Skip to main content
The Innovation logoLink to The Innovation
. 2025 Jun 9;6(9):100985. doi: 10.1016/j.xinn.2025.100985

Frontiers in earth observation for global soil properties assessment linked to environmental and socio-economic factors

José AM Demattê 1,2,, Raul R Poppiel 1,2, Jean Jesus Macedo Novais 1, Nícolas Augusto Rosin 1, Budiman Minasny 3, Igor Y Savin 4,5, Sabine Grunwald 6, Songchao Chen 7,8, Yongsheng Hong 9, Jingyi Huang 10, Sabine Chabrillat 11,12, Quirijn de Jong van Lier 13, Eyal Ben-Dor 14, Cecile Gomez 15,16, Zhang Ganlin 17, Merilyn Taynara Accorsi Amorim 1, Letícia Guadagnin Vogel 1, Jorge Tadeu Fim Rosas 1, Robert Milewski 11, Asa Gholizadeh 18, Arseniy V Zhogolev 4, José Padarian Campusano 3, Yuxin Ma 3,19, Ho Jun Jang 3, Rudiyanto 20, Changkun Wang 9, Rodnei Rizzo 13, Nikolaos Tziolas 6, Nikolaos Tsakiridis 21, Masakazu Kodaira 22, D Nagesh Kumar 23, Subramanian Dharumarajan 24, Yufeng Ge 25, Emmanuelle Vaudour 26, Shamsollah Ayoubi 27, James Kobina Mensah Biney 28, Abdelaziz Belal 29, Salman Naimi Marandi 27, Najmeh Asgari Hafshejani 27, Eleni Kalopesa 21, Danilo Cesar Mello 1, Marcio Rocha Francelino 30, Elsayed Said Mohamed Salama 5,29, Asmaa Abdelbaki 11,31
PMCID: PMC12447651  PMID: 40979290

Abstract

Soil has garnered global attention for its role in food security and climate change. Fine-scale soil-mapping techniques are urgently needed to support food, water, and biodiversity services. A global soil dataset integrated into an Earth observation system and supported by cloud computing enabled the development of the first global soil grid of six key properties at a 90-m spatial resolution. Assessing them from environmental and socio-economic perspectives, we demonstrated that 64% of the world's topsoils are primarily sandy, with low fertility and high susceptibility to degradation. These conditions limit crop productivity and highlight potential risks to food security. Results reveal that approximately 900 Gt of soil organic carbon (SOC) is stored up to 20 cm deep. Arid biomes store three times more SOC than mangroves based on total areas. SOC content in agricultural soils is reduced by at least 60% compared to soils under natural vegetation. Most agricultural areas are being fertilized while simultaneously experiencing a depletion of the carbon pool. By integrating soil capacity with economic and social factors, we highlight the critical role of soil in supporting societal prosperity. The top 10 largest countries in area per continent store 75% of the global SOC stock. However, the poorest countries face rapid organic matter degradation. We indicate an interconnection between societal growth and spatially explicit mapping of soil properties. This soil-human nexus establishes a geographically based link between soil health and human development. It underscores the importance of soil management in enhancing agricultural productivity and promotes sustainable-land-use planning.

Keywords: soil security, digital soil mapping, remote sensing, carbon sequestration, soil health

Graphical abstract

graphic file with name fx1.jpg

Public summary

  • Powerful soil-vegetation satellite proxies generated 90-m soil-property maps.

  • 64% of the world’s topsoil is sandy and highly degradable.

  • Global topsoil stores 900 Gt of carbon, of which 54% is in natural vegetation.

  • Tillage positively affects fertility by making nutrients available.

  • Tillage negatively affects soil health by depleting carbon.

  • The top 10 largest countries collectively store 75% of the global carbon stock.

Introduction

Environmental, agricultural, and climate-change research has historically focused on above-ground factors,1 although soil plays a vital role in the functioning and maintenance of global ecosystems. However, it has been marginalized in conservation and protection strategies2,3 and hotspots of biodiversity studies.4 Until the early 2000s, soil was largely overlooked. Since then, soil has received attention because of growing environmental concerns, becoming a prominent topic in international agreements. As soil is one of the largest carbon sinks on Earth,5,6,7 its management is critical in carbon sequestration.8 Soil addresses societal needs for food and energy9; connects with smart climate initiatives10; and influences global water, energy, and biogeochemical cycles.11 The recognition of the essential nature of the soil has led to an increase in research assessing soil functions, such as water storage and erosion control,12 carbon storage and sequestration potential,13,14,15 surface temperature effects on biodiversity,16 the provision of ecosystem services,17 and soil health assessment.18

Human-accelerated soil degradation underscores the urgent need for comprehensive and global monitoring of soil conditions. Although significant progress has been made in mapping soil properties worldwide,19,20 further research is necessary to unravel the ecosystem processes influencing soil health and to assess how these processes vary across different regions. A major challenge is the limited availability of ground soil observation data. As a result, the finest resolution in global soil-property maps is currently limited to 250 × 250 m,20 which may impact prediction uncertainties, reducing the ability to detect short-range spatial variability and hampering the interpretation of soil properties.21 Furthermore, many global assessments of soil properties focused predominantly on soil organic carbon (SOC)14,19,22 and were related to the global carbon cycle.23 However, carbon dynamics in soils are influenced by numerous factors, particularly mineralogy.24 In addition to SOC, other properties, such as pH, cation exchange capacity (CEC), and clay content, are key ones in soil productivity25 and health.18

In this study, we elaborated global-extent maps of topsoil properties reflecting various soil processes across biological, chemical, and physical dimensions.2,26 SOC stocks (SOCSs) are a key biological indicator with significant environmental and economic implications, serving as a proxy for crop productivity, climate-change mitigation, and carbon market valuation. Within the chemical group, CEC supports nutrient retention and storage, while soil pH influences its availability. In the physical category, clay content influences mineralogy, microbiology, and water retention, while bulk density (BD) directly impacts soil health and plant growth.

To achieve this, we employed a fine grid resolution, utilizing remote sensing (RS) data and machine-learning (ML) techniques. Specifically, we introduce Earth observation bare-soil and vegetation proxies to address gaps around the globe. These are fundamentally rooted in the well-documented intrinsic physical relationship between spectral reflectance and soil in proximal studies27 and for RS studies.18,28,29,30 Vegetation impacts soil composition, particularly data from natural forests.31

We can consider from the current literature some key gaps in digital soil mapping (DSM) at a global scale,32 which are the need to (1) obtain more soil observation points, filling the gaps in the current datasets and also covariates more strongly correlated with soil properties; (2) obtain finer spatial resolution products; (3) connect soils with ecological and socio-economic domains through integrated analyses that consider multiple critical soil properties across diverse climates, ecoregions, countries, and other topics; and (4) validate the products directly in the field, despite the statistical parameters.

Thus, this work contributes to filling these gaps by mapping six key soil properties (SOC content and SOCS, clay, pH, CEC, and BD) at a fine spatial resolution with an innovative system. We also discuss in depth the relationship between these properties with land use, climate, soil management, soil health, economic factors, and other topics across the globe. In collaboration with expert partners worldwide, we pooled together a novel global database consisting of public and private. We bring to light the comparison and discussion between data from different science domains.33 Our research provides compelling insights on soil properties with current global socio-economic challenges, in particular the role of soil in addressing existential issues of climate-change mitigation, water and food security, and biodiversity protection.

Materials and methods

Soil data

We selected key indicators relevant to sustaining healthy and functional soils.29 We leveraged global open soil databases and reached out to multidisciplinary researchers and organizations worldwide to gather over 150,000 georeferenced soil observations from countries around the world (Figure S1, Supplemental Note 1; Tables S1–S3, Supplemental Note 2). However, some had data access restricted due to institutional policies. The data were harmonized, quality checked, and then merged into a unified database for interpolation to a depth interval of 0–20 cm using an equal-area spline to preserve mass.34 We specifically focused on the soil properties with high data availability for spanning the soil pillars: biological, SOC (g kg−1) and SOCS (Gt); chemical, CEC (mmolc kg−1), and pH in water (unitless); and physical, clay content, and BD, g cm−3). The BD was estimated using an improved pedotransfer function.35

Environmental covariates

We carefully selected a (parsimonious) set of environmental covariates (Table 1) to represent key soil formation factors, adopting the SCORPAN (soil, climate, organisms, relief, parent material, age, and spatial position) model framework.36 We used Google Earth Engine (GEE)37 to retrieve these covariates from RS data, where their pixel sizes were resampled to a target size of 90 m to provide comprehensive spatial predictions. The soil-vegetation image represented the factors related to soil, parent material, age, and organisms. It was derived from the Landsat series imagery between 1985 and 2022, processed to retrieve cloud-free surface reflectance of bare surfaces,38 where remaining gaps due to persistent land cover across time were filled with the reflectance of potential natural (permanent) vegetation using the Geospatial Soil Sensing System (GEOS3).29 This consists of continuously validating the spectral information acquired by the system, with fieldwork, until reaching the best of its performance. In conclusion, the new GEOS3 passed through 5 years of constant improvement until reaching the final bare-soil image.

Table 1.

Set of environmental covariates used for spatial predictions of soil properties worldwide

Factor Predictor name Resolution (m) Source Reference
Soil, parent material, age, and organisms soil-vegetation image (six Landsat bands) 30a GEOS3 Demattê et al.29
Climate annual temperature 1,000a CHELSAb Karger et al.39
annual precipitation 1,000a
Relief elevation 90 MERITc Safanelli et al.40
slope 90 Theobald et al.41
multiscale topographic position Index 90 Yamazaki et al.42
Spatial position Euclidean distance to rivers 90 Behrens et al.43
a

The native resolution was resampled to 90 m.

b

Database: CHELSA (Climatologies at High Resolution for the Earth’s Land Surface Areas, https://chelsa-climate.org/)

c

Database: MERIT DEM (Multi-Error-Removed Improved-Terrain Digital Elevation Model, http://hydro.iis.u-tokyo.ac.jp/∼yamadai/MERIT_DEM/)

Annual mean temperature and precipitation were derived from the CHELSA Bioclimatic dataset39 to represent climatic influences on soil properties. Topographic factors, including elevation, slope,40 and multiscale topographic position index,41 were derived from the MERIT Digital Elevation Model.42 The factor related to the spatial position was represented by the Euclidean distance from sites to rivers equal to or greater than the third level of Strahler’s ordination system.43

Soil-property mapping

We processed soil data to match the target spatial resolution by averaging values within a 90 × 90-m area, thus ensuring a single soil-property value per pixel. At each averaged soil data site, covariates were sampled to create a comprehensive dataset for modeling. The approach followed the geospatial mapping pipeline44 using a fully Python-based workflow implemented on Google Colab. The cloud infrastructure, including Google Cloud Storage and GEE, supports data-storage, processing, and computational needs. This framework enabled streamlined handling of our large geospatial datasets, facilitating the integration and analysis of soil data and environmental covariates for mapping of soil properties worldwide.

Model building and optimization

The SCORPAN modeling paradigm36 predicts individual soil properties (SOC, SOCS, clay, CEC, pH, and BD) using stacked environmental variables as inputs. The model training process involved hyperparameter tuning of a random forest (RF) algorithm,45 focusing on three key parameters: variables per split, defined as ranges from 4 to the total number of covariates stepped by 4, controlling the number of covariates considered at each tree split; minimum leaf population, tested values of 3, 6, and 9, setting the minimum sample size for each tree leaf; and maximum nodes, set at 500, 1,000, and 1,500, determining the maximum complexity of each tree.

A grid search across these parameters yielded 27 model combinations. Each model was evaluated using 10-fold cross-validation, stratified by biome to ensure representative performance across ecological regions. Performance metrics, including the coefficient of determination (R2), root-mean-square error (RMSE), and ratio of performance to interquartile range (RPIQ), were calculated. Models were ranked based on the highest mean R2 scores across all folds, and the top-performing model was selected for final predictions.

Spatial prediction

Bootstrap sampling was used to assess prediction uncertainty and generate ensemble soil maps. The site-specific soil data were resampled 100 times, with each iteration ensuring representative sampling across different biomes. For every bootstrap iteration, a new RF model was trained and applied to classify the study area, resulting in a series of predictive maps. These maps were aggregated to calculate the mean prediction; the ensemble mean of all bootstrap predictions, forming the final soil map; and the uncertainty map, calculated by the difference between the 2.5th and 97.5th percentiles across bootstrap predictions, highlighting areas with higher predictive uncertainty.

Data interpretation

To assess the mapped soil properties across multiple global perspectives, we intersected vector polygon data from diverse sources (Table 2) with our soil-property maps. For each polygon, we computed zonal statistics (including mean and sum) of the soil properties for further analysis. This allowed us to analyze spatial variability and distribution of soil properties within defined geographic and socio-environmental frameworks.

Table 2.

List of global data crossed with our soil-property maps

Category Description References
Climate zones updated Köppen-Geiger climate zones Beck et al.46
Terrestrial ecoregions ecoregions focusing on vegetation types within the 14 major world ecoregions Olson et al.47
Countries 196 political subdivisions across the world N/A
Continents six densely populated regions of the planet N/A
World all mapped Earth area N/A
Land use and land cover different land uses and land cover, considering stratification by largest countries by area, continents, climate zones, and biomes Zanaga et al.48
GDP total and per capita mean values from 2021–2028 projections for the 20 wealthiest, intermediate, and poorest countries International Monetary Fund49
Gross product of agriculture total mean value reported from 2022 to 2022 for 20 countries with higher, intermediate, and lower production values countries Food and Agriculture Organization of the United Nations50
Share of the population in extreme poverty share of population living in extreme poverty mean value from 2015 to 2024 (living with less than US$2.15 per day) World Bank51

Statistics for biome and soil properties

Canonical correspondence analysis (CCA) was used to investigate the relationship between soil attribute maps (clay, organic carbon, pH, CEC, BD, and organic carbon stock) and the biome map. The CCA is an ordination technique that seeks to identify canonical axes that maximize the correlation between environmental variables (soil properties) and the distribution of groups (biomes). The model was tested for statistical significance using permutations (999 permutations), and the proportion of variance explained by the canonical axes was calculated. The analysis provided canonical correlations, explained variance, and provided a p value to assess the global significance of the relationship between biome maps and soil-attribute maps.

Results and discussion

Key soil properties at the planetary scale and the model’s performance assessment

Our global soil properties reveal spatial patterns under different environments (Figure 1) and are relevant indicators for sustaining healthy and functional soils.52 We achieve moderate to high model fit with R2 in the validation phase, varying from the lowest (0.36 for clay content) to the highest (0.57 for SOCS). The increasing RPIQ values of 0.42 (SOC) < 0.47 (CEC) < 1.21 (BD) < 1.32 (SOCS) < 1.50 (clay) < 1.86 (pH) suggest that clay and pH were modeled with the highest accuracy (Figures S1 and S2, Supplemental Note 1) for sample density and uncertainty maps.53 These values are consistent with other large-scale DSM at continental and global scales (Tables S1–S3, Supplemental Note 2).

Figure 1.

Figure 1

Maps of soil properties at 0- to 20-cm depth and 90-m resolution with modeling metrics

(A) Soil organic carbon stock (SOCS) in gigatons (Gt); (B) clay content; (C) soil organic carbon content (SOC, g kg−1); (D) pH in water; (E) cation exchange capacity (CEC), and (F) BD. R2, coefficient of determination; StdDev, standard deviation; RMSE, root-mean-square error; RPIQ, ratio of performance to interquartile distance. SOCS amount discounted icecap and rocks.

The performance of prediction models varied between soil properties due to several factors, including sampling density, covariates' importance, and latitudinal pattern (Figure S1, Supplemental Note 1).54,55 The dilemma of global soil legacy databases has been the geographical imbalance in soil observation density. Most uncertainties occurred in the Northern Hemisphere, particularly in Russian and Canadian territories (Figure S2, Supplemental Note 1), where there were fewer soil observation points. Globally, CEC maps had the greatest uncertainty and pH the lowest. The global mapping approach also observed this trend.20,56

An innovation of the present work compared to other global DSM studies is the use of a soil-reflectance proxy. This proxy is of great importance in spatializing soil data because it is a direct measurement of the topsoil reflectance.57 The Geospatial Soil Sensing System (GEOS3) demonstrated accurate results in tests29,30 and is supported by proximal sensing.27 In areas covered by vegetation, the proxy offers an indirect estimation of soil properties.,57 which DSM allowed to fill gaps due to permanent vegetation cover. We instructed researchers to visit a field in their country and conduct observations, which were then compared to the soil-property data collected by the system at the same location (Figure S4, Supplemental Note 1). This new GEOS3 system indicates the importance of creating a model based on several field inspections and different validation procedures (quantitative and qualitative, plus field experience) until reaching the best performance (Video S1, Supplemental Note 3, https://esalqgeocis.wixsite.com/english/videos-1).

The literature published over the last decades shows different predicted spatial patterns for SOCSs, suggesting large uncertainties in SOC or BD measurements and estimation methods (Table S1, Supplemental Note 2). For example, global studies reported topsoil (0–30 cm) SOCSs of 50458 and 577 Gt.23 Interestingly, these published studies found different results even using the same public legacy soil databases.59 In contrast, we created a large database, complementing publicly available legacy data with private information. The modeling reached a global SOCS of 900.1 Gt (Figures 1A and S3A) up to a depth of 0–20 cm with an R2 of 0.57. Our estimate is 32% larger than the stock provided by the FAO (Food and Agriculture Organization of the United Nations), which estimated 680 Gt of SOCS in topsoils,60 and 16% larger than that of Padarian et al. (who estimated 793 Gt for the 0- to 30-cm depth in a grid at a resolution of 500 m).15 Hiederer and Köchy58 estimated SOCSs between 710 and 1,459 Gt for the 0- to 100-cm layer, and Scharleman et al.7 reported a median of 1,437 Gt for the same depth and from 27 individual studies. We reached 900.1 Gt SOCS (0–20 cm), which is within the range estimated by other studies (Table S1, Supplemental Note 2).

We utilized a comprehensive, multi-source dataset spanning the globe, filling important data gaps. Our dataset performed a greater soil observation point density in some regions, such as South America and Asia (Figure S1 , Supplemental Note 1) than for the World Soil Information Service (WoSIS, https://www.isric.org/explore/wosis)59. For example, in Asia, we increased from 4,176 to 16,274 from open-access to open-access + private data, respectively. With this upgrade, we observed a greater variation in the SOC amplitude. This improved geographic representation suggests better reliability in the model. Previous studies have not utilized bare-soil proxies or vegetation data within ML frameworks to predict soil properties worldwide, nor have they mapped at a fine spatial resolution. Literature shows that the smaller pixel sizes result in a gain of spatial details, improving the performance of DSM.61 Despite this, prediction accuracy is inversely correlated with spatial resolution.62,63 These differences can be seen by visual inspection (Figure S4, Supplemental Note 1).

In an entire area with native vegetation in South America, FAO carbon-content maps indicate abruptly different values. Our map indicated similar values in the area, which is consistent with the same land use observed (native vegetation). The visual inspection is impacted by the improvements of a fine-resolution map, compared with the FAO (1,000 m). A consistency between land uses and SOCS (Figure S4B, Supplemental Note 1) was achieved. We also compared our data with soil grids (Figure S4A, Supplemental Note 1). Visual inspection of our maps revealed that certain regions covering South America, Europe, and Asia exhibited higher SOCS values when compared to the 250-m one. A field profile inspection from China (Figure S4C, Supplemental Note 1) presented a very deep and dark pattern caused by a high carbon horizon at the surface, which agrees with the same spot detected on the map.

We found that more than half of the world’s land area (64.5%) consists of sandy soils (soil clay content below 200 g kg−1) (including sand, loamy sand, and sandy loam), while 31.7% of areas have loamy soils and only 3.7% is clayey (clay content over 200 g kg−1). Areas with the highest clay values are in South America, reflecting its geology and greater weathering. Low clay content can be associated with sandy soils, a texture prevalent in desert areas and coastal plains (i.e., southeastern US coastal plains or soils formed in glacial till parent material). These soils typically exhibit low productivity, limiting agricultural production in extensive areas, particularly under technological constraints limitations.21

We observed the lowest soil pH values (acidic soils) in equatorial regions such as sub-Saharan Africa, South America, and Indonesia, as well as in some temperate regions of the Northern Hemisphere rich in biomass (Figure 1). Dry-hot regions such as the Sahara presented the highest pH (alkaline soils). While soil pH is largely controlled by parent rock at local and regional scales, it is also modulated by climate and anthropogenic activities. Water balance controls soil pH in dry climates, resulting in basic soils, while, in wet climates, it produces acid.53 Agricultural management, especially mineral fertilizer applications, has been attributed to a significant decrease in pH values.

A meta-analysis of agricultural soils in China found that, under the long-term application of nitrogen fertilizers in China (1980–2024), soil pH decreased by 15.2%.64 Soil acidification has been linked to increased losses in inorganic carbon and elevated carbon dioxide (CO2) emissions, degrading soil health and crop production.65 Mean pH decreased significantly by 0.3 units with afforestation in a global meta-analysis,66 suggesting that acidification is amplified through land-use change, which can cause significant depletion of microbial diversity and community composition.64 Higher BDs were found in desert areas, which have sandy soils rich in quartz. BD is used to calculate SOC stock and nutrient storage67 and, thus, has major significance in producing accurate global SOCS and SOC sequestration assessments from a local to global scale. Some areas dominated by clayey soils in Brazil presented low BDs, highlighting its relationship with soil compaction and water infiltration capacity reduction.68

The CEC is an indicator of nutrient retention, for which we calculated a mean value of 167.1 mmolc kg−1 (Figure 1). The highest nutrient-holding capacities were found in temperate environments, such as South Australia, Europe, East Asia, and North Africa. CEC usually keeps a direct relationship with soil mineralogy and fertility and represents the total capacity of the soil to hold exchangeable cations required by plants.69

Benchmarking soil properties by climatic zones

We organized the soil properties distribution in Figure 2 according to the 30 climate zones of Köppen-Geiger.46 The highest values of SOC (261.2 g kg−1) and SOCS (268.9 Gt) were found in the humid continental zone with a very cool summer, in contrast to the lowest values (25.6 g kg−1 and 13.4 Gt, respectively) in the warm Mediterranean zone (dry, mild summer). This demonstrates a close relationship between organic matter (OM) and climate, which controls the biomass input and decomposition rates.70

Figure 2.

Figure 2

Soil properties in the Köppen-Geiger climate zones46

(A) Soil organic carbon (SOC) content (blue circles), soil organic carbon stock (SOCS, orange circles); (B) cation exchange capacity (CEC, pink circles) and pH in water (green circles); and (C) clay content (brown circles) and BD (cyan circles). The respective blue, orange, and green spheres represent the SOC, SOCS, and pH differences between the cultivated areas and the native vegetation. Af, tropical rainforest; Am, tropical monsoon; Aw, tropical savanna; BWh, arid desert hot; BWk, arid desert cold; BSh, arid steppe hot; BSk, arid steppe cold; Csa, temperate dry summer hot summer; Csb, temperate dry summer warm summer; Csc, temperate dry summer cold summer; Cwa, temperate dry winter hot summer; Cwb, temperate dry winter warm summer; Cwc, temperate dry winter cold summer; Cfa, temperate no dry season hot summer; Cfb, temperate no dry season warm summer; Cfc, temperate no dry season cold summer; Dsa, cold dry summer hot summer; Dsb, cold dry summer warm summer; Dsc, cold dry summer cold summer; Dsd, cold dry summer very cold winter; Dwa, cold dry winter hot summer; Dwb, cold dry winter warm summer; Dwc, cold dry winter cold summer; Dwd, cold dry winter very cold winter; Dfa, cold no dry season hot summer; Dfb, cold no dry season warm summer; Dfc, cold no dry season cold summer; Dfd, cold no dry season very cold winter; EF, polar frost; ET, polar tundra.

The cold deserts with freezing winters had the highest mean CEC values (297.1 mmolc kg−1). The most acidic soils (pH 4.9) were found in the tropical monsoon climate zone due to geological factors and high-intensity weathering. In contrast, hot deserts with a pronounced dry winter season had a mean pH value of 7.8. The CEC and pH influenced weathering processes through base leaching and acidification.10 Croplands show higher pH values than native areas, up to 1.6 (Figure 2C). This may result from anthropogenic management practices such as fertilizing, irrigating, and liming. In Israel, there are examples of very sandy soils with high pH detected by the mapping system (Figure S5, Supplemental Note 1). Very high- or low-pH soil conditions can reduce the availability of essential plant nutrients and adversely affect plant growth.71 The physical characteristics are also driven by climate, and such patterns are observed in our results (Figure 2C). In tropical climates, soils were more clayey and with a higher BD than in temperate climates. This is in agreement with Zhao et al.72 The higher SOC content (Figure 2A) explains the low BD in temperate regions. Finally, the lower clay contents and higher BD were found in arid climates.

Distribution of soil properties in major ecoregions

The highest SOCS and SOC values were found in the boreal forest/Taiga (247.9 Gt and 250.9 g kg−1, respectively) in temperate climates at the extremes of the Northern Hemisphere (Figure 3A). Deserts, despite having the lowest SOC values due to minimal biomass inputs and high temperatures, contained a significant amount, holding 47.27 Gt. This was four times greater than mangroves, which stored 10.8 Gt but covered a much smaller area. While it is generally acknowledged that cold ecoregions had a high SOCS, followed by tropical rainforests, our findings highlight the importance of desert soils. Despite their seemingly inhospitable nature, desert regions are crucial for soil carbon storage, underscoring the importance of sustainable management practices to preserve and enhance these ecosystems. Nevertheless, the high SOCS values in some areas may be uncertain, primarily due to the modeling from cold deserts, such as in the Gobi Desert in China, which often contain rocks and other features with low SOC content (Figures 1A and 1B).

Figure 3.

Figure 3

Soil properties in 14 world ecoregions47

(A) Soil organic carbon (SOC) content (blue circles) and soil organic carbon stock (SOCS, orange circles), (B) cation exchange capacity (CEC, pink circles) and pH in water (green circles), and (C) clay content (brown circles) and BD (cyan circles). The respective blue, orange, and green spheres represent the SOC, SOCS, and pH differences between the cultivated areas and the native vegetation. Boreal F Taiga, boreal forests/Taiga; Desert X Shrub, deserts and xeric shrublands; Flood G Savanna, flooded grasslands and savannas; Med FW Scrub, Mediterranean forests, woodlands, and scrub; Montane G Shrub, montane grasslands and shrublands; Temp BM Forest, temperate broadleaf and mixed forests; Temp C Forest, temperate conifer forests; Temp GS Shrub, temperate grasslands, savannas, and shrublands; Trop Sub C Forest, tropical and subtropical coniferous forests; Trop Sub DB Forest, tropical and subtropical dry broadleaf forests; Trop Sub GS Shrub, tropical and subtropical grasslands, savannas, and shrublands; Trop Sub MB Forest, tropical and subtropical moist broadleaf forests.

Soil acidity worldwide shows distinct distribution patterns. Alkaline soils were found in arid environments (Figures 3B and S5, Supplemental Note 1), such as Iran with deserts, xeric shrublands, and Mediterranean forests. Conversely, long-time intense weathering processes in tropical ecoregions led to high chemical dissolution and leaching, resulting in acidic soils. Consistently, clay content was higher in tropical ecoregions and lower in temperate ones, such as boreal forests/Taiga and temperate broadleaf and mixed forests. This ecoregion had lower BD values than other areas due to the high OM content (Figure 3C). The association between the soil-property maps and ecoregions was consistent with the distribution of soil classes worldwide.73,74

CCA revealed a significant relationship between soil properties and the world’s biomes. The variance explained by this technique reached 55.3%, indicating that more than half of the variability in soil properties is associated with biome distribution. The canonical correlation for the first axis was 0.338, demonstrating a moderate association between biomes and soil properties, showing the influence of biomes on soil.

SOCS, SOC, and pH affected by LULC

We observed the influence of human activities, especially agriculture, on global soil properties by comparing them across different land-use and land-cover scenarios (LULCs). For example, soils under native vegetation in Asia exhibited the highest values of SOC and SOCS (Figures 4A and 4B), attributed to climate. Tropical regions, such as South America and Africa, had lower SOC and SOCS values despite the broad areas of native vegetation. The impact of human activities on OM depletion is well known. Standardizing this LULC per area unit, native vegetation regions stored 96.1 t ha−1 SOC while cropland stored 59.1 t ha−1. This implies that 37.0 t ha−1 more SOCS were stored in the soil under naturally vegetated areas compared to agricultural land.75 Our values for SOCS are consistent with the results of a similar study that estimated that SOC is lost due to land-use change. Native vegetation covered the largest area on the planet (40.73 million km2) and stored 490 Gt, while cropland stored 79 Gt of SOCS in 15.73 million km2, a difference of 411 Gt (Figure 4A). Grasslands, which encompassed natural and cultivated pasturelands, accounted for 277.0 Gt of SOCS. We highlight the significant role of natural ecosystems in carbon sequestration despite the rise of anthropogenic activities. In the native vegetation areas, the global average SOC is 121.6 g kg−1, in contrast to only 38.1 g kg−1 in croplands (Figure 4B), representing about 60% less SOC. As expected, waterlogged areas showed a significantly higher mean SOC content (250.5 g kg−1) – Figure 4B. These wetland areas are characterized by the accumulation of soil OM due to low levels of aeration, high soil moisture, occasional flooding, and the presence of a water-table fluctuations.76

Figure 4.

Figure 4

Soil properties chart of land use and land cover47 in the six continents and the world

(A) Soil organic carbon stock (SOCS, orange circles), (B) soil organic carbon content (SOC, blue circles), and (C) pH in water (green circles). The respective orange, blue, and green spheres represent SOCS, SOC, and pH differences between the cultivated areas and the native vegetation. NV, native vegetation; CL, cropland; CL-NV, cropland minus native vegetation; GL, grassland; WL, wetland; BS, bare surface.

We observed that soils under native vegetation presented lower pH values than croplands in all regions (Figure 4B). Anthropogenic activities affect soil chemistry (pH) on both sides, for good (when fertilizers are applied) and for depletion (when they are not applied). Thus, this higher pH on agricultural lands may be due to soil anthropic management (e.g., liming). Agricultural practices that make soils less acidic include liming by limestone, which increases the pH value to a suitable level.70 During this process, soil tillage is intense and exposed to degradation such as irradiation, variation of temperature, carbon oxidation, and erosion, which generates negative effects on SOC. On the other hand, most of the areas with continuous bare surfaces are found in desert or arid biomes that were predominated by rocks, low precipitation amounts, and no plowing activities. These biomes are characterized by minimal leaching, maintaining the bases in the soil profile, which explains the high pH (7.4).

The use of conservationist agriculture practices, such as no tillage, increases SOC levels77 and mitigates soil health. This strategy can protect soils from mechanical and weather impacts and maintain biota and water inside the system. Other practices should be agroforestry, biochar, and biodynamic agriculture. These practices generate long-term economic gains through ecosystem services.78 Crop diversification further boosts agroecosystem productivity by promoting the “rotation effect.”79 Figure S6 in Supplemental Note 1 shows a global SOCS map analysis across biomes47 and LULC48 by a heatmap.

A comprehensive review of 377 studies published between 1985 and 2017 found that cover crops increased cash crop yields by 1.15–7.89 times, likely due to improved soil processes.80 On the other hand, inappropriate management practices, causing hidden soil hunger, can cause yield losses exceeding 40% in cereal grains.81 Identifying antagonistic or synergistic soil-property spatialization (as presented in this paper) is key to optimizing management strategies.

Geopolitical distribution of soil properties

We observed the soil properties of the top 10 largest countries in land area per continent (Figure 5). Russia, Canada, the United States, China, Brazil, Australia, Kazakhstan, Mongolia, Argentina, India, the Democratic Republic of Congo, Indonesia, Peru, and Papua New Guinea together account for 800 Gt of SOCS, representing 75% of the total SOCS on the Earth’s surface. These values are proportional to the ones of the Global SOC Map,60 which shows 70% located in these countries (Figure S7, Supplemental Note 1). Russia and Canada had the highest mean SOC values, with more than 220 g kg−1. This reveals that the cold climates prevailing in these countries contribute to the slow degradation of OM, allowing SOC to accumulate in the soil.46

Figure 5.

Figure 5

Soil properties chart of the top 10 largest countries in land area per continent

(A) Soil organic carbon (SOC) content (blue circles) and soil organic carbon stock (SOCS, orange circles), (B) cation exchange capacity (CEC, pink circles) and pH in water (green circles), and (C) clay content (brown circles) and BD (cyan circles). The respective blue, orange, and green spheres represent the SOC, SOCS, and pH differences between the cultivated areas and the native vegetation.

Crop management and site-specific environmental conditions mainly determine the geographic variation of CEC and pH. CEC was highest in Greenland (Denmark) with 288.9 mmolc kg−1 and lowest in the Democratic Republic of Congo (Africa) with 89.5 mmolc kg−1 (Figure 5B). This figure also shows pH values to range from a mean of 4.8 in Guyana (South America) to 7.7 in Algeria and Libya (Africa), Saudi Arabia, and Iran (Asia). These results (Figures 5A and 5B) illustrate the differences in some soil properties (SOC, SOCS, and pH) between arable land and native vegetation. This arrangement emphasizes that soils from areas with native vegetation tend to have more carbon-related properties than arable lands, confirmed by visual inspection (Figures S3A and S3B, Supplemental Note 1).

Soils from specific volcanic regions showed the highest clay contents (Figure 5C), such as Cuba (284.6 g kg−1), French Polynesia (279.4 g kg−1), and Fiji (268.9 g kg−1) in Oceania. In contrast, the lowest values for clay content occur in Finland (Europe), with a mean value of only 80.5 g kg−1. The mean BD of the soils (Figure 5C) reveals high density in Saudi Arabia and Iran (approximately 1.6 gcm−3) and low density in Russia (1.1 gcm−3). We illustrate the soil-property contents in these countries, which determine soil health and fertility that subsequently drive particular agricultural and forest management.

Distribution of soil properties by socio-economic factors

Among human-induced factors, socio-economic aspects have been understudied concerning soils on a global scale.82 This overview reveals how countries utilize their resources to maximize their economic capacity and productivity. In addition, identifying soil properties that constrain capacity and productivity provides insights into where targeted management activities (i.e., soil amendments or carbon-smart management practices) impact the sustainable use of resources.

The gross domestic product (GDP) is crucial for assessing the economic size and health of a country and allows the monitoring of economic trends over time.49 We present relationships between SOCS and pH by GDP as well as SOC and CEC by GDP per capita for the top 20 richest, middle-income, and poorest countries, illustrating a second-order polynomial trendline with equations R2 and RMSE of the model for total and per capita GDP (Figures 6A and 6B). SOCSs, for instance, presented 0.56 of R2 and low RMSE (US$1.13 billion). Although there is no direct correlation between soil pH and countries' GDP (R2 = 0.05), areas in the Northern Hemisphere show less acidic soils than in the Southern. This may be explained by the predominant alkaline minerals in soils and the drier climate, constraining the water balance in this region. Water balance creates a threshold in soil pH that regulates the soil’s capacity for nutrient supply and crop productivity at the global scale.53 There was an abrupt transition from alkaline to acid soil pH when mean annual precipitation exceeded mean annual potential evapotranspiration in a global dataset of 60,291 pH measurements.

Figure 6.

Figure 6

Relationship between soil properties and the top 20 richest, intermediate, and poorest countries

(A) Decreasing order according to the gross domestic product (GDP) in billions of dollars, and (B) decreasing order by GDP per capita in thousands of dollars. The data considered the mean of the International Monetary Fund’s49 projections for GDP from 2021 to 2028. Soil organic carbon stock (SOCS, orange circles), pH in water (green circles), soil organic carbon (SOC) content (blue circles), and cation exchange capacity (CEC, pink circles). Gt, gigatons; USA, United States of America; UK, United Kingdom; Herz., Bosnia and Herzegovina; DRC, Democratic Republic of Congo; R2, coefficient of determination; RMSE, root-mean-square error.

The concept of resource quality is linked to the dynamic human settlement of the Earth, where more prosperous nations have often developed under more favorable conditions, including optimal climate conditions and soil properties critical to soil fertility, quality, health, and, thus, crop productivity. For instance, American societies emerged along rivers and water resources over time (1790–2010), suggesting that favorable water resources support the prosperity of human societies.83 In analogy to the historical coevolution of water-human systems, our study suggests that societal growth expressed through mapping spatially explicit relations between soil properties and economic success (GDP) is interlinked. In essence, this soil-human nexus provides a direct, geographically primed link between soil health and human prosperity and growth. As a consequence, regions with limited resources could benefit from soil management and tillage practices that can alter soil conditions. Especially, climate-smart management practices, such as no-tillage, reduced tillage, biochar applications, and cover crops,10 hold promise to enhance SOC, and optimize pH and CEC, boosting crop productivity and economic growth.

Secondly, we observe an exponential trend between a country’s economic strength and the level of SOCS (Figure 6A). This trend supports the hypothesis that the oldest communities have historically settled in regions with milder, cooler, and temperate climates, where SOC has accumulated in the soil over longer periods, increasing its fertility. In contrast, the world’s poorest regions are generally found in tropical areas where soil organic matter degrades rapidly.46 Many of the wealthiest nations, corresponding to the highest SOCS values, are also among the largest countries in terms of surface area. As SOCS calculations are based on surface layers, larger countries naturally have an advantage in carbon storage compared to their smaller counterparts. The classification of countries by GDP per capita was necessary to constrain the influence of country size (Figure 6B). Higher-income nations had higher SOC and CEC than the poorest ones. For example, the USA, with 90.8-Gt SOCS, had a mean SOC of 97.9 g kg−1. This country was the wealthiest in the total GDP and placed eighth considering the GDP per capita. The R2 for SOC and CEC achieved 0.26 and 0.11, with higher RMSE, demonstrating low potential to predict these socio-economic variables.

Brazil, with a per capita income of US$9,550, was in the middle of the ranking with a mean SOC of only 29.6 g kg−1 and the highest SOCS (34.2 Gt). Although Mali is the eighth-largest country in Africa, it is classified as a low-income country. Sahara Desert covers a large part of its territory, resulting in very low SOC and SOCS values (11 g kg−1 and 2.5 Gt, respectively).

To increase the discussion on this matter, we observed that the richest countries with higher agriculture gross production value (GPV) had greater SOCS and clay, as the share of the population in poverty correlated with SOC and CEC (Figures S8A and S8B, Supplemental Note 1). Results show a tendency but also indicate the complexity of understanding all the variables involved. Earlier, it was already observed that population pressure and soil quality were strongly negatively correlated in Uganda.84 However, they also found strong evidence concerning the positive correlation between population pressure and agricultural intensification. A complex discussion arises as observations indicate that rural poverty in degrading agricultural land increased in low-income countries and sub-Saharan Africa and South Asia.85

The key to avoiding these problems is to relate soil properties to economic, social, and environmental factors. According to the International Food Policy Research Institute (IFPRI),86 nearly 40%–75% of the world’s agricultural land experiences important impacts on productivity due to soil degradation. In conclusion, the spatial variability (Figure 1) and the relation of the soil properties variations with land use (Figure S6, Supplemental Note 1) shed light on this urgent discussion. It is important to note that evaluating the distribution of soil properties concerning GDP, agriculture GPV, and poverty rates are just some approaches. Other socio-economic data, such as human development and health indices, can also be examined to uncover connections between soil and various sectors.

Conclusion

A comprehensive analysis of global soil properties reveals new and significant variations in six soil properties due to heterogeneity in climate, parent materials, topography, and other factors.2 This approach enabled us to understand the connection between soil health, climatic, biotic, and socio-economic factors that inform targeted human management. We present new results on global SOC mapping using an enhanced dataset of soil and environmental covariates to inspire discussion within the global mapping community. Our soil maps (Figures 1A–1F) focus on profoundly important properties to achieve soil health,18,52,87 soil mapping,9 soil productivity,25 and circular economy in soil.88 Besides, high-resolution soil spatial information may contribute to agriculture, such as fertilization, the carbon credit market, and precision agriculture.89,90

We adopted a transdisciplinary approach that integrates soils with socio-economic and environmental domains. These insights are critical for formulating targeted soil management and conservation strategies to achieve Sustainable Development Goals (SDGs).91 For example, enhancing soil health in tropical regions with strategies to accumulate more SOC in low-fertility and marginal soils could reduce poverty and hunger by improving agricultural yields and livelihoods (No Poverty, SDG1; Zero Hunger, SDG2). Targeting practices that maintain and promote carbon sequestration in soils supports regional climate action, such as low-carbon agriculture (Climate Action, SDG13).

Since we invested substantial effort and time (5+ years) in compiling a new global soil dataset, it is expected to advance findings beyond previous global DSM-derived datasets, mainly from public legacy databases. Thus, new insights into global soil patterns are necessary. The inclusion of RS-derived global spatiotemporal Earth observation data as powerful environmental covariates to improve soil property models is an important advancement. This strategy extrapolates the model to account for the unique regional characteristics.

Future works should prioritize space-time mapping approaches to enhance dynamic soil-attribute prediction and generate comprehensive time-series maps. Dynamic tracking of soil changes due to climate variability and land-use change is hampered by scarce or inconsistent soil-sampling data. Coordinated efforts, including private collaborations to enrich global soil datasets with accurate data information, are essential. Soil minerals in the clay fraction, such as goethite and hematite, have a great influence on SOCS dynamics and sequestration potential92,93 and, thus, should be addressed in future DSM works. Finally, the present paper contributes to addressing global issues of concern outlined by the Intergovernmental Panel on Climate Change (IPCC; https://www.ipcc.ch/srccl/). Our results are poised to provide the foundation for global soil monitoring and can guide the strategic placement of soil observatories in different geographic regions.

Thus, we successfully developed the first high-resolution global soil grids for key soil properties, with a 90-m pixel size, underpinned by a robust and new Earth observation system. This achievement provides a valuable resource for researchers and policymakers, offering improved spatial detail information. Furthermore, by integrating these soil properties with socio-environmental perspectives, we have gone beyond traditional soil-mapping studies and laid the foundation for more informed, context-sensitive decision-making in land management, environmental conservation, and sustainable development. Integrating multi-temporal satellite proxies, such as bare soil and vegetation, is a crucial technique for achieving fine-scale, global coverage without gaps in ground data. This method effectively modeled key soil properties with excellent predictive performance, taking into account the vast and diverse soil-environmental conditions around the world.

The findings demonstrated that 64% of the world’s topsoils are predominantly sandy, with restricted fertility and susceptible to degradation-limiting crop productivity and highlighting potential risks to food security. We identified that the mean SOC content in cultivated soils is at least 60% lower than under natural vegetation. Global soil pH values were 5.5 in native vegetation and 6.6 in croplands, reflecting the application of fertility soil-management practices. We also revealed that the top 10 largest countries per continent in terms of land area collectively store 75% of the global SOCS, emphasizing that such soil wealth represents a form of land capital closely linked to national economic strength. Overall, we underscore the need to integrate soil and socio-economic factors to bolster carbon marketing, conservation, and climate-smart practices to improve soil health and to discuss sustainability, resilience, and food and human security. This approach represents a significant, challenging step forward in these fields and also paves the way for a deeper understanding of the complex interactions between soil systems and socio-environmental factors.

Resource availability

Material availability

This study did not generate new unique material/reagents.

Data and code availability

The maps generated in this study are publicly available and can be accessed online at https://esalqgeocis.wixsite.com/english/jose-a-m-dematte-global-soil-properties-90m. They can also be accessed and downloaded via the Earth Engine API at https://ee-geocis.projects.earthengine.app/view/global-soil-properties. Any code used in this study is available from the corresponding author upon reasonable request.

Funding and Acknowledgments

This research was supported by the São Paulo Research Foundation (FAPESP) under grants 2014-22262-0 and 2021/05129-8 as well as the from the Center for Carbon Research in Tropical Agriculture (CCARBON) at the University of São Paulo, under grant 2021/10573-4. We also must mention the support of the MII Project of the Russian Federation (reg. 123030300031-6). The first author acknowledges CNPq for research scholarship (307190-2021-8). We extend our gratitude to international researchers and institutions for contributing soil data, especially the US Department of Agriculture Natural Resources Conservation Service for their coordination, collection, and provision of soil sample data in the USA. We thank the GeoCiS research group (https://esalqgeocis.wixsite.com/english) for their technical support.

Author contributions

Conceptualization, J.A.M.D. and R.R.P.; Formal analysis, R.R.P., J.J.M.N., and N.A.R.; Methodology, R.R.P. and J.J.M.N.; Software, R.R.P. and J.J.M.N.; Visualization, R.R.P., J.J.M.N., M.T.A.A., and N.A.R.; Data curation, J.A.M.D., R.R.P., J.J.M.N., and N.A.R.; Roles/Writing - original draft, J.A.M.D., J.J.M.N., and N.A.R.; Writing - review & editing, J.A.M.D., R.R.P., J.J.M.N., N.A.R., B.M., I.Y.S., S.G., S.C., Y.H., J.H., S.C.C., Q.d.J.v.L., E.B., C.G., Z.G., M.T.A.A., L.G.V., J.T.F.R., R.M., A.G., A.V.Z., J.P.C., Y.M., H.J.J., R., C.W., R.R., N.T., N.Ts., M.K., D.N.K., S.D., Y.G., E.V., S.A., J.K.M.B., A.B., S.N.M., N.A.H., E.K., D.C.M., M.R.F., E.S.M.S., and A.A.; Validation, J.A.M.D., R.R.P., N.A.R., J.J.M.N.; Funding acquisition, Resources, Supervision, J.A.M.D. All authors contributed to the manuscript and approved the final version.

Declaration of interests

The authors declare no conflict of interest.

Published Online: June 9, 2025

Footnotes

Supplemental information

Document S1. Figures S1–S8, Tables S1–S3, and Notes S1–S3
mmc1.pdf (2.8MB, pdf)
Document S2. Article plus supplemental information
mmc2.pdf (41.3MB, pdf)

References

  • 1.Bardgett R.D., van der Putten W.H. Belowground biodiversity and ecosystem functioning. Nature. 2014;515:505–511. doi: 10.1038/nature13855. [DOI] [PubMed] [Google Scholar]
  • 2.Lal R., Bouma J., Brevik E., et al. Soils and sustainable development goals of the United Nations: An International Union of Soil Sciences perspective. Geoderma Reg. 2021;25 doi: 10.1016/j.geodrs.2021.e00398. [DOI] [Google Scholar]
  • 3.Eekhout J.P.C., de Vente J. Global impact of climate change on soil erosion and potential for adaptation through soil conservation. Earth Sci. Rev. 2022;226 doi: 10.1016/j.earscirev.2022.103921. [DOI] [Google Scholar]
  • 4.Guerra C.A., Berdugo M., Eldridge D.J., et al. Global hotspots for soil nature conservation. Nature. 2022;610:693–698. doi: 10.1038/s41586-022-05292-x. [DOI] [PubMed] [Google Scholar]
  • 5.Friedlingstein P., O’Sullivan M., Jones M.W., et al. Global Carbon Budget 2023. Earth Syst. Sci. Data. 2023;15:5301–5369. doi: 10.5194/essd-15-5301-2023. [DOI] [Google Scholar]
  • 6.Georgiou K., Jackson R.B., Vindušková O., et al. Global stocks and capacity of mineral-associated soil organic carbon. Nat. Commun. 2022;13:3797. doi: 10.1038/s41467-022-31540-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Scharlemann J.P., Tanner E.V., Hiederer R., et al. Global soil carbon: Understanding and managing the largest terrestrial carbon pool. Carbon Manag. 2014;5:81–91. doi: 10.4155/cmt.13.77. [DOI] [Google Scholar]
  • 8.Wang F., Harindintwali J.D., Wei K., et al. Climate change: Strategies for mitigation and adaptation. Innov. Geosci. 2023;1 doi: 10.59717/j.xinn-geo.2023.100015. [DOI] [Google Scholar]
  • 9.McBratney A., Field D.J., Koch A. The dimensions of soil security. Geoderma. 2014;213:203–213. doi: 10.1016/j.geoderma.2013.08.013. [DOI] [Google Scholar]
  • 10.Paustian K., Lehmann J., Ogle S., et al. Climate-smart soils. Nature. 2016;532:49–57. doi: 10.1038/nature17174. [DOI] [PubMed] [Google Scholar]
  • 11.Vereecken H., Amelung W., Bauke S.L., et al. Soil hydrology in the Earth system. Nat. Rev. Earth Environ. 2022;3:573–587. doi: 10.1038/s43017-022-00324-6. [DOI] [Google Scholar]
  • 12.Boretti A., Rosa L. Reassessing the projections of the World Water Development Report. npj Clean Water. 2019;2 doi: 10.1038/s41545-019-0039-9. [DOI] [Google Scholar]
  • 13.Grunwald S. Artificial intelligence and soil carbon modeling demystified: Power, potentials, and perils. Carbon Footprints. 2022;1:6. doi: 10.20517/cf.2022.03. [DOI] [Google Scholar]
  • 14.Padarian J., Stockmann U., Minasny B., et al. Monitoring changes in global soil organic carbon stocks from space. Remote Sens. Environ. 2022;281 doi: 10.1016/j.rse.2022.113260. [DOI] [Google Scholar]
  • 15.Padarian J., Minasny B., McBratney A., et al. Soil carbon sequestration potential in global Croplands. PeerJ. 2022;10 doi: 10.7717/peerj.13740. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Lembrechts J.J., van den Hoogen J., Aalto J., et al. Global maps of soil temperature. Glob. Change Biol. 2022;28:3110–3144. doi: 10.1111/gcb.16060. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Lenton T.M., Xu C., Abrams J.F., et al. Quantifying the human cost of global warming. Nat. Sustain. 2023;6:1237–1247. doi: 10.1038/s41893-023-01132-6. [DOI] [Google Scholar]
  • 18.Poppiel R.R., Cherubin M.R., Novais J.J.M., et al. Soil health in Latin America and the Caribbean. Commun. Earth Environ. 2025;6:141. doi: 10.1038/s43247-025-02021-w. [DOI] [Google Scholar]
  • 19.Batjes N.H. Harmonized soil property values for broad-scale modeling (WISE30sec) with estimates of global soil carbon stocks. Geoderma. 2016;269:61–68. doi: 10.1016/j.geoderma.2016.01.034. [DOI] [Google Scholar]
  • 20.Poggio L., de Sousa L.M., Batjes N.H., et al. SoilGrids 2.0: producing soil information for the globe with quantified spatial uncertainty. Soil. 2021;7:217–240. doi: 10.5194/soil-7-217-2021. [DOI] [Google Scholar]
  • 21.Lotfollahi L., Delavar M.A., Biswas A., et al. Modeling the Spatial Distribution of Sand, Silt, and Clay Particles Based on Global Soil Map and Limited Data. J. Desert Res. 2023;23:123–137. doi: 10.22059/jdesert.2023.95533. [DOI] [Google Scholar]
  • 22.Xiang D., Wang G., Tian J., et al. Global patterns and edaphic-climatic controls of soil carbon decomposition kinetics predicted from incubation experiments. Nat. Commun. 2023;14:2171. doi: 10.1038/s41467-023-37900-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Lin Z., Dai Y., Mishra U., et al. Global and regional soil organic carbon estimates: Magnitude and uncertainties. Pedosphere. 2024;34:685–698. doi: 10.1016/j.pedsph.2023.06.005. [DOI] [Google Scholar]
  • 24.Dynarski K.A., Bossio D.A., Scow K.M. Dynamic stability of soil carbon: Reassessing the “permanence” of soil carbon sequestration. Front. Environ. Sci. 2020;8:1–14. doi: 10.3389/fenvs.2020.514701. [DOI] [Google Scholar]
  • 25.Greschuk L.T., Demattê J.A.M., Silvero N.E.Q., et al. A soil productivity system reveals most Brazilian agricultural lands are below their maximum potential. Sci. Rep. 2023;13 doi: 10.1038/s41598-023-39981-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Evangelista S.J., Field D.J., McBratney A.B., et al. A proposal for the assessment of soil security: Soil functions, soil services and threats to soil. Soil Secur. 2023;10 doi: 10.1016/j.soisec.2023.100086. [DOI] [Google Scholar]
  • 27.Peng Y., Ben-Dor E., Biswas A., et al. Spectroscopic solutions for generating new global soil information. Innovation. 2025;6 doi: 10.1016/j.xinn.2025.100839. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Rogge D., Bauer A., Zeidler J., et al. Building an exposed soil composite processor (SCMaP) for mapping spatial and temporal characteristics of soils with Landsat imagery (1984–2014) Remote Sens. Environ. 2018;205:1–17. doi: 10.1016/j.rse.2017.11.004. [DOI] [Google Scholar]
  • 29.Demattê J.A.M., Fongaro C.T., Rizzo R., et al. Geospatial Soil Sensing System (GEOS3): A powerful data mining procedure to retrieve soil spectral reflectance from satellite images. Remote Sens. Environ. 2018;212:161–175. doi: 10.1016/j.rse.2018.04.047. [DOI] [Google Scholar]
  • 30.Xue J., Zhang X., Huang Y., et al. A two-dimensional bare soil separation framework using multi-temporal Sentinel-2 images across China. Int. J. Appl. Earth Obs. Geoinf. 2024;134 doi: 10.1016/j.jag.2024.104181. [DOI] [Google Scholar]
  • 31.Durán J., Delgado-Baquerizo M. Vegetation structure determines the spatial variability of soil biodiversity across biomes. Sci. Rep. 2020;10 doi: 10.1038/s41598-020-78483-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Arrouays D., Poggio L., Salazar Guerrero O.A., et al. Digital soil mapping and GlobalSoilMap: Main advances and ways forward. Geoderma Reg. 2020;21 doi: 10.1016/j.geodrs.2020.e00265. [DOI] [Google Scholar]
  • 33.Barman K.G., Caron S., Claassen T., et al. Towards a benchmark for scientific understanding in humans and machines. Minds Mach. 2024;34 doi: 10.1007/s11023-024-09657-1. [DOI] [Google Scholar]
  • 34.Bishop T.F.A., McBratney A.B., Laslett G.M. Modelling soil attribute depth functions with equal-area quadratic smoothing splines. Geoderma. 1999;91:27–45. doi: 10.1016/S0016-7061(99)00003-8. [DOI] [Google Scholar]
  • 35.Chen S., Dai L., Shuai Q., et al. Global Soil Bulk Density DataBase (GSBDDB) (V1.0) Zenodo. 2022 doi: 10.5281/zenodo.7075158. [DOI] [Google Scholar]
  • 36.McBratney A.B., Mendonça Santos M.L., Minasny B. On digital soil mapping. Geoderma. 2003;117:3–52. doi: 10.1016/S0016-7061(03)00223-4. [DOI] [Google Scholar]
  • 37.Gorelick N., Hancher M., Dixon M., et al. Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sens. Environ. 2017;202:18–27. doi: 10.1016/j.rse.2017.06.031. [DOI] [Google Scholar]
  • 38.Demattê J.A.M., Rizzo R., Rosin N.A., et al. A global soil spectral grid based on space sensing. Sci. Total Environ. 2025;968 doi: 10.1016/j.scitotenv.2025.178791. [DOI] [PubMed] [Google Scholar]
  • 39.Karger D.N., Conrad O., Böhner J., et al. Climatologies at high resolution for the Earth land surface areas. Sci. Data. 2017;4 doi: 10.1038/sdata.2017.122. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Safanelli J., Poppiel R., Ruiz L., et al. Terrain analysis in Google Earth Engine: A method adapted for high-performance global-scale analysis. ISPRS Int. J. GeoInf. 2020;9:400. doi: 10.3390/ijgi9060400. [DOI] [Google Scholar]
  • 41.Theobald D.M., Harrison-Atlas D., Monahan W.B., et al. Ecologically-relevant maps of landforms and physiographic diversity for climate adaptation planning. PLoS One. 2015;10 doi: 10.1371/journal.pone.0143619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Yamazaki D., Ikeshima D., Tawatari R., et al. A high accuracy map of global terrain elevations. Geophys. Res. Lett. 2017;44:5844–5853. doi: 10.1002/2017GL072874. [DOI] [Google Scholar]
  • 43.Behrens T., Schmidt K., Viscarra Rossel R.A., et al. Spatial modelling with Euclidean distance fields and machine learning. Eur. J. Soil Sci. 2018;69:757–770. doi: 10.1111/ejss.12687. [DOI] [Google Scholar]
  • 44.van den Hoogen J., Gerten D., van Bodegom P.M., et al. A geospatial mapping pipeline for ecologists. bioRxiv. 2021 https://www.researchgate.net/publication/353150945_A_geospatial_mapping_pipeline_for_ecologists Preprint at. [Google Scholar]
  • 45.Breiman L. Random Forests. Mach. Learn. 2001;45:5–32. doi: 10.1023/A:1010933404324. [DOI] [Google Scholar]
  • 46.Beck H.E., McVicar T.R., Vergopolan N., et al. High-resolution (1 km) Köppen-Geiger maps for 1901–2099 based on constrained CMIP6 projections. Sci. Data. 2023;10:724. doi: 10.1038/s41597-023-02549-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Olson D.M., Dinerstein E., Wikramanayake E.D., et al. Terrestrial ecoregions of the world: A new map of life on Earth. Bioscience. 2001;51:933. doi: 10.4236/oje.2022.123014. [DOI] [Google Scholar]
  • 48.Zanaga D., Van De Kerchove R., Daems D., et al. ESA WorldCover 10 m 2021 v200. Zenodo. 2022 doi: 10.5281/zenodo.7254220. [DOI] [Google Scholar]
  • 49.International Monetary Fund – IMF World Economic Outlook Database, October 2023. 2023. https://www.imf.org/en/Publications/WEO/weo-database/2023/October
  • 50.Food and Agriculture Organization of the United Nations Gross production value of the agricultural sector. 2024. https://ourworldindata.org/grapher/value-of-agricultural-production Our World in Data (with major processing)
  • 51.World Bank Share of population living in extreme poverty. 2024. https://ourworldindata.org/poverty Our World in Data (with major processing)
  • 52.Lehmann J., Bossio D.A., Kögel-Knabner I., et al. The concept and future prospects of soil health. Nat. Rev. Earth Environ. 2020;1:544–553. doi: 10.1038/s43017-020-0080-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Slessarev E.W., Lin Y., Bingham N.L., et al. Water balance creates a threshold in soil pH at the global scale. Nature. 2016;540:567–569. doi: 10.1038/nature20139. https://www.nature.com/articles/nature20139 [DOI] [PubMed] [Google Scholar]
  • 54.Brus D.J. Sampling for digital soil mapping: A tutorial supported by R scripts. Geoderma. 2019;338:464–480. doi: 10.1016/j.geoderma.2018.07.036. [DOI] [Google Scholar]
  • 55.Somarathna P.D.S.N., Minasny B., Malone B.P. More data or a better model? Figuring out what matters most for the spatial prediction of soil carbon. Soil Sci. Soc. Am. J. 2017;81:1413–1426. doi: 10.2136/sssaj2016.11.0376. [DOI] [Google Scholar]
  • 56.Hengl T., de Jesus J.M., MacMillan R.A., et al. SoilGrids1km — Global Soil Information Based on Automated Mapping. PLoS One. 2014;9 doi: 10.1371/journal.pone.0105992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.van Wesemael B., Abdelbaki A., Ben-Dor E., et al. A European soil organic carbon monitoring system leveraging Sentinel 2 imagery and the LUCAS soil database. Geoderma. 2024;452 doi: 10.1016/j.geoderma.2024.117113. [DOI] [Google Scholar]
  • 58.Hiederer R., Köchy M. Publ. Off. Eur. Union; 2011. Global Soil Organic Carbon Estimates and the Harmonized World Soil Database. [DOI] [Google Scholar]
  • 59.Batjes N.H., Ribeiro E., van Oostrum A. Standardised soil profile data to support global mapping and modelling (WoSIS snapshot 2019) Earth Syst. Sci. Data. 2020;12:299–320. doi: 10.5194/essd-12-299-2020. [DOI] [Google Scholar]
  • 60.United Nations Global Soil Organic Carbon Map (GSOCmap) FAO Tech. Rep. 2018;1:162. https://www.fao.org/3/I8891EN/i8891en.pdf [Google Scholar]
  • 61.Piedallu C., Pedersoli E., Chaste E., et al. Optimal resolution of soil properties maps varies according to their geographical extent and location. Geoderma. 2022;412 doi: 10.1016/j.geoderma.2022.115723. [DOI] [Google Scholar]
  • 62.Garosi Y., Ayoubi S., Nussbaum M., et al. Effects of different sources and spatial resolutions of environmental covariates on prediction of soil organic carbon using machine learning in a semi-arid region of Iran. Geoderma Reg. 2022;29 doi: 10.1016/j.geodrs.2022.e00513. [DOI] [Google Scholar]
  • 63.Guo L., Shi T., Linderman M., et al. Exploring the influence of spatial resolution on the digital mapping of soil organic carbon by airborne hyperspectral VNIR Imaging. Remote Sens. 2019;11:1032. doi: 10.3390/rs11091032. [DOI] [Google Scholar]
  • 64.Zhang S., Zhu Q., de Vries W., et al. Effects of soil amendments on soil acidity and crop yields in acidic soils: A world-wide meta-analysis. J. Environ. Manage. 2023;345 doi: 10.1016/j.jenvman.2023.118531. [DOI] [PubMed] [Google Scholar]
  • 65.Raza S., Zamanian K., Ullah S., et al. Inorganic carbon losses by soil acidification jeopardize global efforts on carbon sequestration and climate change mitigation. J. Clean. Prod. 2021;315 doi: 10.1016/j.jclepro.2021.128036. [DOI] [Google Scholar]
  • 66.Berthrong S.T., Jobbágy E.G., Jackson R.B. A global meta-analysis of soil exchangeable cations, pH, carbon, and nitrogen with afforestation. Ecol. Appl. 2009;19:2228–2241. doi: 10.1890/08-1730.1. [DOI] [PubMed] [Google Scholar]
  • 67.Xu L., He N., Yu G. Methods of evaluating soil bulk density: Impact on estimating large scale soil organic carbon storage. Catena. 2016;144:94–101. doi: 10.1016/j.catena.2016.05.001. [DOI] [Google Scholar]
  • 68.AL-Shammary A.A.G., Kouzani A.Z., Kaynak A., et al. Soil Bulk Density Estimation Methods: A Review. Pedosphere. 2018;28:581–596. doi: 10.1016/S1002-0160(18)60034-7. [DOI] [Google Scholar]
  • 69.Khaledian Y., Brevik E.C., Pereira P., et al. Modeling soil cation exchange capacity in multiple countries. Catena. 2017;158:194–200. doi: 10.1016/j.catena.2017.07.002. [DOI] [Google Scholar]
  • 70.Stockmann U., Padarian J., McBratney A., et al. Global soil organic carbon assessment. Global Food Secur. 2015;6:9–16. doi: 10.1016/j.gfs.2015.07.001. [DOI] [Google Scholar]
  • 71.Neina D. The role of soil pH in plant nutrition and soil remediation. Appl. Environ. Soil Sci. 2019;2019:1–9. doi: 10.1155/2019/5794869. [DOI] [Google Scholar]
  • 72.Zhao X., Yang Y., Shen H., et al. Global soil–climate–biome diagram: linking surface soil properties to climate and biota. Biogeosciences. 2019;16:2857–2871. doi: 10.5194/bg-16-2857-2019. [DOI] [Google Scholar]
  • 73.Powell T.L., Davis T.S., Cousins A.B. Global influence of soil texture on ecosystem water limitation. Nature. 2021;593:549–552. doi: 10.1038/s41586-024-08089-2. [DOI] [Google Scholar]
  • 74.Turek M.E., Poggio L., Batjes N.H., et al. Global mapping of volumetric water retention at 100, 330, and 15 000 cm suction using the WoSIS database. Int. Soil Water Conserv. Res. 2023;11:225–239. doi: 10.1016/j.iswcr.2022.08.001. [DOI] [Google Scholar]
  • 75.Sanderman J., Hengl T., Fiske G.J. Soil carbon debt of 12,000 years of human land use. Proc. Natl. Acad. Sci. USA. 2017;114:9575–9580. doi: 10.1073/pnas.1706103114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 76.Bansal S., Creed I.F., Tangen B.A., et al. Practical guide to measuring wetland carbon pools and fluxes. Wetlands. 2023;43:105. doi: 10.1007/s13157-023-01722-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 77.de Freitas I.C., Ribeiro J.M., Araújo N.C.A., et al. Agrosilvopastoral systems and well-managed pastures increase soil carbon stocks in the Brazilian Cerrado. Rangel. Ecol. Manag. 2020;73:776–785. doi: 10.1016/j.rama.2020.08.001. [DOI] [Google Scholar]
  • 78.Koudahe K., Allen S.C., Djaman K. Critical review of the impact of cover crops on soil properties. Int. Soil Water Conserv. Res. 2022;10:343–354. doi: 10.1016/j.iswcr.2022.03.003. [DOI] [Google Scholar]
  • 79.Bowles T.M., Mooshammer M., Socolar Y., et al. Long-term evidence shows that crop rotation diversification increases agricultural resilience to adverse growing conditions in North America. One Earth. 2020;2:284–293. doi: 10.1016/j.oneear.2020.02.007. [DOI] [Google Scholar]
  • 80.Daryanto S., Fu B., Wang L., et al. Quantitative synthesis on the ecosystem services of cover crops. Earth Sci. Rev. 2018;185:357–373. doi: 10.1016/j.earscirev.2018.06.013. [DOI] [Google Scholar]
  • 81.Noulas C., Tziouvalekas M., Karyotis T. Zinc in soils, water and food crops. J. Trace Elem. Med. Biol. 2018;49:252–260. doi: 10.1016/j.jtemb.2018.02.009. [DOI] [PubMed] [Google Scholar]
  • 82.Grunwald S. Grand challenges in pedometrics-AI research. Front. Soil Sci. 2021;1:1–8. doi: 10.3389/fsoil.2021.714323. [DOI] [Google Scholar]
  • 83.Fang Y., Jawitz J.W. The evolution of human population distance to water in the USA from 1790 to 2010. Nat. Commun. 2019;10:430. doi: 10.1038/s41467-019-08366-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 84.Mugizi F.M.P., Matsumoto T. A curse or a blessing? Population pressure and soil quality in Sub-Saharan Africa: Evidence from rural Uganda. Ecol. Econ. 2021;179 doi: 10.1016/j.ecolecon.2020.106851. [DOI] [Google Scholar]
  • 85.Barbier E.B., Hochard J.P. Land degradation and poverty. Nat. Sustain. 2018;1:623–631. doi: 10.1038/s41893-018-0155-4. [DOI] [Google Scholar]
  • 86.Food and Agriculture Organization of the United Nations . Vol. 1. FAO; Rome: 2023. p. 150. (The State of Food and Agriculture 2023: Revealing the True Cost of Food to Transform Agrifood Systems). [DOI] [Google Scholar]
  • 87.Novais J.J.M., Melo B.M.D., Neves Junior A.F., et al. Online analysis of Amazon’s soils through reflectance spectroscopy and cloud computing can support policies and the sustainable development. J. Environ. Manage. 2025;375 doi: 10.1016/j.jenvman.2025.124155. [DOI] [PubMed] [Google Scholar]
  • 88.Nwaogu C., Minasny B., Field D.J., et al. Conceptualizing core aspects of circular economy in soil: A critical review and analysis. Crit. Rev. Environ. Sci. Technol. 2025;55:805–835. doi: 10.1080/10643389.2025.2454689. [DOI] [Google Scholar]
  • 89.Lawrence P.G., Roper W., Morris T.F., et al. Guiding soil sampling strategies using classical and spatial statistics: A review. Agron. J. 2020;112:493–510. doi: 10.1002/agj2.20048. [DOI] [Google Scholar]
  • 90.Padhiary M., Saha D., Kumar R., et al. Enhancing precision agriculture: A comprehensive review of machine learning and AI vision applications in all-terrain vehicle for farm automation. Smart Agric. Technol. 2024;8 doi: 10.1016/j.atech.2024.100483. [DOI] [Google Scholar]
  • 91.United Nations. Transforming Our World . Vol. 1. United Nations; New York: 2015. p. 41.https://sdgs.un.org/publications/transforming-our-world-2030-agenda-sustainable-development-17981 (The 2030 Agenda for Sustainable Development). [Google Scholar]
  • 92.Wieder W.R., Bonan G.B., Allison S.D. Global soil carbon projections are improved by modelling microbial processes. Nat. Clim. Change. 2013;3:909–912. doi: 10.1038/nclimate1951. [DOI] [Google Scholar]
  • 93.Rodríguez-Albarracín H.S., Demattê J.A.M., Rosin N.A., et al. Soil organic carbon sequestration potential explained by mineralogical and microbiological activity using spectral transfer functions. Sci. Total Environ. 2024;947 doi: 10.1016/j.scitotenv.2024.174652. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Figures S1–S8, Tables S1–S3, and Notes S1–S3
mmc1.pdf (2.8MB, pdf)
Document S2. Article plus supplemental information
mmc2.pdf (41.3MB, pdf)

Data Availability Statement

The maps generated in this study are publicly available and can be accessed online at https://esalqgeocis.wixsite.com/english/jose-a-m-dematte-global-soil-properties-90m. They can also be accessed and downloaded via the Earth Engine API at https://ee-geocis.projects.earthengine.app/view/global-soil-properties. Any code used in this study is available from the corresponding author upon reasonable request.


Articles from The Innovation are provided here courtesy of Elsevier

RESOURCES