Search

The turnover time (τ) of global soil organic carbon is central to the functioning of terrestrial ecosystems. Yet our spatially-explicit understanding of depth-dependent variations and environmental controls of τ at a global scale remain incomplete. In this study, we combine multiple state-of-the-art observation-based datasets, including over ninety thousand geo-referenced soil profiles, the latest root observations distributed globally, and large amounts of satellite-derived environmental variables, to generate global maps of apparent τ in topsoil (0–0.3 m) and subsoil (0.3–1 m) layers with a spatial resolution of 30 arcsec (~1 km at the Equator). We show that subsoil τ (385203485 years [mean with a variation range from 2.5th to 97.5th percentile]) is over eight times longer than topsoil τ (1511137 years). The cross-validation shows that the fitted machine learning models effectively captured the variabilities in τ, with R 2 values of 0.87 and 0.70 for topsoil and subsoil τ mapping, respectively. The prediction uncertainties of the τ maps were quantified for better user applications. The environmental controls on top- and subsoil τ were investigated at global, biome, and local scales. Our analyses illustrate that how temperature, water availability, physio-chemical properties and depth exert jointly impacts on τ. The data-driven approaches allow us to identify their interactions, thereby enriching our comprehension of mechanisms driving nonlinear τ–environment relationships from global to local scales. The distributions of dominating factors of τ at local scales were mapped for identifying context-dependent controls on τ across different regions. We further reveal that the current Earth system models may underestimate τ by comparing model-derived maps with our observation-derived τ maps. The resulting maps with new insights demonstrated in this study facilitate the future modelling efforts of carbon cycle–climate feedbacks and supporting effective carbon management. The dataset is archived and freely available at https://doi.org/10.5281/zenodo.14560239 (Zhang, 2025). [ABSTRACT FROM AUTHOR]

In digital soil mapping, modelling soil thickness poses a challenge due to the prevalent issue of right‐censored data. This means that the true soil thickness exceeds the depth of sampling, and neglecting to account for the censored nature of the data can lead to poor model performance and underestimation of the true soil thickness. Survival analysis is a well‐established domain of statistical modelling that can deal with censored data. The random survival forest is a notable example of a survival‐related machine learning approach used to address right‐censored soil property data in digital soil mapping. Previous studies that employed this model either focused on mapping the probability of soil thickness exceeding certain depths, and thereby not mapping soil thickness itself, or dismissed it due to perceived poor performance. In this study, we propose an alternative survival model to map soil thickness that is based on the inverse probability of censoring weighting. In this approach, calibration data are weighted by the inverse of the probability that soil thickness exceeds a certain depth, that is, a survival probability. These weights can then be used with most machine learning models. We used the weights with a regular random forest, and compared it with a random survival forest, and other strategies for handling right‐censored data, through a comprehensive synthetic simulation study and two real‐world case studies. The results suggest that the weighted random forest model produces competitive predictions, establishing it as a viable option for mapping right‐censored soil property data. [ABSTRACT FROM AUTHOR]

In response to the growing societal awareness of the critical role of healthy soils, there has been an increasing demand for accurate and high-resolution soil information to inform national policies and support sustainable land management decisions. Despite advancements in digital soil mapping and initiatives like GlobalSoilMap, quantifying soil variability and its uncertainty across space, depth and time remains a challenge. Therefore, maps of key soil properties are often still missing on a national scale, which is also the case in the Netherlands. To meet this challenge and fill this data gap, we introduce BIS-4D, a high-resolution soil modeling and mapping platform for the Netherlands. BIS-4D delivers maps of soil texture (clay, silt and sand content), bulk density, pH, total nitrogen, oxalate-extractable phosphorus, cation exchange capacity and their uncertainties at 25 m resolution between 0 and 2 m depth in 3D space. Additionally, it provides maps of soil organic matter and its uncertainty in 3D space and time between 1953 and 2023 at the same resolution and depth range. The statistical model uses machine learning informed by soil observations amounting to between 3815 and 855 950, depending on the soil property, and 366 environmental covariates. We assess the accuracy of mean and median predictions using design-based statistical inference of a probability sample and location-grouped 10-fold cross validation (CV) and prediction uncertainty using the prediction interval coverage probability. We found that the accuracy of clay, sand and pH maps was the highest, with the model efficiency coefficient (MEC) ranging between 0.6 and 0.92 depending on depth. Silt, bulk density, soil organic matter, total nitrogen and cation exchange capacity (MEC of 0.27 to 0.78), and especially oxalate-extractable phosphorus (MEC of -0.11 to 0.38) were more difficult to predict. One of the main limitations of BIS-4D is that prediction maps cannot be used to quantify the uncertainty in spatial aggregates. We provide an example of good practice to help users decide whether BIS-4D is suitable for their intended purpose. An overview of all maps and their uncertainties can be found in the Supplement. Openly available code and input data enhance reproducibility and help with future updates. BIS-4D prediction maps can be readily downloaded at 10.4121/0c934ac6-2e95-4422-8360-d3a802766c71. BIS-4D fills the previous data gap of the national-scale GlobalSoilMap product in the Netherlands and will hopefully facilitate the inclusion of soil spatial variability as a routine and integral part of decision support systems. [ABSTRACT FROM AUTHOR]

In response to the growing societal awareness of the critical role of healthy soils, there is an increasing demand for accurate and high-resolution soil information to inform national policies and support sustainable land management decisions. Despite advancements in digital soil mapping and initiatives like GlobalSoilMap, quantifying soil variability and its uncertainty across space, depth, and time remains a challenge. Therefore, maps of key soil properties are often still missing on a national scale, which is also the case in the Netherlands. To meet this challenge and fill this data gap, we introduce BIS-4D, a high resolution soil modelling and mapping platform for the Netherlands. BIS-4D delivers maps of soil texture (clay, silt and sand content), bulk density, pH, total nitrogen, oxalate-extractable phosphorus, cation exchange capacity and their uncertainties at 25m resolution between 0 - 2m depth in 3D space. Additionally, it provides maps of soil organic matter and its uncertainty in 3D space and time between 1953 - 2023 at the same resolution and depth range. The statistical model uses machine learning informed by soil observations numbering between 3815 - 855 950, depending on the soil property, and 366 environmental covariates. We assess the accuracy of mean and median predictions using design-based statistical inference of a probability sample and location-grouped 10-fold cross-validation, and prediction uncertainty using the prediction interval coverage probability. We found that the accuracy of clay, sand and pH maps was highest, with the model efficiency coefficient (MEC) ranging between 0.6 - 0.92 depending on depth. Silt, bulk density, soil organic matter, total nitrogen and cation exchange capacity (MEC = 0.27 - 0.78), and especially oxalate-extractable phosphorus (MEC = -0.11 - 0.38), were more difficult to predict. One of the main limitations of BIS-4D is that prediction maps cannot be used to quantify the uncertainty of spatial aggregates. A step-by-step manual helps users decide whether BIS-4D is suitable for their intended purpose, an overview of all maps and their uncertainties can be found in the supplementary information (SI), openly available code and input data enhance reproducibility and future updates, and BIS-4D prediction maps can be easily downloaded at https://doi.org/10.4121/0c934ac6-2e95-4422-8360-d3a802766c71 (Helfenstein et al., 2024a). BIS-4D fills the previous data gap of a national scale GlobalSoilMap product in the Netherlands and will hopefully facilitate the inclusion of soil spatial variability as a routine and integral part of decision support systems. [ABSTRACT FROM AUTHOR]

Spatial soil applications frequently involve binomial variables. If relevant environmental covariates are available, using a Bayesian generalized linear model (BGLM) might be a solution for mapping such discrete soil properties. The geostatistical extension, a Bayesian generalized linear geostatistical model (BGLGM), adds spatial dependence and is thus potentially better equipped. The objective of this work was to evaluate whether it pays off to extend from a BGLM to a BGLGM for mapping binary soil properties, evaluated in terms of prediction accuracy and modelling complexity. As motivating example, we mapped the presence/absence of the Pleistocene sand layer within 120 cm from the land surface in the Dutch province of Flevoland, using the BGLGM implementation in the R‐package geoRglm. We found that BGLGM yields considerably better statistical validation metrics compared to a BGLM, especially with – as in our case – a large (n = 1,000) observation sample and few relevant covariates available. Also, the inferred posterior BGLGM parameters enable the quantification of spatial relationships. However, calibrating and applying a BGLGM is quite demanding with respect to the minimal required sample size, tuning the algorithm, and computational costs. We replaced manual tuning by an automated tuning algorithm (which eases implementing applications) and found a sample composition that delivers meaningful results within 50 h calculation time. With the gained insights and shared scripts spatial soil practitioners and researchers can – for their specific cases – evaluate if using BGLGM is feasible and if the extra gain is worth the extra effort. Highlights: Does adding spatial correlation to a Bayesian GLM for mapping a binary soil variable pay off?We aim to make spatial Bayesian hierarchical modelling accessible for pedometricians.Most hierarchical models work well when enough observations are provided, even without covariates.Including spatial correlation might sometimes be worth the extra effort and computational costs. [ABSTRACT FROM AUTHOR]

There is a growing demand for high‐quality soil data. However, soil measurements are subject to many error sources. We aimed to quantify uncertainties in synthetic and real‐world wet chemistry soil data through a linear mixed‐effects model, including batch and laboratory effects. The use of synthetic data allowed us to investigate how accurately the model parameters were estimated for various experimental measurement designs, whereas the real‐world case served to explore if estimates of the random effect variances were still accurate for unbalanced datasets with few replicates. The variance estimates for synthetic pHH2O data were unbiased, but limited laboratory information led to imprecise estimates. The same was observed for unbalanced synthetic datasets, where 20, 50 and 80% of the data were removed randomly. Removal led to a sharp increase of the interquartile range (IQR) of the variance estimates for batch effect and the residual. The model was also fitted to real‐world pHH2O and total organic carbon (TOC) data, provided by the Wageningen Evaluating Programmes for Analytical Laboratories (WEPAL). For pHH2O, the model yielded unbiased estimates with relatively small IQRs. However, the limited number of batches with replicate measurements (5.8%) caused the batch effect to be larger than expected. A strong negative correlation between batch effect and residual variance suggested that the model could not distinguish well between these two random effects. For TOC, batch effect was removed from the model as no replicates were available within batches. Again, unbiased model estimates were obtained. However, the IQRs were relatively large, which could be attributed to the smaller dataset with only a single replicate measurement. Our findings demonstrated the importance of experimental measurement design and replicate measurements in the quantification of uncertainties in wet chemistry soil data. Highlights: Accurate uncertainty quantification depends on the experimental measurement design.Linear mixed‐effects models can be used as a tool to quantify uncertainty in wet chemistry soil data.Lack of replicate measurements leads to poor estimates of error variance components.Measurement error in wet chemistry soil data should not be ignored. [ABSTRACT FROM AUTHOR]

Spatially resolved estimates of change in soil organic carbon (SOC) stocks are necessary for supporting national and international policies aimed at achieving land degradation neutrality and climate change mitigation. In this work we report on the development, implementation and application of a data‐driven, statistical method for mapping SOC stocks in space and time, using Argentina as a pilot. We used quantile regression forest machine learning to predict annual SOC stock at 0–30 cm depth at 250 m resolution for Argentina between 1982 and 2017. The model was calibrated using over 5,000 SOC stock values from the 36‐year time period and 35 environmental covariates. We preprocessed normalized difference vegetation index (NDVI) dynamic covariates using a temporal low‐pass filter to allow the SOC stock for a given year to depend on the NDVI of the current as well as preceding years. Predictions had modest temporal variation, with an average decrease for the entire country from 2.55 to 2.48 kg C m−2 over the 36‐year period (equivalent to a decline of 211 Gg C, 3.0% of the total 0–30 cm SOC stock in Argentina). The Pampa region had a larger estimated SOC stock decrease from 4.62 to 4.34 kg C m−2 (5.9%) during the same period. For the 2001–2015 period, predicted temporal variation was seven‐fold larger than that obtained using the Tier 1 approach of the Intergovernmental Panel on Climate Change and United Nations Convention to Combat Desertification. Prediction uncertainties turned out to be substantial, mainly due to the limited number and poor spatial and temporal distribution of the calibration data, and the limited explanatory power of the covariates. Cross‐validation confirmed that SOC stock prediction accuracy was limited, with a mean error of 0.03 kg C m−2 and a root mean squared error of 2.04 kg C m−2. In spite of the large uncertainties, this work showed that machine learning methods can be used for space–time SOC mapping and may yield valuable information to land managers and policymakers, provided that SOC observation density in space and time is sufficiently large. Highlights: We tested the use of machine learning for space–time mapping of soil organic carbon (SOC) stock.Predictions for Argentina from 1982 to 2017 showed a 3% decrease of the topsoil SOC stock over time.The machine learning model predicted a greater temporal variation than the IPCC Tier 1 approach.Accurate machine learning SOC stock prediction requires dense soil sampling in space and time. [ABSTRACT FROM AUTHOR]