As essential yet trace-level components of living systems, micronutrients (e.g., Fe, Mn, Zn, and V) sustain fundamental ecological processes, including photosynthesis (Fe, Mn; Fischer et al., 2015; Schmidt et al., 2020), respiration (Fe; Dallman, 1986), enzymatic/redox functions (Zn, V; Hänsch and Mendel, 2009), and biological nitrogen fixation (V; O'Hara, 2001). Crucially, micronutrient gradients in soils propagate through trophic chains, directly influencing human nutrition and health; deficiencies exacerbate global malnutrition burdens (Fageria et al., 2002; White and Broadley, 2005). Despite their pivotal role in ecosystem stability and food security (Prestele et al., 2016; Stehfest et al., 2019), critical knowledge gaps persist regarding the distribution patterns and drivers of soil micronutrients from regional to global scales.

The Tibetan Plateau represents a uniquely important yet underrepresented region for addressing these knowledge gaps. As the high-altitude landmass and the most extensive alpine permafrost region at low to mid-latitude (Yao et al., 2012), the Tibetan Plateau supports vast areas of cold- and nutrient-limited ecosystems, where plant productivity, microbial activity, and biogeochemical cycling are particularly sensitive to micronutrient availability (Han et al., 2022; Tian et al., 2019). In such environments, micronutrients may act as co-limiting or even primary limiting factors alongside nitrogen and phosphorus, yet their spatial distributions and environmental controls remain poorly quantified (Dong et al., 2023; Pan et al., 2024). Moreover, the Tibetan Plateau is experiencing rapid climate change characterized by accelerated warming and increasing moisture availability, which is expected to fundamentally alter soil weathering regimes, redox conditions, and mineral-organic interactions (Yu et al., 2024; Cheng et al., 2024). Without robust baseline assessments of soil micronutrients, it remains difficult to evaluate how ongoing climate change may reshape nutrient limitation patterns and ecosystem functioning across high-altitude regions.

Soil micronutrient supply originates from coupled physicochemical weathering and biological mediation, critically regulated by local climate and topography (Ochoa-Hueso et al., 2020; Hartmann and Six, 2023). In cold-arid high-altitude regions, particularly the Tibetan Plateau, extreme environmental interactions uniquely govern micronutrient cycling. Cryogenic processes such as glacial erosion and freeze-thaw cycles, accelerate physical bedrock weathering to mobilize lithogenic micronutrient reservoirs, while aridity concurrently constrains chemical weathering and elemental release (Mu et al., 2020, 2016). Low temperatures suppress biological turnover and synergize with aridity to compromise pedogenesis through clay deficits and diminished mineral reactive sites, thereby reducing elemental retention capacity (Dijkstra et al., 2004). The combined effects of these opposing processes jointly determine the overall abundance and spatial heterogeneity of soil micronutrients across alpine soils, likely differing from the patterns of soil micronutrients observed in temperate and tropical regions.

Despite the ecological significance of soil micronutrient, our understanding of their spatial distribution and controlling factors across the Tibetan Plateau remains limited. Previous studies have been largely confined to localized transects or site-specific investigations (e.g., the Heihe River Basin and the Tibetan Plateau Highway), offering limited spatial coverage and representativeness, and providing insufficient insight into how climate, vegetation, topography and lithology jointly regulate micronutrient patterns at the regional scale (Zhang et al., 2012; Guan et al., 2017; Bu et al., 2016). To address these knowledge gaps, we conducted a large-scale field investigation across the Tibetan Plateau, establishing 526 sampling sites distributed across representative temperature and moisture gradients (Fig. 1). The sampling design encompassed the plateau's dominant vegetation types and lithological classes. Using this dataset, we analyzed distribution patterns and key controlling factors for four essential micronutrients (Fe, Mn, Zn, V). We then applied a Random Forest algorithm to generate high-resolution spatial distribution maps of these micronutrients, representing the first comprehensive quantification at this scale and resolution.

We added comparisons of our results with the few available observations on the Tibetan Plateau, as well as broader datasets from other regions, to better place our findings in a wider context (Table S2). In terms of data range and mean values, the contents and magnitudes of the soil micronutrients we observed are generally consistent with previous reports from the Plateau and global grassland soils, indicating the reliability of our datasets. For the Plateau, the mean Zn, and V contents in this study were slightly lower (Fig. S2). These discrepancies are more likely due to differences in study regions. Our sampling covered a much broader area, whereas previous studies focused mainly on local regions such as the Heihe Basin (Bu et al., 2016). Given the substantial spatial heterogeneity of soil micronutrients across the Tibetan Plateau, such differences are expected and further highlight the necessity of exploring soil micronutrient patterns at the plateau scale.

The vegetation-type contrasts likely reflect ecosystem controls on topsoil (0–10 cm) geochemistry (Fig. 3). These contrasts are consistent with ecosystem-mediated controls on topsoil geochemistry, where denser vegetation generally enhances near-surface nutrient cycling through greater biomass inputs, rhizosphere processes, and organic matter accumulation (Hinsinger, 2000). Increased organic inputs and reactive organo-mineral associations can promote retention of Fe, Mn, and Zn and, indirectly, V that is often associated with Fe-bearing phases (Lehmann and Kleber, 2015). In contrast, arid steppe and desert systems typically exhibit weaker soil development and lower organic inputs, reducing reactive surfaces and binding sites for micronutrients, resulting in lower surface pools (Kabata-Pendias, 2010). Cropland tends to show intermediate levels, likely reflecting mixing within the plough layer and partial homogenization of surface concentrations (Alloway, 2008). Lithological class imposes a first order control on surface soil micronutrient variability (Fig. 4), consistent with the well-established role of parent material in determining baseline micronutrient pools through differences in primary mineral assemblages and weathering products. In our dataset, acidic metamorphic-derived soils are consistently enriched in Fe, Mn, Zn, and V relative to several other lithologies, whereas acidic igneous and fluvial facies substrates tend to show lower contents, with carbonate, eolian facies, and clastic sedimentary rocks generally occupying intermediate levels. Such contrasts are expected where lithology governs not only the supply of Fe and Mn-bearing minerals but also the abundance of secondary oxides and clay surfaces formed during pedogenesis, which strongly influence trace-element retention and depletion intensity at regional scales (Lara et al., 2018; Spinola et al., 2022). The coherent behaviour of V alongside Fe is also mechanistically plausible because vanadate is efficiently retained by Fe and Mn oxides, making V distributions closely coupled to the abundance and reactivity of these mineral phases (Abernathy et al., 2022; Larsson et al., 2017).

The distribution of soil micronutrients across the Tibetan Plateau demonstrates significant spatial heterogeneity (Figs. 2 and 7), aligning with patterns observed in Europe for elements such as Zn (Van et al., 2023). This variability is predominantly governed by the interactions among soil, topography, and climate (Figs. 5a–d and 6). Soils act as the principal sink for micronutrients entering the terrestrial environment through both natural processes and anthropogenic inputs. Consistent with continental-scale assessments, clay content emerged as the dominant control on the spatial distribution of soil Zn in Europe, with coarse-textured soils showing markedly lower Zn concentrations (Van et al., 2023). Similar texture effects have been reported for other elements: soil Cu is positively related to clay and negatively related to sand (Ballabio et al., 2018), sensitivity analyses show that soil Se increases with increasing clay (Jones et al., 2017), whereas surface-soil Hg decreases once sand exceeds ∼ 30 % (Ballabio et al., 2021). These patterns agree with our observation that ecosystems developed on coarse substrates contain the lowest levels of micronutrients. This is because clay serves as an adsorption site for metal elements, whereas sandy soils are relatively weak adsorbents for micronutrients (Mesquita, 1998). In poorly developed soils with low clay content, such as those in arid climates, the lack of capacity to retain soil organic matter leads to a significant reduction in the amount of micronutrients, making them less likely to accumulate (Moreno-Jimenez et al., 2019). In addition to soil covariates, factors such as topography also play a significant role in the micronutrient content of surface soils on the Tibetan Plateau. Accumulated local effect (ALE) curves indicate pronounced, nonlinear elevation responses in soil micronutrients, with Fe and Mn generally decreasing with elevation, whereas Zn and V show threshold-like behaviour, remaining relatively stable at mid-elevations but changing sharply at higher elevations (Fig. S3). Contents of Fe, Mn, Zn, and V increased with slope steepness, particularly below 10°. This pattern suggests that topography governs both physical and chemical soil-forming processes by controlling erosion, leaching, and moisture dynamics (Jenny, 1941; Montgomery, 2007). Steeper slopes enhance runoff and erosion, removing fine particles enriched in Fe and Mn oxides and promoting elemental redistribution along hillslopes, while also accelerating bedrock weathering and releasing lithogenic elements such as Fe and V (Yoo et al., 2006). In contrast, gentler slopes favor organic matter accumulation and reducing conditions that can immobilize or leach redox-sensitive elements. Overall, slope-driven variations in hydrology, weathering, and redox status collectively shape the spatial heterogeneity of soil micronutrients across the Plateau (Quesada et al., 2010).

Notably, precipitation emerges as the primary predictor for micronutrients such as Fe, Mn, Zn, and V, with all elements exhibiting characteristic U-shaped responses, with minima occurring at 300–500 mm. This pattern likely reflects distinct weathering regimes across precipitation gradients. In arid regions (< 300 mm), physical weathering processes, including freeze-thaw fracturing and aeolian erosion, predominate, allowing micronutrients to accumulate near the surface due to evaporation effects (Pachauri et al., 2014; Moreno-Jimenez et al., 2023). In transitional precipitation regimes (300–500 mm), intensified chemical weathering occurs; however, leaching fluxes surpass the rates of parent material weathering, leading to soil elemental depletion (Anderson, 2019; Bluth and Kump, 1994; Hartmann et al., 2014). In humid regions (> 500 mm), enhanced chemical weathering results in the formation of secondary clay minerals (e.g., montmorillonite, illite), whose negatively charged surfaces facilitate elemental retention through ionic adsorption and co-precipitation mechanisms (Alloway, 2009). To further verify these nonlinear effects, accumulated local effects analyses were performed and revealed consistent U-shaped responses for Fe, Mn, Zn, and V, confirming the robustness of the partial dependence results (Fig. S4). This pattern likely reflects a trade-off between weathering intensity, leaching losses, and elemental retention under contrasting hydrological conditions. Additionally, MAP showed a significant positive correlation with CIA (Fig. S5), indicating that higher precipitation enhances chemical weathering. However, this relationship alone does not fully explain the observed U-shaped nutrient responses, as direct indicators of leaching fluxes and mineral retention are lacking. We therefore consider this U-shaped pattern, supported by multiple lines of evidence, as an open question for future investigation, emphasizing the need for complementary datasets (e.g., redox status, mineral composition) to refine its mechanistic interpretation. Although lithological differences are detectable for some elements, they do not alter the overall spatial patterns (Fig. 4). In summary, lithology modulates the baseline levels of certain micronutrients but is not the dominant factor shaping their regional distributions, which are more strongly governed by external environmental drivers such as climate, weathering intensity, sediment transport, and biotic cycling.

Our findings suggest that the aridity index is a significant determinant of soil micronutrient distribution. Specifically, elemental contents tend to decrease when the aridity index falls below a certain threshold (Fig. 6). This trend likely reflects reduced input or retention of elements under arid conditions. Grouped comparisons across aridity classes (humid, dry sub-humid, semi-arid, arid, and hyper-arid) were performed. The results show that the medians and means of Fe, Mn, Zn, and V decrease with increasing aridity (Fig. S6). Drought conditions may modify soil redox states, thereby influencing element speciation, adsorption capacity, mobility, and ultimately, leaching behavior (Brady and Weil, 2016; Loveland and Webb, 2003; Carter and Stewart, 1995). Also, arid environments may indirectly impact micronutrients through alterations in soil pH and soil organic matter content (Moreno-Jimenez et al., 2019). Previous research has shown that droughts induced by climate change can restrict the availability of essential micronutrients, such as iron and zinc (Moreno-Jimenez et al., 2019; Bista et al., 2018). This limitation, along with other adverse effects like diminished water availability, poses substantial threats to vital ecological processes and services in drylands, including food production (Gupta et al., 2008; Graham, 1991). Furthermore, ongoing regional warming and associated hydroclimatic shifts may interact with hydrology and redox processes, organic-matter cycling, and vegetation patterns, thereby altering the availability and spatial variability of certain micronutrients (Myers et al., 2014; Pachauri et al., 2014).

Multiple biogeochemical processes are sensitive to micronutrient availability. Biological nitrogen fixation relies on Mo and Fe as essential cofactors of nitrogenase, while the methane cycle depends on Ni and Cu through their roles in methanogenesis and methane oxidation, respectively (Thauer, 2019; Stefan et al., 2020). Permafrost thaw and thermokarst development can further activate Fe-Mn redox cycling, altering metal mobility (Chauhan et al., 2024). Collectively, soil micronutrients depletion may cascade through nutrient cycling and ecosystem feedbacks, amplifying the impacts of ongoing environmental change across the Plateau.

Although the laboratory XRF measurements used in this study were validated against ICP-MS and generally showed strong correlations, several limitations should be noted. For Fe, Mn, and V, the two methods produced highly consistent results, supporting the robustness of the dataset. In contrast, Zn exhibited systematic deviations from the 1:1 line, with XRF tending to slightly underestimate contents relative to ICP-MS (Fig. S1). These deviations likely arise from element-specific detection sensitivities of XRF. To address this, we calibrated Zn, using the regression relationships between the two methods and then repeated spatial mapping. The results before and after calibration were not significantly different, confirming that these deviations do not affect our main findings (Figs. S7 and S8). Furthermore, the primary scientific purpose of this dataset is to characterize spatial patterns, ecological gradients, and multivariate relationships across the Plateau, rather than to provide ultra-precise quantification of element contents at the individual-sample level. For these objectives, XRF-derived contents, supported by calibration, cross-validation, and transparent uncertainty reporting, provide reliable and ecologically interpretable information. Nonetheless, we acknowledge that incorporating ICP-based measurements on a larger sample set would further strengthen the accuracy of the dataset, and future efforts should consider combining multiple analytical approaches to minimize methodological uncertainties.

Our random forest regression models demonstrated robust predictive capability (e.g. cross-validated R2 range from 0.65 to 0.77 for Fe, Mn, and Zn; Table S3), V exhibit weaker predictive performance (R2=0.39). The spatial predictions inevitably contain uncertainties arising from both the Random Forest model and the input datasets. Model-based uncertainty was quantified as the inter-tree variance among predictions (RF-SD; Fig. S9), while additional uncertainty may stem from sparse meteorological observations and limited soil sampling across the Tibetan Plateau. Large regions in the west (including parts of the Changtang Plateau and Ngari) are sparsely inhabited and difficult to access, with extensive high-elevation terrain and access restrictions that constrain fieldwork. Consequently, model estimates in these areas are supported by fewer observations and should be interpreted with greater caution. We therefore recommend that users consider the accompanying uncertainty layer when applying the gridded products, especially for analyses focusing on the western Plateau or on fine-scale local interpretation. Nevertheless, model accuracy could be further improved to more extensive field sampling and refinement of input data. Specifically, targeted collection of soil micronutrient data and associated covariates in underrepresented high-altitude regions of the Tibetan Plateau is necessary to address existing spatial gaps. Additionally, systematic reduction of uncertainties inherent in gridded environmental datasets is essential, as these uncertainties propagate errors into micronutrient predictions. Continued advancement in both field observations and foundational geospatial dataset is crucial for improving the reliability of regional-scale element mapping.

We note that the reported contents represent a fixed-depth (0–10 cm) surface layer, which may correspond to different dominant horizons among land uses (organic-rich surface in some forests, mineral topsoil in grasslands/meadows, and the plough layer in croplands). Users should consider this context when applying the dataset, especially for process inference or when comparing across ecosystems with contrasting surface horizons. Where applications require deeper profiles or horizon-specific interpretation, we recommend integrating our maps with local profile data. Because croplands occupy a very small fraction of the Plateau and are represented by a limited number of samples in our area-weighted design, cropland-related results should be interpreted primarily as a cross-ecosystem contrast at the plateau scale. Users addressing cropland-specific questions are advised to use these values with caution and, where possible, complement them with targeted cropland sampling or independent field observations.

The gridded soil micronutrient (Fe, Mn, Zn, and V) maps for Tibetan Plateau can be downloaded from 10.11888/Terre.tpdc.303242 (Huo et al., 2026). Users should note that laboratory XRF provides robust estimates for major elements (e.g., Fe), whereas trace-level elements (particularly Zn) carry higher uncertainty even after ICP-MS calibration. The dataset is intended primarily for plateau-scale analyses and cross-ecosystem spatial comparisons, caution is advised when using absolute Zn contents for applications requiring high analytical precision (e.g., bioavailability thresholds, detailed mass-balance calculations, or benchmarking requiring near-reference values), and users should consult the accompanying calibration statistics and uncertainty diagnostics.

This study delivers a comprehensive assessment of spatial distribution patterns for four soil micronutrients (Fe, Mn, Zn, and V) across the Tibetan Plateau, revealing pronounced regional-scale heterogeneity. Soil covariates were the primary contributors to the spatial variation of micronutrient contents, followed by topographic and climatic factors, whereas vegetation and grazing disturbance had relatively minor effects. These findings highlight the coupled effects of climate, vegetation, and parent material on micronutrient biogeochemical cycling within the complex environmental context of the Tibetan Plateau. Using five predictor groups (climate, vegetation, soil covariates, topography, and grazing disturbances), we generated high-resolution spatial maps for key soil micronutrients (Fe, Mn, Zn, and V) via machine learning. These maps provide validated initial conditions for process-based models simulating micronutrient cycling, advances understanding of elemental distribution in alpine ecosystems.