Introduction

Elevation strongly influences soil fertility by affecting temperature, precipitation, erosion, and the movement and deposition of soil materials. These factors shape the soil’s physical and chemical properties, which in turn determine its agricultural potential. As global food demand rises, understanding how soil fertility varies with elevation has become essential for improving nutrient management, ensuring food security, and adapting to climate change1,2. Such knowledge also supports the design of targeted nutrient management strategies to address deficiencies and prevent nutrient toxicity in cultivated soils3.

In mountainous regions such as Yunnan Province, China, large differences in elevation create diverse microclimates and soil conditions. Effective soil fertility management in these landscapes requires accurate information on soil biochemical and physical characteristics along elevation gradients4,5.

The Nutrient Value Index (NVI), introduced by Parker et al.6, provides a systematic way to evaluate nutrient availability and detect variability across landscapes. It has been widely used in agricultural studies to identify fertility heterogeneity across elevation zones7,8. For example, Beg and Chaurey9 applied the NVI in the Indian Himalayas and found that nitrogen (N), phosphorus (P), and potassium (K) contents varied markedly with altitude. Similar findings have been reported across Asia and Africa, confirming that elevation can significantly influence key nutrients such as magnesium (Mg), sulfur (S), and zinc (Zn)10,11.

In Yunnan, complex topography and strong elevation gradients make uniform fertilization practices ineffective12. Although the region has a long agricultural tradition, climate change, land degradation, and socio-economic pressures are threatening its productivity13,14. Several studies have reported clear elevation-related differences in soil organic carbon, available phosphorus, and other nutrients, all of which affect crop yield and long-term soil health5,15.

This study investigates how elevation affects soil fertility in Yunnan’s agricultural lands. We analyzed soil samples from four elevation ranges (1600–2000, 2000–2400, 2400–2800, and > 2800 m) using the NVI to assess nutrient variability. The results provide insights into the relationship between altitude and soil fertility, offering guidance for improving soil management, optimizing fertilizer use, and supporting sustainable agriculture in mountainous regions of Yunnan.

Materials and methods

Description of the study area

The study was conducted across 1,438 km2 of agricultural land in Yunnan Province, China, located between 100°50′00″–105°00′25″ E and 20°15′00″–30°00′00″ N (Fig. 1). The region lies within a subtropical highland climate zone under the Köppen classification and experiences a monsoon climate with distinct wet and dry seasons. Annual rainfall averages about 1,200 mm, with the wet season from May to October and the dry season from November to April. The mean annual temperature is approximately 15 °C, varying considerably with elevation.

Fig. 1
figure 1

Location of the study area and sampling point distribution.

Full size image

Elevations range from 1,600 to over 2,800 m, producing diverse microclimates that strongly influence soil-forming processes such as leaching and nutrient accumulation13. The main crops include maize, rice, and tea, all sensitive to soil fertility limitations, particularly at higher altitudes12. Understanding the effects of elevation on soil properties is essential for developing suitable soil management and fertilization strategies in this mountainous agricultural region.

The soils in the study area are mainly classified as Alfisols, Ultisols, and Oxisols under the USDA system and occur widely across the region’s elevation gradient. Farmers have traditionally relied on surface irrigation, but declining water availability has encouraged the adoption of more efficient systems such as drip and sprinkler irrigation. Common fertilizers include urea [CO(NH₂)₂], diammonium phosphate [(NH₄)₂HPO₄], and sulfur-based compounds, complemented by organic amendments like humic acid and foliar nutrient sprays. These inputs help mitigate the nutrient deficiencies typical of Yunnan’s agricultural soils2,12.

Methods

In this study, soil fertility was evaluated using two classification approaches: the Common method and the Gomez method. The use of both techniques enabled a comparison between a regionally standardized system and a statistically derived, elevation-sensitive approach. The Common method, based on FAO, USDA, and Chinese soil classification standards, applies conventional agronomic thresholds to group soil properties into three categories—low, optimal, and high—making it practical for field management and fertilizer recommendations in Yunnan. In contrast, the Gomez method6 is a data-driven statistical system that defines class boundaries using the dataset’s mean and standard deviation, allowing the detection of subtle fertility variations across elevation zones. By integrating both methods, this study combines a globally recognized, quantitative framework with a locally calibrated, management-oriented one, enhancing both the precision and practical relevance of soil fertility assessment in Yunnan and other mountainous agricultural regions with similar elevation-related variability.

Land units determination and soil sampling

Land units for soil sampling were selected based on the elevation gradient of the study area, which was divided into four categories: 1600–2000 m, 2000–2400 m, 2400–2800 m, and above 2800 m above sea level (m asl). The highest elevation zone (> 2800 m) covered approximately 27,770 ha. In total, 350 surface soil samples were randomly collected from agricultural lands across these zones to ensure comprehensive representation of the region’s soil characteristics. Sampling distribution included 85 samples from 1600–2000 m, 91 from 2000–2400 m, 99 from 2400–2800 m, and 75 from areas above 2800 m. This stratified random sampling design was intended to capture the influence of elevation on soil fertility and related physical and chemical properties4,16.

Laboratory analysis and data evaluation

Soil texture was determined using the hydrometer method17. Electrical conductivity (EC, dS m⁻1), pH, and organic matter (OM, %) were analyzed following the standard procedures of Sparks et al.18. Calcium carbonate (CaCO₃, %) was measured by acid reaction with HCl19. Total nitrogen (TN, %) was analyzed using the Kjeldahl method20, available phosphorus (AP, mg kg⁻1) by spectrophotometry following Olsen’s method21, and available potassium (AK, mg kg⁻1) using a flame photometer according to USDA-NRCS22. Micronutrient availability, including Fe, Cu, Zn, and Mn, was assessed with the DTPA extraction technique23. Soil fertility was then classified into three levels—low, optimal, and high—using both the Common method, based on established agronomic thresholds, and the Gomez method6, which defines class boundaries statistically from the dataset’s mean and standard deviation. Classification thresholds for each property across elevation zones are summarized in Eq. (1) and Table 1.

$${\text{X}}_{\text{i}}\le \text{X}-\text{SD }\left(\text{Low}\right)$$
$$\text{X}-\text{SD}<{\text{X}}_{\text{i}}<\text{X}+\text{SD }(\text{Medium})$$
$${\text{X}}_{\text{i}}\ge \text{X}+\text{SD }(\text{High})$$

where \({\text{X}}_{\text{i}}\) represents the value of the soil property, X is the mean of that property, and SD is the standard deviation.

Table 1 Classification criteria for soil properties.
Full size table

The classification thresholds in Table 1 integrate both globally recognized standards (FAO, USDA-NRCS, IUSS) and regionally adapted references for Asian and Chinese agricultural soils24,Sparks et al., 2023;25. Global standards were included to ensure international comparability, while regional thresholds better represent the pedoclimatic and agronomic conditions of China and neighboring regions. For this study, focused on Yunnan Province, regionally adapted thresholds were used as the primary basis for management recommendations, with global criteria serving only as secondary benchmarks. This combined approach ensures that the interpretations are locally meaningful yet remain consistent with broader international frameworks.

After classifying soil attributes, the Nutrient Index Value (NIV) was computed using the formula outlined by Parker et al.6, as shown in Eq. (2). The classification criteria for the NIV are provided in Table 2:

$$NIV=\frac{1\times {S}_{Low}+2\times {S}_{Medium}+3\times {S}_{High}}{{S}_{Low}+{S}_{Medium}+{S}_{High}}$$

where \({S}_{Low}\), ​\({S}_{Medium}\) and \({S}_{High}\) are the number of soil samples classified as low, optimal, and high, respectively.

Table 2 Nutrient index with range and remarks.
Full size table

Nutrient index value (NIV) mapping using GIS-based IDW interpolation

To visualize the spatial variability of soil fertility, Nutrient Index Values (NIVs) for key parameters (EC, pH, CaCO₃, OM, and TN) were interpolated using the Inverse Distance Weighting (IDW) algorithm in ArcGIS 10.8. The IDW method estimates unknown values based on the proximity and weighting of measured points, generating continuous surfaces that illustrate localized nutrient distribution patterns. Each sampling location was geo-referenced using WGS-84 coordinates and classified into Low, Optimal, or High fertility levels according to both the Gomez and Common methods. Spatial interpolation was conducted separately for each parameter and elevation zone (1600–2000, 2000–2400, 2400–2800, and > 2800 m) to account for altitudinal effects on soil fertility. The resulting maps display interpolated nutrient gradients using standardized color scales: red for Low, yellow to green for Optimal, and blue for High fertility. Legends, units, and labels were standardized and translated into English for clarity. These GIS-based maps clearly depict the NIV–Gomez and NIV–Common patterns and visualize IDW-derived gradients for major soil properties across Yunnan’s elevation zones.

Results

Statistical analysis

The statistical summary of soil properties is presented in Table 3. Soil texture showed wide variation, with sand ranging from 5–78%, silt from 10–68%, and clay from 5–48%, resulting in eight textural classes: clay, silty clay, silty clay loam, clay loam, silt loam, loam, sandy clay loam, and sandy loam. This variability strongly influences soil structure and nutrient dynamics, thereby affecting crop growth and productivity8. As described in Sect. 2.3, total nitrogen (TN) values represent total forms, while phosphorus (P), potassium (K), and micronutrients (Fe, Mn, Zn, and Cu) are expressed as available forms.

Table 3 Descriptive statistics of soil properties in the study area.
Full size table

Electrical conductivity (EC) ranged from 0.18 to 1.75 dS/m, with a mean of 0.62 dS/m and a coefficient of variation (CV) of 56.45%, indicating non-saline soil conditions. Soil pH values ranged from 7.30 to 8.40 (mean 7.92; CV 2.27%), showing the lowest variability among all measured parameters and reflecting slightly to moderately alkaline conditions35. Calcium carbonate (CaCO₃) content varied between 0.70 and 35.20% (mean 17.80%,CV 51.12%), further supporting moderate alkalinity. Organic matter (OM) ranged from 0.10 to 3.80% (mean 1.20%; CV 54.17%), placing most soils in the moderate fertility category, as values below 1% are considered very low and those above 3% high36. Total nitrogen (TN) ranged from 0.005 to 0.20% (mean 0.06%,CV 66.67%), while available phosphorus (AP) exhibited substantial variability, from 0.60 to 950 mg/kg (mean 40 mg/kg; CV 428.6%). Available potassium (AK) averaged 420 mg/kg, corresponding to moderate to high fertility levels. Mean concentrations of micronutrients—iron (Fe), manganese (Mn), zinc (Zn), and copper (Cu)—were 2.45, 5.10, 0.48, and 0.25 mg/kg, respectively, indicating generally low to moderate micronutrient availability across the study area7.

Classification of soil properties based on the NIV index

Table 4 summarizes the classification of soil properties using the Nutrient Index Value (NIV) under the Gomez and Common methods. Both approaches classified electrical conductivity (EC) as Low, with 72% and 91% of samples in the Gomez and Common methods, respectively. For pH, the Gomez method identified an optimal range of 7.5–8.2, classifying 88% of samples as Optimal, whereas the Common method categorized all samples (100%) as High (> 7.4). Calcium carbonate (CaCO₃) classifications also differed, with 83% of samples ranked as Optimal by Gomez and 62% by Common. The threshold between Low and Optimal was lower in Gomez (6%) than in Common (31%).

Table 4 Percentage of soil sample categories for soil properties and NIV index using two classification methods.
Full size table

Organic matter (OM) showed notable discrepancies: 65% of samples were Optimal under Gomez but 79% were Low under Common. Total nitrogen (TN) was consistently Low in both methods, with 78% (Gomez) and 88% (Common) of samples falling into this class. For available phosphorus (P), 47% of samples were Optimal under Gomez, while 82% were Low under Common. Potassium (K) was mostly Optimal in Gomez (58%) but predominantly High in Common (75%).

Micronutrient classifications further highlighted methodological differences. Iron (Fe) was Low in both systems (94% in Gomez; 100% in Common), while manganese (Mn) was Optimal in Gomez (77%) but Low in Common (97%). Zinc (Zn) was Optimal in Gomez (93%) and Low in Common (68%), and copper (Cu) was Low in both methods (60% and 100% for Gomez and Common, respectively). Soil texture, evaluated only through the Gomez method, showed clay as Low (73%), sand as High (59%), and silt as Optimal (72%).

Overall, substantial discrepancies between the Gomez and Common classifications underscore the importance of selecting an appropriate method according to analytical objectives. The Gomez method offers finer differentiation across elevation gradients, while the Common method provides simpler, management-oriented categories. Together, these results enhance the interpretability and reliability of soil fertility assessments across diverse landscapes.

Figures 2 and 3 illustrate the spatial variation of major soil fertility properties within Yunnan Province, with all interpolations and spatial boundaries clipped strictly to the province’s official administrative limits. This ensures that the maps represent true provincial-scale patterns rather than regional trends from other parts of South China.

Fig. 2
figure 2

GIS-based spatial representation of soil fertility levels for major soil properties in Yunnan Province, interpolated using the IDW method and classified according to the Gomez method.

Full size image
Fig. 3
figure 3

GIS-based spatial representation of soil fertility levels for major soil properties in Yunnan Province, interpolated using the IDW method and classified according to the Common method.

Full size image

The Gomez method classifies soils into three categories—Low, Medium (Optimal), and High—based on the dataset’s mean and standard deviation. Figure 2 displays the spatial distribution of key parameters, including electrical conductivity (EC), pH, calcium carbonate (CaCO₃), organic matter (OM), and total nitrogen (TN) across four elevation zones (1600–2000, 2000–2400, 2400–2800, and > 2800 m). The legend provides parameter-specific thresholds, enabling precise and elevation-sensitive interpretation of fertility gradients (Table 5; Figs. 2, 3). This statistically grounded approach captures fine-scale variability and is well suited for detailed scientific and spatial analyses.

Table 5 Percentage of soil sample categories and NIV for soil properties by two classification methods (Gomez and Common) at different elevations.
Full size table

In contrast, the Common method (Fig. 3) applies broader threshold ranges that align more closely with practical field conditions. It simplifies fertility classification for management and planning purposes but lacks the elevation-specific detail offered by the Gomez approach.

The main differences between the two methods are:

  1. 1.

    EC classification – The Gomez method employs lower threshold values, making it more sensitive to minor salinity variations, while the Common method uses wider ranges suitable for general soil assessment.

  2. 2.

    Organic matter and total nitrogen – The Gomez method defines higher thresholds for OM and TN, capturing greater variability, whereas the Common method adopts more conservative limits for routine management use.

  3. 3.

    Elevation-based mapping – The Gomez method provides detailed classification by altitude, which the Common method does not include.

Overall, the Gomez method is better suited for research and precision soil analysis, whereas the Common method is more practical for on-farm management decisions. Integrating both approaches enhances the understanding of spatial fe.

ANOVA and Tukey HSD post-hoc test

The one-way analysis of variance (ANOVA) showed that elevation had no statistically significant effect (p > 0.05) on most soil properties, including pH, organic matter (OM), total nitrogen (TN), potassium (K), and micronutrients (Fe, Mn, Zn, and Cu). Among all parameters, available phosphorus (P) displayed the strongest elevation-related variation, with a borderline significance level (p = 0.076). Tukey’s post-hoc test indicated that soils at 2000–2400 m generally contained less available P than those at 2400–2800 m (mean difference = –17.5 mg kg⁻1, p = 0.075). However, this difference was not significant at the 5% threshold (Table 6). No other pairwise comparisons among elevation zones revealed significant variation for phosphorus or other nutrients. Overall, these results indicate that although descriptive statistics suggested some nutrient variation with altitude, inferential tests did not confirm strong elevation-driven effects on soil fertility.

Table 6 Results of ANOVA and Tukey HSD Post-hoc test to evaluate the effects of elevation (altitudinal zones) on soil nutrient variations.
Full size table

Discussion

This study analyzed soil fertility across elevation gradients in Yunnan Province, China, using the Nutrient Index Value (NIV) through both the Gomez and Common classification methods. A two-step approach was employed: first, NIV values were calculated without accounting for elevation, and then recalculated by incorporating elevation ranges (1600–2000, 2000–2400, 2400–2800, and > 2800 m). Understanding how soil fertility varies with elevation and environmental conditions is critical for optimizing agricultural productivity and developing sustainable management strategies (Li et al., 2021; 5).

Electrical conductivity (EC) values indicated non-saline conditions across all elevations, consistent with earlier studies showing minimal salinitylimitations in Yunnan’s farmlands7,32. Soil pH exhibited a decreasing trend with elevation under the Gomez method but an opposite pattern under the Common method, reflecting methodological differences in classification. Both methods, however, confirmed that most soils are slightly to moderately alkaline (pH > 7.5), which can restrict nutrient availability and crop growth (Liu et al., 2021).

Calcium carbonate (CaCO₃) was generally within low to optimal ranges, supporting favorable buffering capacity. Moderate CaCO₃ levels (5–15%) are known to enhance soil structure, water retention, and nutrient transport35. Organic matter (OM) levels were low to moderate, with slightly higher values above 2800 m, likely due to cooler temperatures and reduced microbial decomposition. Intensive cultivation practices in lower zones may accelerate OM depletion, underscoring the importance of conservation tillage and organic amendments to maintain fertility2,13.

Total nitrogen (TN) was consistently low across all elevations, likely reflecting both low OM content and inefficient fertilizer use15. Available phosphorus (AP) and potassium (AK) also showed wide variability, with phosphorus particularly limited at mid-elevations (2000–2400 m). These findings emphasize the need for balanced nutrient management and elevation-specific fertilizer recommendations to enhance nutrient availability.

Micronutrient analyses further revealed deficiencies in iron (Fe) and copper (Cu), while zinc (Zn) and manganese (Mn) were generally adequate, particularly under the Gomez classification. The Common method, by contrast, categorized all micronutrients as low. These results align with previous research identifying widespread Fe and Cu deficiencies in Yunnan’s agricultural soils37,38, suggesting that targeted micronutrient supplementation could significantly improve soil fertility.

The one-way ANOVA results confirmed that elevation had no statistically significant effect on most soil properties (p > 0.05). Only phosphorus showed a borderline relationship (p = 0.076), with slightly higher concentrations at 2400–2800 m than at 2000–2400 m. While these differences were not significant at the 5% level, they indicate potential elevation-related influences on P availability. High within-zone variability may have masked other subtle patterns, highlighting the need for larger datasets and more advanced analytical models in future research.

Overall, the study demonstrates that although descriptive analyses reveal clear fertility constraints across altitudinal zones, inferential tests indicate that most of these differences are not statistically driven by elevation. This finding underscores that local management practices, rather than altitude alone, play a dominant role in shaping soil fertility. Future studies incorporating larger sample sizes, replicated sampling, and multivariate or mixed-effects analyses will help clarify these interactions. Integrating descriptive and statistical approaches will enable more precise, site-specific soil fertility management to sustain productivity and long-term soil health in Yunnan’s mountainous agricultural systems.

Limitations and perspectives

A primary limitation of this study is the lack of strong statistical evidence for elevation effects on soil nutrient variation. High within-zone variability likely obscured subtle differences, and the reliance on descriptive classifications with limited inferential support reduced the strength of the conclusions. In addition, the analysis was restricted to surface soils (0–30 cm) and did not account for subsoil nutrient dynamics, seasonal changes, or crop-specific nutrient uptake. Future studies should employ larger and more spatially diverse datasets, replicated sampling across multiple seasons, and advanced analytical approaches such as multivariate ANOVA or mixed-effects models. Integrating these methods with crop productivity data and management practices would help clarify the relationship between elevation and soil fertility and improve the precision of management recommendations in mountainous agricultural systems.

Despite these limitations, the study underscores the need for elevation-sensitive soil fertility management in Yunnan. Addressing nutrient imbalances through targeted organic amendments and micronutrient supplementation, alongside adaptive land-use and conservation practices, will be critical for maintaining soil health, improving crop productivity, and ensuring sustainable agriculture in this topographically complex region.

Conclusion

This study provides a comprehensive assessment of soil fertility variations across different elevation gradients within Yunnan Province, China. By applying both the Gomez and Common classification methods, we identified consistent nutrient limitations—particularly alkaline pH, low organic matter, and deficiencies in nitrogen, phosphorus, and certain micronutrients. These factors, rather than elevation itself, emerged as the primary constraints on soil fertility in the province.

The results demonstrate that while topography contributes to subtle spatial differences, most soil fertility variations are driven by management practices and local soil characteristics. Consequently, improving soil health in Yunnan’s agricultural regions requires site-specific nutrient management, increased organic inputs, and targeted micronutrient supplementation.

The Gomez method proved more effective in detecting subtle, elevation-related variations and is recommended for precision soil fertility studies, whereas the Common method remains practical for routine field applications. Future research should combine long-term monitoring, crop yield assessments, and advanced statistical modeling to refine elevation-based management practices and support sustainable agriculture in Yunnan’s mountainous landscapes.