategy that plants use to cope with resource-limited environments (Elser et al., 1996, 2000). Some species in dry-hot valleys have been observed to adopt this strategy (Lin et al., 2019; Jingwen et al., 2020; Hang et al., 2022), and M. alba appears to employ a similar mechanism to enhance its growth under these conditions.

Slope aspect significantly influences soil nutrient levels by regulating water and energy availability and energy input, thereby shaping local abiotic and biotic environments (Bale et al., 1998; Yang et al., 2020; Qin et al., 2021). This is consistent with the findings of this study, where the primary effects on soil organic carbon, soil nitrogen, C:N, C:P, and N:P were attributed to slope differences. Previous studies have also shown that shady slopes, characterized by lower temperatures, reduced solar radiation, smaller temperature fluctuations, and higher topsoil water retention, promote organic matter accumulation, leading to more fertile compared to sunny slopes (Sharma et al., 2010; Zhaoyang et al., 2019; Tan et al., 2020; Yang et al., 2020; Liu et al., 2024). Research conducted in the dry-hot valleys of the Jinsha River (Chang-ming et al., 2024) and Minjiang River (Nan, 2016; Yang et al., 2020) similarly indicates that soil fertility is higher on shady slopes, which is consistent with the results of this study.

In mountainous areas, altitude is the primary factor driving spatial heterogeneity. Soil nutrient levels and stoichiometry vary significantly between higher and lower altitudes due to differences in climatic conditions, precipitation patterns, vegetation, and microbiome composition (Jeyakumar et al., 2020; Chang-ming et al., 2023). In this study, altitude was identified as the main factor influencing soil phosphorus content and the C:P ratio. Numerous studies have demonstrated that in dry-hot valleys, increasing altitude results in reduced dry-hot winds and higher precipitation, leading to a steady accumulation of soil organic carbon, nitrogen, and phosphorus, and ultimately enhancing soil fertility (Chunming et al., 2003; Mullen, 2011; Peng et al., 2011). However, previous studies have not thoroughly examined how soil nutrient changes with altitude differ between sunny and shady slopes. In this study, we found that soil nutrient levels responded differently to altitude depending on slope aspect. On sunny slopes, soil organic carbon exhibited a significant increasing trend with altitude, aligning with findings from other research in dry-hot valleys (Chunming et al., 2003; Mullen, 2011; Peng et al., 2011), In contrast, on shady slopes, soli phosphorus decreased with increasing altitude, indicating a reduction in soil nutrient levels. The levels of soil carbon, nitrogen, and phosphorus are influenced by various processes, including nutrient input, mineralization, immobilization, and leaching (Johnson et al., 1998; Wang et al., 2001; Soon and Arshad, 2002). As altitude increases, higher precipitation promotes vegetation growth and enhances soil organic matter accumulation (Chang-ming et al., 2024). However, unlike soil carbon and nitrogen, soil phosphorus primarily originates from the parent material rather than SOM (Tian et al., 2010). At higher altitudes, increased precipitation may lead to greater nutrient leaching, which in turn reduces soil phosphorus levels (Maojie, 2011; Liu et al., 2019). A similar decline in soil phosphorus with increasing altitude has also been observed in other studies on dry-hot valleys (Xueju, 2005; Zhen-heng and Yuan-bo, 2018). Similarly, another study found that soil nutrient levels responded differently to altitude between sunny and shady slopes, with soil nitrogen increasing with altitude on sunny slopes but decreasing on shady slopes (Zhen-heng and Yuan-bo, 2018).

Stoichiometric homeostasis is a key parameter in ecological stoichiometry (Zhou et al., 2024). Plants with strong stoichiometric homeostasis are relatively conservative in nutrient use, whereas those with weaker homeostasis can flexibly use nutrients (Yu et al., 2010). Thus, the level of stoichiometric homeostasis reflects plant ecological adaptation mechanisms (Wirtz and Kerimoglu, 2016; Peng et al., 2017). In this study, M. alba exhibited strong homeostasis, as no significant correlation was found between the nutrient contents and stoichiometry ratios of M. alba leaves and soil. Additionally, M. alba leaves and soil responded differently to slope aspects and altitude. Moreover, the differences in nutrient content and stoichiometry between M. alba and soil were not constrained within a specific range. The 1/H calculation results indicated that M. alba was strictly homeostatic. These findings clearly support our expectation that plants in resource-limited environments, such as dry-hot valleys, exhibit strong stoichiometric homeostasis. Plants employ various metabolic and physiological mechanisms to maintain stable nutrient levels when nutrients limit their growth. The level of homeostatic flexibility largely depends on how effectively they use these limited resources (Hessen et al., 2004; Peng et al., 2017; Hong-bo et al., 2021; Su and Shangguan, 2022). Previous studies have reported that plants in dry-hot valleys can maintain stoichiometric stability by increasing nutrient absorption and resorption, which may lead to a decoupling of plant and soil stoichiometry (Jingwen et al., 2020). Plants with stronger stoichiometric homeostasis are better adapted to environmental changes (Chen et al., 2016; Peng et al., 2017; Wei et al., 2021). Our study demonstrated that M. alba, with its strict stoichiometric homeostasis, is well-equipped to thrive in a dry-hot environment.