Using process-oriented model output to enhance machine learning-based soil organic carbon prediction in space and time

Monitoring and modelling soil organic carbon (SOC) in space and time can help us to better understand soil carbon dynamics and is of key importance to support climate change research and policy. Although machine learning (ML) has attracted a lot of attention in the digital soil mapping (DSM) community for its powerful ability to learn from data and predict soil properties, such as SOC, it is better at capturing soil spatial variation than soil temporal dynamics. By contrast, process-oriented (PO) models benefit from mechanistic knowledge to express physiochemical and biological processes that govern SOC temporal changes. Therefore, integrating PO and ML models seems a promising means to represent physically plausible SOC dynamics while retaining the spatial prediction accuracy of ML models. In this study, a hybrid modelling framework was developed and tested for predicting topsoil SOC stock in space and time for a regional cropland area located in eastern China. In essence, the hybrid model uses predictions of the PO model in unsampled years as additional training data of the ML model, with a weighting parameter assigned to balance the importance of SOC values from the PO model and real measurements. The results indicated that temporal trends of SOC stock modelled by PO and ML models were largely different, while they were notably similar between the PO and hybrid models. Cross-validation showed that the hybrid model had the best performance (RMSE = 0.29 kg m-2), with a 19 % improvement compared with the ML model. We conclude that the proposed hybrid framework not only enhances space-time soil carbon mapping in terms of prediction accuracy and physical plausibility, it also provides insights for soil management and policy decisions in the face of future climate change and intensified human activities.

Tree-based algorithms for spatial modeling of soil particle distribution in arid and semi-arid region

Accurate estimation of particle size distribution across a large area is crucial for proper soil management and conservation, ensuring compatibility with capabilities and enabling better selection and adaptation of precision agricultural techniques. The study investigated the performance of tree-based models, ranging from simpler options like CART to sophisticated ones like XGBoost, in predicting soil texture over a wide geographic region. Models were constructed using remotely sensed plant and soil indexes as covariates. Variable selection employed the Boruta approach. Training and testing data for machine learning models consisted of particle size distribution results from 622 surface soil samples collected in southeastern Turkey. The XGBoostClay model emerged as the most accurate predictor, with an R2 value of 0.74. Its superiority was further underlined by a 21.36% relative improvement in XGBoostClay RMSE compared to RFClay and 44.5% compared to CARTClay. Similarly, the R2 values for XGBoostSilt and XGBoostSand models reached 0.71 and 0.75 in predicting sand and silt content, respectively. Among the considered covariates, the normalized ratio vegetation index and slope angle had the highest impact on clay content (21%), followed by topographic position index and simple ratio clay index (20%), while terrain ruggedness index had the least impact (18%). These results highlight the effectiveness of Boruta approach in selecting an adequate number of variables for digital mapping, suggesting its potential as a viable option in this field. Furthermore, the findings of this study suggest that remote sensing data can effectively contribute to digital soil mapping, with tree-based model development leading to improved prediction performance.

Spatial prediction of lime requirements by adjusting aluminium saturation in Sub-Saharan Africa croplands

Acidic soils cover over 30 % of Sub-Saharan Africa cropland. Acidic soils deprive crops of calcium, magnesium, potassium, molybdenum, and phosphorus due to aluminium (Al), manganese, and iron toxicities. Thus, liming is required to adjust the level of exchangeable Al3+ to the desired level of Al saturation of the crops grown. Lime requirement (LR) was quantified using soil dataset from Africa soil information service (AfSIS). Spatial variations of LR of cereals, pulses and cash crops were predicted using random forest algorithm. Our results revealed that mean of LR Mg CaCO3 (1 Mg = 106 g) ha-1 for cereal crops were 6.34, 6.35, and 4.41 for maize, sorghum, and upland rice, respectively. Mean of LR (Mg ha-1) for pulses were 6.28, 5.19, and 4.90 for common beans, soybeans, and cowpeas, respectively. Mean of LR Mg CaCO3 (1 Mg = 106 g) ha-1 for cash crops were 3.41 and 6.29 for coffee and cotton, respectively. Spatial variation showed that LR in croplands was higher in tropical humid regions than in semi-arid and arid regions and ranged from 0 to 8.8 Mg ha-1. The results of 10-fold cross validation for high model performance of LR for tested crops were coefficient of determination (R2) of 0.61, a root mean square error (RMSE) of 0.5, and a mean absolute error (MAE) of 0.31, maize LR with RMSE = 0.9, MAE = 0.24, and R2 = 0.51, and cotton LR with RMSE = 0.5, MAE = 0.31, and R2 = 0.60. We recommend predicting lime requirement in acidic soils of Sub-Saharan Africa by adjusting Al saturation up to the tolerance of the grown crop, updating soil surveys in Sab Saharan Africa, and using digital soil mapping to monitor soil acidity and lime requirement.

Improved soil carbon stock spatial prediction in a Mediterranean soil erosion site through robust machine learning techniques

Soil serves as a reservoir for organic carbon stock, which indicates soil quality and fertility within the terrestrial ecosystem. Therefore, it is crucial to comprehend the spatial distribution of soil organic carbon stock (SOCS) and the factors influencing it to achieve sustainable practices and ensure soil health. Thus, the present study aimed to apply four machine learning (ML) models, namely, random forest (RF), k-nearest neighbors (kNN), support vector machine (SVM), and Cubist model tree (Cubist), to improve the prediction of SOCS in the Srou catchment located in the Upper Oum Er-Rbia watershed, Morocco. From an inventory of 120 sample points, 80% were used for training the model, with the remaining 20% set aside for model testing. Boruta's algorithm and the multicollinearity test identified only nine (9) factors as the controlling factors selected as input data for predicting SOCS. As a result, spatial distribution maps for SOCS were generated for all models, then compared, and further validated using statistical metrics. Among the models tested, the RF model exhibited the best performance (R2 = 0.76, RMSE = 0.52 Mg C/ha, NRMSE = 0.13, and MAE = 0.34 Mg C/ha), followed closely by the SVM model (R2 = 0.68, RMSE = 0.59 Mg C/ha, NRMSE = 0.15, and MAE = 0.34 Mg C/ha) and Cubist model (R2 = 0.64, RMSE = 0.63 Mg C/ha, NRMSE = 0.16, and MAE = 0.43 Mg C/ha), while the kNN model had the lowest performance (R2 = 0.31, RMSE = 0.94 Mg C/ha, NRMSE = 0.24, and MAE = 0.63 Mg C/ha). However, bulk density, pH, electrical conductivity, and calcium carbonate were the most important factors for spatially predicting SOCS in this semi-arid region. Hence, the methodology used in this study, which relies on ML algorithms, holds the potential for modeling and mapping SOCS and soil properties in comparable contexts elsewhere.

Unleashing the sequestration potential of soil organic carbon under climate and land use change scenarios in Danish agroecosystems

Future global climate changes are expected to increase soil organic carbon (SOC) decomposition. However, the combined effect of C inputs, land use changes, and climate on SOC turnover is still unclear. Exploring this SOC-climate-land use interaction allows us to understand the SOC stabilization mechanisms and examine whether the soil can act as a source or a sink for CO2. The current study estimates the SOC sequestration potential in the topsoil layer of Danish agricultural lands by 2038, considering the effect of land use change and future climate scenarios using the Rothamsted Carbon (RothC) model. Additionally, we quantified the loss vulnerability of existing and projected SOC based on the soil capacity to stabilize OC. We used the quantile random forest model to estimate the initial SOC stock by 2018, and we simulated the SOC sequestration potential with RothC for a business-as-usual (BAU) scenario and a crop rotation change (LUC) scenario under climate change conditions by 2038. We compared the projected SOC stocks with the carbon saturation deficit. The initial SOC stock ranged from 10 to 181 Mg C ha-1 in different parts of the country. The projections showed a SOC loss of 8.1 Mg C ha-1 for the BAU scenario and 6 Mg C ha-1 after the LUC adoption. This SOC loss was strongly influenced by warmer temperatures and clay content. The proposed crop rotation became a mitigation measure against the negative effect of climate change on SOC accumulation, especially in sandy soils with a high livestock density. A high C accumulation in C-saturated soils suggests an increase in non-complexed SOC, which is vulnerable to being lost into the atmosphere as CO2. With these results, we provide information to prioritize areas where different soil management practices can be adopted to enhance SOC sequestration in stable forms and preserve the labile-existing SOC stocks.

Soil organic carbon stock retrieval from Sentinel-2A using a hybrid approach

Digital soil maps find application in numerous fields, making their accuracy a crucial factor. Mapping soil properties in homogeneous landscapes where the soil surface is concealed, as in forests, presents a complex challenge. In this study, we evaluated the spatial distribution of soil organic carbon stocks (SOCstock) under forest vegetation using three methods: regression kriging (RK), random forest (RF), and RF combined with ordinary kriging of residuals (RFOK) in combination with Sentinel-2A satellite data. We also compared their accuracies and identified key influencing factors. We determined that SOCstock ranged from 0.6 to 10.9 kg/m2 with an average value of 4.9 kg/m2. Among the modelling approaches, we found that the RFOK exhibited the highest accuracy (RMSE = 1.58 kg/m2, NSE = 0.33), while the RK demonstrated a lack of spatial correlation of residuals, rendering this method inapplicable. An analysis of variable importance revealed that the SWIR B12 band of the Sentinel-2A satellite contributed the most to RFOK predictions. We concluded that the RFOK hybrid approach outperformed the others, potentially serving as a foundation for digital soil mapping under similar environmental conditions. Therefore, it is essential to consider spatial correlations when mapping soil properties in ecosystems that are inaccessible for capturing the spectral response of the soil surface.

Digital mapping of soil organic carbon density in China using an ensemble model

The soil organic carbon stock (SOCS) is considered as one of the largest carbon reservoirs in terrestrial ecosystems, and small changes in soil can cause significant changes in atmospheric CO2 concentration. Understanding organic carbon accumulation in soils is crucial if China is to meet its dual carbon target. In this study, the soil organic carbon density (SOCD) in China was digitally mapped using an ensemble machine learning (ML) model. First, based on SOCD data obtained at depths of 0-20 cm from 4356 sampling points (15 environmental covariates), we compared the performance of four ML models, namely random forest (RF), extreme gradient boosting (XGBoost), support vector machine (SVM), and artificial neural network (ANN) models, in terms of coefficient of determination (R2), mean absolute error (MAE), and root mean square error (RMSE) values. Then, we ensembled four models using Voting Regressor and the principle of stacking. The results showed that ensemble model (EM) accuracy was high (RMSE = 1.29, R2 = 0.85, MAE = 0.81), so that it could be a good choice for future research. Finally, the EM was used to predict the spatial distribution of SOCD in China, which ranged from 0.63 to 13.79 kg C/m2 (average = 4.09 (±1.90) kg C/m2). The SOC storage amount in surface soil (0-20 cm) was 39.40 Pg C. This study developed a novel, ensemble ML model for SOC prediction, and improved our understanding of the spatial distribution of SOC in China.

Effects of optical and radar satellite observations within Google Earth Engine on soil organic carbon prediction models in Spain

The modeling and mapping of soil organic carbon (SOC) has advanced through the rapid growth of Earth observation data (e.g., Sentinel) collection and the advent of appropriate tools such as the Google Earth Engine (GEE). However, the effects of differing optical and radar sensors on SOC prediction models remain uncertain. This research aims to investigate the effects of different optical and radar sensors (Sentinel-1/2/3 and ALOS-2) on SOC prediction models based on long-term satellite observations on the GEE platform. We also evaluate the relative impact of four synthetic aperture radar (SAR) acquisition configurations (polarization mode, band frequency, orbital direction and time window) on SOC mapping with multiband SAR data from Spain. Twelve experiments involving different satellite data configurations, combined with 4027 soil samples, were used for building SOC random forest regression models. The results show that the synthesis mode and choice of satellite images, as well as the SAR acquisition configurations, influenced the model accuracy to varying degrees. Models based on SAR data involving cross-polarization, multiple time periods and "ASCENDING" orbits outperformed those involving copolarization, a single time period and "DESCENDING" orbits. Moreover, combining information from different orbital directions and polarization modes improved the soil prediction models. Among the SOC models based on long-term satellite observations, the Sentinel-3-based models (R² = 0.40) performed the best, while the ALOS-2-based model performed the worst. In addition, the predictive performance of MSI/Sentinel-2 (R² = 0.35) was comparable with that of SAR/Sentinel-1 (R² = 0.35); however, the combination (R² = 0.39) of the two improved the model performance. All the predicted maps involving Sentinel satellites had similar spatial patterns that were higher in northwest Spain and lower in the south. Overall, this study provides insights into the effects of different optical and radar sensors and radar system parameters on soil prediction models and improves our understanding of the potential of Sentinels in developing soil carbon mapping.

Modeling the spatial variation of calcium carbonate equivalent to depth using machine learning techniques

Inorganic carbon is the largest source of carbon in terrestrial surface, particularly in arid and semiarid regions, including the Chahardowli Plain in western Iran. Inorganic carbon plays an equal or greater role than organic soil carbon in these areas, although less attention has been paid in quantifying their variability. The objective of this study was to model and map calcium carbonate equivalent (CCE) presenting inorganic carbon in soil using machine learning and digital soil mapping techniques. Chahardowli Plain in foothills of the Zagros Mountains in the southeast of Kurdistan Province in Iran was taken as a case study area. CCE was measured at 0-5, 5-15, 15-30, 30-60, and 60-100 cm depths following GloalSoilMap.net project specifications. A total of 145 samples were collected from 30 soil profiles using the conditional Latin hypercube (cLHS) method of sampling. Relationships between CCE and environmental predictors were modeled using random forest (RF) and decision tree (DT) models. In general, the RF model performed slightly superior than the DT model. The mean value of CCE increased with soil depth, from 3.5% (0-5 cm) to 63.8% (30-60 cm). Remote sensing (RS) variables and terrestrial variables were equally important. The importance of RS variables was higher at the surface than terrestrial variables, and vice versa. The most significant variables were Channel Network Base Level (CNBL) variable and Difference Vegetation Index (DVI) with the same variable importance value (21.1%). In areas affected by river activities, the use of the CNBL and vertical distance to channel networks (VDCN) as variables in digital soil mapping (DSM) could increase the accuracy of soil property prediction maps. The VDCN played a principal role in soil distribution in the study area by affecting the rate of discharge and, thus, erosion and sedimentation. A high percentage of carbonate in parts of the region could exacerbate nutrient deficiencies for most crops and provide information for sustainably managing agricultural activity.

Synergetic use of DEM derivatives, Sentinel-1 and Sentinel-2 data for mapping soil properties of a sloped cropland based on a two-step ensemble learning method

Understanding the spatial variability of soil organic matter (SOM), soil total nitrogen (STN), soil total phosphorus (STP), and soil total potassium (STK) is important to support site-specific agronomic management, food production, and climate change adaptation. High-resolution remote sensing imageries have emerged as an innovative solution to investigate the spatial variation in agricultural soils with machine learning (ML) algorithms. However, the predictive power of the individual and combined effects of Sentinel-1 (S1) synthetic aperture radar (SAR) and Sentinel-2 (S2) multispectral images for mapping soil properties, especially STN, STP, and STK, have rarely been investigated. Moreover, single ML model may achieve unstable performance for predicting multiple soil properties due to strong spatial heterogeneity. This study explored the combine use of S1, S2, and DEM derivatives to map SOM, STN, STP, and STK content of a sloped cropland of northeastern China. A two-step method with a weighted sum of four ML models was proposed to improve the accuracy and robustness in predicting multiple soil properties. Our results showed that single ML model has various performance in predicting the four soil properties. The optimal ML models could explain approximately 56 %, 53 %, 56 % and 37 % of the variability of SOM, STN, STP, and STK, respectively. Using the weights estimated through a 10-fold cross-validation procedure, the two-step ensemble learning model was retrained and showed more robust performance than the four ML models, in which the prediction accuracy was improved by 2.38 %, 1.40 %, 3.52 %, and 3.29 % for SOM, STN, STP, and STK, respectively. Our results also showed that the optical S2 derived features, especially the two S2 short-wave infrared bands, enhanced vegetation index, and soil adjusted vegetation index, were more important for soil property prediction than S1 data and DEM derivatives. Compared with individual sensor, a combination of S1 and S2 data yielded more accurate predictions of STN and STP but not for SOM and STK. The results of this study highlight the potential of high-resolution S1 and S2 data and the two-step method for soil property prediction at farmland scale.

Incorporating agricultural practices in digital mapping improves prediction of cropland soil organic carbon content: The case of the Tuojiang River Basin

Accurate mapping of soil organic carbon (SOC) in cropland is essential for improving soil management in agriculture and assessing the potential of different strategies aiming at climate change mitigation. Cropland management practices have large impacts on agricultural soils, but have rarely been considered in previous SOC mapping work. In this study, cropland management practices including carbon input (CI), length of cultivation (LC), and irrigation (Irri) were incorporated as agricultural management covariates and integrated with natural variables to predict the spatial distribution of SOC using the Extreme Gradient Boosting (XGBoost) model. Additionally, we evaluated the performance of incorporating agricultural management practice variables in the prediction of cropland topsoil SOC. A case study was carried out in a traditional agricultural area in the Tuojiang River Basin, China. We found that CI was the most important environmental covariate for predicting cropland SOC. Adding cropland management practices to natural variables improved prediction accuracy, with the coefficient of determination (R²), the root mean squared error (RMSE) and Lin's Concordance Correlation Coefficient (LCCC) improving by 16.67%, 17.75% and 5.62%, respectively. Our results highlight the effectiveness of incorporating agricultural management practice information into SOC prediction models. We conclude that the construction of spatio-temporal database of agricultural management practices derived from inventories is a research priority to improve the reliability of SOC model prediction.

Digital mapping of soil pH and carbonates at the European scale using environmental variables and machine learning

Soil pH and carbonates (CaCO₃) are important indicators of soil chemistry and fertility, and the prediction of their spatial distribution is critical for the agronomic and environmental management. Digital soil mapping (DSM) techniques are widely accepted for the geospatial analysis of the soil properties. They are rapid and cost-efficient approaches that can provide quantitative prediction. However, the digital mapping of soil pH and CaCO₃ are not well studied, especially at a continental scale. In this research, we mapped the soil pH and CaCO₃ at the European scale using multisource environmental variables and machine learning approaches. Moderate Resolution Imaging Spectroradiometer (MODIS) products, terrain attributes, and climatic variables were considered. Meanwhile, nine machine learning algorithms, namely, three linear and six nonlinear models, were used for the spatial prediction of soil pH and CaCO₃. The land use and cover area frame statistical survey (LUCAS) 2015 topsoil dataset provided by the European Soil Data Centre was utilised. The performances of different models were compared and analysed in terms of coefficient of determination (R²), root mean square error (RMSE), and ratio of performance to deviation (RPD). Specifically, nonlinear machine learning models outperformed the linear ones, and extremely randomized trees (ERT) gave the most satisfactory result for soil pH (R² = 0.70, RMSE = 0.75, and RPD = 1.84) and CaCO₃ (R² = 0.53, RMSE = 93.49 g/kg, and RPD = 1.46). The results revealed that MODIS products and climatic variables were important in predicting soil pH and CaCO₃. Moreover, spatial distribution of soil pH and CaCO₃ in Europe were mapped at 250 m resolution, and the areas with high CaCO₃ content always showed high soil pH value.

Comparison and application of SOFM, fuzzy c-means and k-means clustering algorithms for natural soil environment regionalization in China

Soil attributes and their environmental drivers exhibit different patterns in different geographical directions, along with distinct regional characteristics, which may have important effects on substance migration and transformation such as organic matter and soil elements or the environmental impacts of pollutants. Therefore, regional soil characteristics should be considered in the process of regionalization for environmental management. However, no comprehensive evaluation or systematic classification of the natural soil environment has been established for China. Here, we established an index system for natural soil environmental regionalization (NSER) by combining literature data obtained based on bibliometrics with the analytic hierarchy process (AHP). Based on the index system, we collected spatial distribution data for 14 indexes at the national scale. In addition, three clustering algorithms-self-organizing feature mapping (SOFM), fuzzy c-means (FCM) and k-means (KM)-were used to classify and define the natural soil environment. We imported four cluster validity indexes (CVI) to evaluate different models: Davies-Bouldin index (DB), Silhouette index (Sil) and Calinski-Harabasz index (CH) for FCM and KM, clustering quality index (CQI) for SOFM. Analysis and comparison of the results showed that when the number of clusters was 13, the FCM clustering algorithm achieved the optimal clustering results (DB = 1.16, Sil = 0.78, CH = 6.77 × 10⁶), allowing the natural soil environment of China to be divided into 12 regions with distinct characteristics. Our study provides a set of comprehensive scientific research methods for regionalization research based on spatial data, it has important reference value for improving soil environmental management based on local conditions in China.

Machine learning based estimation of field-scale daily, high resolution, multi-depth soil moisture for the Western and Midwestern United States

Background

High-resolution soil moisture estimates are critical for planning water management and assessing environmental quality. In-situ measurements alone are too costly to support the spatial and temporal resolutions needed for water management. Recent efforts have combined calibration data with machine learning algorithms to fill the gap where high resolution moisture estimates are lacking at the field scale. This study aimed to provide calibrated soil moisture models and methodology for generating gridded estimates of soil moisture at multiple depths, according to user-defined temporal periods, spatial resolution and extent.

Methods

We applied nearly one million national library soil moisture records from over 100 sites, spanning the U.S. Midwest and West, to build Quantile Random Forest (QRF) calibration models. The QRF models were built on covariates including soil moisture estimates from North American Land Data Assimilation System (NLDAS), soil properties, climate variables, digital elevation models, and remote sensing-derived indices. We also explored an alternative approach that adopted a regionalized calibration dataset for the Western U.S. The broad-scale QRF models were independently validated according to sampling depths, land cover type, and observation period. We then explored the model performance improved with local samples used for spiking. Finally, the QRF models were applied to estimate soil moisture at the field scale where evaluation was carried out to check estimated temporal and spatial patterns.

Results

The broad-scale QRF model showed moderate performance (R² = 0.53, RMSE = 0.078 m³/m³) when data points from all depth layers (up to 100 cm) were considered for an independent validation. Elevation, NLDAS-derived moisture, soil properties, and sampling depth were ranked as the most important covariates. The best model performance was observed for forest and pasture sites (R² > 0.5; RMSE < 0.09 m³/m³), followed by grassland and cropland (R² > 0.4; RMSE < 0.11 m³/m³). Model performance decreased with sampling depths and was slightly lower during the winter months. Spiking the national QRF model with local samples improved model performance by reducing the RMSE to less than 0.05 m³/m³ for grassland sites. At the field scale, model estimates illustrated more accurate temporal trends for surface than subsurface soil layers. Model estimated spatial patterns need to be further improved and validated with management data.

Conclusions

The model accuracy for top 0-20 cm soil depth (R² > 0.5, RMSE < 0.08 m³/m³) showed promise for adopting the methodology for soil moisture monitoring. The success of spiking the national model with local samples showed the need to collect multi-year high frequency (e.g., hourly) sensor-based field measurements to improve estimates of soil moisture for a longer time period. Future work should improve model performance for deeper depths with additional hydraulic properties and use of locally-selected calibration datasets.

Exploring the potential of UAV hyperspectral image for estimating soil salinity: Effects of optimal band combination algorithm and random forest

Hyperspectral remote sensing by unmanned aerial vehicle (UAV) is an important technical tool for rapid, accurate, and real-time monitoring of soil salinity in arid zone agroecosystems. However, the key to effective soil salinity (electrical conductivity, EC) prediction by UAV visible and near-infrared (Vis-NIR) spectroscopy depends on the selection of effective features selection techniques and robust prediction characteristics algorithms. Therefore, in this study, two advanced feature selection methods and two commonly used modeling methods were applied to predict and characterize the spatial patterns of soil salinity (EC). The aim of this study was to explore the predictive performance of different feature band selection methods and to identify a robust soil salinity mapping strategy. The results demonstrated that standard normal variate (SNV) pre-processing broadened the absorption characteristics of the spectrum. Compared with competitive adaptive reweighted sampling (CARS), the optimal band combination algorithm (OBCA) strengthened the correlation with soil salinity and had a higher variable importance in the modeling. Random forest (RF) was more stable in mapping the spatial pattern of surface soil salinity compared to the partial least squares regression model (PLSR). Our results confirm the effectiveness of OBCA and RF in the developing UAV remote sensing models for surface soil salinity estimation and mapping.

ALOS-2 L-band SAR backscatter data improves the estimation and temporal transferability of wildfire effects on soil properties under different post-fire vegetation responses

Remote sensing techniques are of particular interest for monitoring wildfire effects on soil properties, which may be highly context-dependent in large and heterogeneous burned landscapes. Despite the physical sense of synthetic aperture radar (SAR) backscatter data for characterizing soil spatial variability in burned areas, this approach remains completely unexplored. This study aimed to evaluate the performance of SAR backscatter data in C-band (Sentinel-1) and L-band (ALOS-2) for monitoring fire effects on soil organic carbon and nutrients (total nitrogen and available phosphorous) at short term in a heterogeneous Mediterranean landscape mosaic made of shrublands and forests that was affected by a large wildfire. The ability of SAR backscatter coefficients and several band transformations of both sensors for retrieving soil properties measured in the field in immediate post-fire situation (one month after fire) was tested through a model averaging approach. The temporal transferability of SAR-based models from one month to one year after wildfire was also evaluated, which allowed to assess short-term changes in soil properties at large scale as a function of pre-fire plant community type. The retrieval of soil properties in immediate post-fire conditions featured a higher overall fit and predictive capacity from ALOS-2 L-band SAR backscatter data than from Sentinel-1 C-band SAR data, with the absence of noticeable under and overestimation effects. The transferability of the ALOS-2 based model to one year after wildfire exhibited similar performance to that of the model calibration scenario (immediate post-fire conditions). Soil organic carbon and available phosphorous content was significantly higher one year after wildfire than immediately after the fire disturbance. Conversely, the short-term change in soil total nitrogen was ecosystem-dependent. Our results support the applicability of L-band SAR backscatter data for monitoring short-term variability of fire effects on soil properties, reducing data gathering costs within large and heterogeneous burned landscapes.

Identification of soil type in Pakistan using remote sensing and machine learning

Soil study plays a significant role in the cultivation of crops. To increase the productivity of any crop, one must know the soil type and properties of that soil. The conventional soil type identification, grid sampling and hydrometer method require expert intervention, more time and extensive laboratory experimentation. Digital soil mapping, while applying remote sensing, offers soil type information and has rapidity, low cost, and spatial resolution advantages. This study proposes a model to identify the soil type using remote sensing data. Spectral data of the Upper Indus Plain of Pakistan Pothwar region and Doabs were acquired using fifteen Landsat eight images dated between June 2020 to August 2020. Bare soil images were obtained to identify the soil type classes Silt Loam, Loam, Sandy Loam, Silty Clay Loam and Clay Loam. Spectral data of band values, reflectance band values, corrective reflectance band values and vegetation indices are practiced studying the reflectance factor of soil type. Regarding multi-class classification, Random Forest and Support Vector Machine are two popular techniques used in the research community. In the present work, we used these two techniques aided with Logistic Model Tree with 10-fold cross-validation. The classification with the best performance is achieved using the spectral data, with an overall accuracy of 86.61% and 84.41% for the Random Forest and Logistic Model Tree classification, respectively. These results may be applied for crop cultivation in specific areas and assist decision-makers in better agricultural planning.

Accuracy and uncertainty of geostatistical models versus machine learning for digital mapping of soil calcium and potassium

Accuracy and uncertainty of models used for digital soil mapping are important for assessing confidence of predictions and reliable land use planning and management. In this study, two approaches of geostatistical (spatial) and machine learning (ML) models were evaluated for predictive mapping of soil calcium (Ca) and potassium (K). Two spatial models including empirical Bayesian kriging (EBK) and sequential Gaussian simulation (SGS) were compared with machine learning models: Cubist, random forest (RF) and support vector machine (SVM) in terms of their accuracy and uncertainty for mapping soil Ca and K. The study area is in Nowley, New South Wales, Australia, with an area of 2083 ha and a variety of soil types and farming systems. For the models training process, 240 soil samples data and for validation 102 independent samples data were used. For accuracy assessment R², root mean square error (RMSE), concordance and bias and for uncertainty assessment confidence limits were investigated. Also, in order to compare the outcomes for the two soil properties with different measurement units, mean absolute percentage error (MAPE) and relative uncertainty (RU) as accuracy and uncertainty measures, respectively, were evaluated. Results showed that for K map SGS had the highest R² (0.74) and lowest RMSE (1.96), followed by EBK with R² = 0.72 and RMSE = 2.02. For Ca map, EBK model showed the highest accuracy (R² = 0.46; RMSE = 3.21), followed by SVM and SGS with comparable accuracies. Comparing the two soil properties, Ca map showed higher MAPE and RU, compared to K map. The lowest MAPE was obtained for EBK model (for K = 39) and SGS model (for K = 44). Also, the lowest RU values were found for EBK and SGS models. Among the ML models, SVM showed lower sensitivity to higher variance in data input. In general, the spatial models outperformed the ML models with regard to both accuracy and uncertainty. An additional conclusion is that considering the data variance in the two soil properties, geostatistical models with lower RU and MAPE were relatively less susceptible to data variance, compared to ML models.

Effects of hyperspectral data with different spectral resolutions on the estimation of soil heavy metal content: From ground-based and airborne data to satellite-simulated data

Soil heavy metal distribution maps can provide decision-making information for pollution control and agricultural management. However, the estimation of heavy metals is sensitive to the resolution of the soil spectra due to their sparse content in soils. The purposes of this study were to test the sensitivity of Ni, Zn and Pb prediction results to variations in spectral resolution, then to map their spatial distributions over a large area. In addition, the effectiveness of spectral feature extraction was investigated. In total, 92 soil samples and corresponding field soil spectra were obtained from the Tongwei-Zhuanglang area in Gansu Province, China. Airborne HyMap hyperspectral image of this area was acquired simultaneously. Three satellite image spectra (AHSI_GF-5, Hyperion, AHSI_{ZY-1 02D}) were simulated using the field spectra which were measured under real environmental conditions rather than laboratory conditions. The combination of genetic algorithm and partial least squares regression (GA-PLSR) was used as prediction algorithm. The models calibrated by HyMap image full spectral bands had the highest accuracies (R_P² = 0.8558, 0.8002, and 0.8592 for Ni, Zn, and Pb, respectively) because of high consistency. For field spectra and three simulated satellite spectra, models calibrated by simulated AHSI_GF-5 spectra performed best because of appropriate resolution (5 nm in the visible near-infrared [VNIR] and 10 nm in the short-wave infrared [SWIR]). The spectral feature extraction method only improved prediction accuracy of the field spectra, indicating that this method benefited from higher spectral resolution. The mapping of the spatial distribution of soil heavy metals over a large area was realized based on HyMap image. According to the results of the satellite simulation spectra, this study proposes to use GF-5 hyperspectral image to estimate heavy metals content. The outcomes provide a reference for the utilization of aerial and satellite hyperspectral images in prediction of soil heavy metal concentrations.

Soil carbon sequestration potential in global croplands

Improving the amount of organic carbon in soils is an attractive alternative to partially mitigate climate change. However, the amount of carbon that can be potentially added to the soil is still being debated, and there is a lack of information on additional storage potential on global cropland. Soil organic carbon (SOC) sequestration potential is region-specific and conditioned by climate and management but most global estimates use fixed accumulation rates or time frames. In this study, we model SOC storage potential as a function of climate, land cover and soil. We used 83,416 SOC observations from global databases and developed a quantile regression neural network to quantify the SOC variation within soils with similar environmental characteristics. This allows us to identify similar areas that present higher SOC with the difference representing an additional storage potential. We estimated that the topsoils (0-30 cm) of global croplands (1,410 million hectares) hold 83 Pg C. The additional SOC storage potential in the topsoil of global croplands ranges from 29 to 65 Pg C. These values only equate to three to seven years of global emissions, potentially offsetting 35% of agriculture's 85 Pg historical carbon debt estimate due to conversion from natural ecosystems. As SOC store is temperature-dependent, this potential is likely to reduce by 14% by 2040 due to climate change in a "business as usual" scenario. The results of this article can provide a guide to areas of focus for SOC sequestration, and highlight the environmental cost of agriculture.

Digital soil mapping of soil total nitrogen based on Landsat 8, Sentinel 2, and WorldView-2 images in smallholder farms in Yellow River Basin, China

Predicting spatial explicit information of soil nutrients is critical for sustainable soil management and food security under climate change and human disturbance in agricultural land. Digital soil mapping (DSM) techniques can utilize soil-landscape information from remote sensing data to predict the spatial pattern of soil nutrients, and it is important to explore the effects of remote sensing data types on DSM. This research utilized Landsat 8 (LT), Sentinel 2 (ST), and WorldView-2 (WV) remote sensing data and employed partial least squares regression (PLSR), random forest (RF), and support vector machine (SVM) algorithms to characterize the spatial pattern of soil total nitrogen (TN) in smallholder farm settings in Yellow River Basin, China. The overall relationships between TN and spectral indices from LT and ST were stronger than those from WV. Multiple red edge band-based spectral indices from ST and WV were relevant variables for TN, while there were no red band-based spectral indices from ST and WV identified as relevant variables for TN. Soil moisture and vegetation were major driving forces of soil TN spatial distribution in this area. The research also concluded that farmlands of crop rotation had relatively higher TN concentration compared with farmlands of monoculture. The soil prediction models based on WV achieved relatively lower model performance compared with those based on ST and LT. The effects of remote sensing data spectral resolution and spectral range on enhancing soil prediction model performance are higher than the effects of remote sensing data spatial resolution. Soil prediction models based on ST can provide location-specific soil maps, achieve fair model performance, and have low cost. This research suggests DSM research utilizing ST has relatively high prediction accuracy, and can produce soil maps that are fit for the spatial explicit soil management for smallholder farms.

Spatial modeling of soil organic carbon using remotely sensed indices and environmental field inventory variables

The relationship between soil organic carbon (SOC) and environmental parameters was investigated in the Galazchai Watershed, Iran. Therefore, correlating the SOC amounts with remote sensing (RS) indices, topographic variables, and soil texture was analyzed. Some 125 soil samples gather from the upper 30 cm, and the weight of each sample was about 0.5 kg. The RS indices, consisting of difference vegetation index (DVI), enhanced vegetation index (EVI), optimized soil adjusted vegetation index (OSAVI), normalized difference vegetation index (NDVI), and soil adjusted vegetation index (SAVI), were used. Topographic variables included slope, elevation, aspect, and topographical wetness index (TWI), as well as clay and silt contents. The ordinary least square (OLS) and the geographically weighted regression (GWR) were employed to develop the SOC relationship considering different combinations of the variables. Results showed that none of the combinations of variables accurately estimated SOC (R² < 0.32 and p value > 0.001). However, EVI with GWR (R² = 0.291) and OSAVI, clay, slope, and aspect with GWR (R²= 0.32) better estimated SOC. Therefore, results showed that the study remotely sensed indices and environmental field inventory variables could not favorably predict the SOC content. These results can be attributed to the low SOC values varying from 0.917 to 3.355%, with a mean of 2.194 ± 0.522 in the study watershed. However, studies using more uniformly distributed and denser sampling in the study area and other methods to investigate the relationship between variables are recommended.

Exploring the likely relationship between soil carbon change and environmental controls using nonrevisited temporal data sets: Mapping soil carbon dynamics across China

The prediction of soil organic carbon (SOC) changes in response to environmental change is often limited by a scarcity of revisited temporal data, which constrains scientific understanding and realistic predictions of soil carbon change. The present study improved the potential of nonrevisited temporal data in the prediction of SOC stocks (SOCS) variations. We proposed a method to develop predictions of SOCS change using two independent temporal data sets (pertaining to the 1980s and 2010s) in China based on the digital soil mapping technique. Changes in SOCS over time at the site level were analyzed via the interpolation of missing SOCS values in each data set. Quantitative SOCS change predictions were generated by modeling the relationship between SOCS change and variables that represent changes in climate, vegetation indices, and land cover. The scale-dependent response of SOCS change to these environmental dynamics was assessed. On average, a slight increase was observed from 3.70 kg m^-2 in the 1980s to 4.53 kg m^-2 in the 2010s. The proposed approach attained moderate accuracy with an R² value of 0.32 and a root mean squared error (RMSE) of 1.73 kg m^-2. We found that changes in climate factors were dominant controls of SOCS change over time at the country scale. At the regional scale, the controlling factors of SOCS change were distinct and variable. Our case study may be of value in the application of independent temporal data sets to analyze soil carbon change on multiple scales. The method may be used to resolve questions of soil carbon change projections and provide an alternative solution to predict likely changes in soil carbon in response to future environmental change when no temporal data are available.

Digital mapping of zinc in urban topsoil using multisource geospatial data and random forest

This study aimed to map the spatial patterns of Zn in urban topsoil by using multisource geospatial data and machine learning method. Geological map, digital elevation models, and Landsat images were used to extract data related to geology, relief, and land use types and a vegetation index. Urban functional types were derived from the fusion of Systeme Probatoire d'Observation de la Terre 5 images, points of interest, and real-time Tencent user data. A geodetector was adopted to select key environmental covariates. Random forest (RF) and geographically weighted regression (GWR) were employed to model and map Zn concentrations in urban topsoil. The results showed that urban functional type, geology, NDVI, elevation, slope, and aspect were key environmental covariates. Compared with land use types, urban functional types could better reflect the spatial variation in Zn. The RF and GWR models were established using the key environmental covariates, with leave-one-out cross-validated R values of 0.68 and 0.58 and root mean square errors of 0.51 and 0.57, respectively. The results indicated that digital mapping of Zn in urban topsoil by using multisource geospatial data and RF was feasible. RF might be more suitable to fit the stochastic characteristics of Zn in urban topsoils than GWR, which considers deterministic trends in modeling.

Selecting optimal calibration samples using proximal sensing EM induction and γ-ray spectrometry data: An application to managing lime and magnesium in sugarcane growing soil

Calcium (Ca) and magnesium (Mg) are essential for growth of sugarcane leaves and roots, as well as respiration and nitrogen metabolism, respectively. To assist farmers decide suitable application rates of lime and Mg fertiliser, respectively, the Australian sugarcane industry established the Six-Easy-Steps nutrient management guidelines based on topsoil (0-0.3 m) Ca (cmol(+) kg^-1) and Mg (cmol(+) kg^-1). Given the heterogeneous nature of soil, digital soil mapping (DSM) methods can be employed to allow for the precise application rate to be determined. In this study, we examine statistical models (i.e., ordinary kriging [OK], linear mixed model [LMM], quantile regression forests [QRF], support vector machine [SVM], and Cubist regression kriging [CubistRK]) to predict topsoil and subsoil (0.6-0.9) Ca and Mg, employing digital data in combination (i.e., proximal sensing electromagnetic induction (EMI) [e.g., 1mPcon, 1mHcon, etc.], gamma-ray [γ-ray] spectrometry [i.e., TC, K, U and Th] and digital elevation model [DEM] derivatives). We also investigate various sampling designs (i.e., spatial coverage [SCS], feature space coverage [FSCS], conditioned Latin hypercube [cLHS] and simple random sampling [SRS]) and calibration sample size (i.e., n = 180, 150, 120, 90, 60 and 30). The predictions are assessed using Lin's concordance correlation coefficient (LCCC) and ratio of performance to interquartile distance (RPIQ) with an independent validation dataset (i.e., n = 40). The best results were for prediction of subsoil Mg, utilising CubistRK (LCCC = 0.82) with the largest calibration sample size (n = 180), followed by LMM (0.79), SVM (0.76), QRF (0.70) and OK (0.65). This was generally the case for topsoil and subsoil Ca. We also conclude that no single sampling design was universally better, and 180 samples were necessary for predicting topsoil Ca and Mg with moderate agreement (0.65 < LCCC < 0.80). However, with FSCS, a minimum of 120 samples were enough to calibrate a CubistRK model and achieve substantial (LCCC > 0.80) agreement for predicting subsoil Ca and Mg. With respect to soil use and management according to the Six-Easy-Steps, the sandy soil in the north and south require large application rate of lime (3.5 t/ha) and Mg (125 kg/ha), respectively. Conversely, varying amounts of fertiliser rates of lime (2.0, 1.5 and 1 t/ha) and Mg (50 kg/ha) were recommended where Vertosols were previously mapped.

Trend analysis of global usage of digital soil mapping models in the prediction of potentially toxic elements in soil/sediments: a bibliometric review

The rising and continuous pollution of the soil from anthropogenic activities is of great concern. Owing to this concern, the advent of digital soil mapping (DSM) has been a tool that soil scientists use in this era to predict the potentially toxic element (PTE) content in the soil. The purpose of this paper was to conduct a review of articles, summarize and analyse the spatial prediction of potentially toxic elements, determine and compare the models' usage as well as their performance over time. Through Scopus, the Web of Science and Google Scholar, we collected papers between the year 2001 and the first quarter of 2019, which were tailored towards the spatial PTE prediction using DSM approaches. The results indicated that soil pollution emanates from diverse sources. However, it provided reasons why the authors investigate a piece of land or area, highlighting the uncertainties in mapping, number of publications per journal and continental efforts to research as well as published on trending issues regarding DSM. This paper reveals the complementary role machine learning algorithms and the geostatistical models play in DSM. Nevertheless, geostatistical approaches remain the most preferred model compared to machine learning algorithms.

Large-scale digital mapping of topsoil total nitrogen using machine learning models and associated uncertainty map

Understanding the spatial distribution of soil nutrients and factors affecting their concentration and availability is crucial for soil fertility management and sustainable land utilization while quantifying factors affecting soil nitrogen distribution in Qorveh-Dehgolan plain is mostly lacking. This study, thus, aimed at digital modeling and mapping the spatial distribution of topsoil total nitrogen (TN) in Qorveh-Dehgolan plain with an area of 150,000 ha using random forest (RF), decision tree (DT), and cubist (CB) algorithms. A total of 130 observation points were collected from a depth of 0 to 30 cm from topsoil surfaces based on a random sampling pattern. Then, soil physicochemical properties, calcium carbonate equivalent, organic carbon, and topsoil total nitrogen were measured. A number of 51 environmental variables including 31 geomorphometric attributes derived from a digital elevation model with 12.5-m spatial resolution, 13 spectral indices and reflectance from SENTINEL-2 satellite (MSIsensor), and five soil properties and two spatial variables of latitude and longitude were used as covariates for digital mapping of topsoil total nitrogen. The most appropriate covariates were then selected by the Boruta algorithm in the R software environment. A standard deviation map was produced to show model uncertainty. The covariate selection resulted in the separation of 14 effective covariates in the spatial prediction of topsoil total nitrogen by using the data mining algorithms. The validation of digital mapping of topsoil total nitrogen by RF, DT, and CB models using 20% of independent data showed root mean square error (RMSE) of 0.032, 0.035, and 0.043%; mean absolute error (MAE) of 0.0008, 0.001, and 0.002%; and based on the coefficients of determination of 0.42, 0.38, 0.35, respectively. Relative importance (RI) of environmental covariates using the %IncMSE index indicated the importance of two geomorphometric variables of midslope position and normalized height along with SAVI and NDVI remote sensing variables in the spatial modeling and distribution of total nitrogen in the studied lands. The RF prediction and associated uncertainty maps, with show high accuracy and low standard deviation in the most part of study area, reveled low overfitting and overtraining in soil-landscape modeling; so, this model can lead to the development of a digital map of soil surface properties with acceptable accuracy for sustainable land utilization.

Prediction of soil organic carbon and the C:N ratio on a national scale using machine learning and satellite data: A comparison between Sentinel-2, Sentinel-3 and Landsat-8 images

Soil organic carbon (SOC) and soil carbon-to-nitrogen ratio (C:N) are the main indicators of soil quality and health and play an important role in maintaining soil quality. Together with Landsat, the improved spatial and temporal resolution Sentinel sensors provide the potential to investigate soil information on various scales. We analyzed and compared the potential of satellite sensors (Landsat-8, Sentinel-2 and Sentinel-3) with various spatial and temporal resolutions to predict SOC content and C:N ratio in Switzerland. Modeling was carried out at four spatial resolutions (800 m, 400 m, 100 m and 20 m) using three machine learning techniques: support vector machine (SVM), boosted regression tree (BRT) and random forest (RF). Soil prediction models were generated in these three machine learners in which 150 soil samples and different combinations of environmental data (topography, climate and satellite imagery) were used as inputs. The prediction results were evaluated by cross-validation. Our results revealed that the model type, modeling resolution and sensor selection greatly influenced outputs. By comparing satellite-based SOC models, the models built by Landsat-8 and Sentinel-2 performed the best and the worst, respectively. C:N ratio prediction models based on Landsat-8 and Sentinel-2 showed better results than Sentinel-3. However, the prediction models built by Sentinel-3 had competitive or better accuracy at coarse resolutions. The BRT models constructed by all available predictors at a resolution of 100 m obtained the best prediction accuracy of SOC content and C:N ratio; their relative improvements (in terms of R²) compared to models without remote sensing data input were 29.1% and 58.4%, respectively. The results of variable importance revealed that remote sensing variables were the best predictors for our soil prediction models. The predicted maps indicated that the higher SOC content was mainly distributed in the Alps, while the C:N ratio shared a similar distribution pattern with land use and had higher values in forest areas. This study provides useful indicators for a more effective modeling of soil properties on various scales based on satellite imagery.

Mapping the geogenic radon potential for Germany by machine learning

The radioactive gas radon (Rn) is considered as an indoor air pollutant due to its detrimental effects on human health. In fact, exposure to Rn belongs to the most important causes for lung cancer after tobacco smoking. The dominant source of indoor Rn is the ground beneath the house. The geogenic Rn potential (GRP) - a function of soil gas Rn concentration and soil gas permeability - quantifies what "earth delivers in terms of Rn" and represents a hazard indicator for elevated indoor Rn concentration. In this study, we aim at developing an improved spatial continuous GRP map based on 4448 field measurements of GRP distributed across Germany. We fitted three different machine learning algorithms, multivariate adaptive regression splines, random forest and support vector machines utilizing 36 candidate predictors. Predictor selection, hyperparameter tuning and performance assessment were conducted using a spatial cross-validation where the data was iteratively left out by spatial blocks of 40 km*40 km. This procedure counteracts the effect of spatial auto-correlation in predictor and response data and minimizes dependence of training and test data. The spatial cross-validated performance statistics revealed that random forest provided the most accurate predictions. The predictors selected as informative reflect geology, climate (temperature, precipitation and soil moisture), soil hydraulic, soil physical (field capacity, coarse fraction) and soil chemical properties (potassium and nitrogen concentration). Model interpretation techniques such as predictor importance as well as partial and spatial dependence plots confirmed the hypothesized dominant effect of geology on GRP, but also revealed significant contributions of the other predictors. Partial and spatial dependence plots gave further valuable insight into the quantitative predictor-response relationship and its spatial distribution. A comparison with a previous version of the German GRP map using 1359 independent test data indicates a significantly better performance of the random forest based map.

Mapping farmland soil organic carbon density in plains with combined cropping system extracted from NDVI time-series data

The accurate mapping of farmland soil organic carbon density (SOCD) is crucial for evaluating carbon (C) sequestration potential and forecasting climate change. Natural factors such as soil types and topographical factors are important variables in mapping soil properties. Moreover, cropping systems are important components of agricultural activities and are significantly correlated with soil properties. Therefore, integrating cropping systems and natural factors can improve the accuracy of mapping farmland SOCD. This study aimed to obtain and incorporate cropping system information in mapping SOCD in plains by combining normalized difference vegetation index (NDVI) time-series data and the regression Kriging (RK) method. We collected 230 topsoil samples in Jianghan Plain, China and (i) obtained the spatial patterns of crops in summer and winter using NDVI time-series data derived from HJ-1A/1B satellite images, (ii) investigated the differences in SOCD under different cropping systems, and (iii) evaluated the performance of the RK_CS model in integrating cropping systems and natural factors into mapping SOCD. ANOVA results showed significant differences in SOCD under different cropping systems. Specifically, the SOCD of single rice was higher than that of rice-wheat rotation and dry crops. Meanwhile, the regression results showed that SOCD was affected by natural factors and cropping system, with the latter playing a major role. The integration of soil types, slope and cropping systems explained approximately 26.3% of the variation in SOCD. Model validation results confirmed the effectiveness of the RK_CS model. The findings reveal single cropping rice sequences more C than other cropping systems. Cropping system is an important environmental variable in improving mapping farmland SOCD in plains.

Characterizing soil salinity at multiple depth using electromagnetic induction and remote sensing data with random forests: A case study in Tarim River Basin of southern Xinjiang, China

Tarim River Basin is experiencing heavy soil degeneration in a long term because of the extreme natural conditions, added with improper human activities such as reclamation and rejected field repeatedly, which hindered the soil health. One of the mainly form is soil salinization. Spatial distribution and variation of soil salinity is essential both for agricultural resource management and local economic development. However, knowledge of the spatial distribution of soil salinization in this region has not been updated since 1980s while land use and climate have undergone major changed. Electromagnetic induction (EMI) has been successfully used to directly measurement the spatial distribution of targeting soil property at field- scale, and apparent electrical conductivity (ECa, mS m^-1) has become a surrogate of soil salinity (EC, dS m^-1) studied by many researchers at local scale. However, the effectiveness of this equipment has not been verified in the typical soil salinization areas in southern Xinjiang, especially on a large scale. This study was aimed to test the performance of ECa jointed with Random Forest (RF) for soil salinity regional-scale mapping at a typical arid area, taking Tarim River Basin as an example. The result showed that ECa together with environmental derivative variables and with RF were suited for regional-scale soil salinity mapping. Predicted accuracy of EC was higher at surface (0-20 cm, R² = 0.65, RMSE = 5.59) and deeper soil depth (60-80 cm, R² = 0.63, RMSE = 2.00, and 80-100 cm, R² = 0.61, RMSE = 1.73), lower at transitional zone (20-40 cm, R² = 0.55, RMSE = 2.66, and 40-60 cm, R² = 0.51, RMSE = 2.49). When ECa is involved in modeling, the prediction accuracy of multiple depths of EC is improved by 13.33%-61.54%, of which the most obvious depths are 60-80 cm and 0-20 cm. The results of variable importance show that SoilGrids were also favored the power EC model. Hence, we strongly recommended to joint EMI reads with remote sensing imagery for soil salinity monitoring at large scale in southern Xinjiang. These EC and ECa map can provide a data source for environmental modeling, a benchmark against which to evaluate and monitor water and salt dynamics, and a guide for the design of future soil surveys.

The first version of nation-wide open 3D soil database for Sri Lanka

Soil data for Sri Lanka are available through semi-detailed series maps that were developed based on limited soil profile data combined with expert knowledge. This data plays a vital role in decisions at national and regional levels. However, the present format of this database does not allow for their wider use in crop simulation modelling and other related agricultural research that require finer scale data. This is due to the fact that cross-country profile data are not harmonised based on standard depths. Several attempts were made to produce digital soil data for Sri Lanka at different geographic scales, however, a completely harmonised data that covers variability across depths and properties is yet to be made available. In this article, we describe the first version of the open digital soil database that was developed using a database of 122 locations across the country. Soil properties were harmonised for standard depths using equal-area quadratic smoothing splines. Out of several interpolation methods that were evaluated for univariate interpolation, maps which were produced with the least overall error (RMSE) in the process of cross-validation were selected. The newly developed digital soil database contains 9 soil properties; pH, bulk density, cation exchange capacity, organic carbon, volumetric moisture content at 0.33 and 15 bars levels, sand silt and clay content. Moreover, the data are available for five standard depth layers as 0–5, 5–15, 15–30, 30–60 and 60–100 cm in raster format at 1 km spatial resolution. Both interpolated property maps and their error maps were stored in an open repository and made available for public use. The first version of all maps is also showcased online through open web mapping services. The repository will be gradually updated with higher resolution and more accurate maps as more samples become available and better interpolation method are used. This data could provide complementary information for insight generation at finer scales where limited local informaiton about soils hinders agricultural development.

Use of infrared spectroscopy and geospatial techniques for measurement and spatial prediction of soil properties

The main aim of this research was to assess the use of mid-infrared (MIR) spectroscopy and geostatistical model for the evaluation and mapping of the spatial variability of some selected soil properties across a field. It is with the view of aiding site-specific soil management decisions. The performance of the model for the prediction of the components (soil parameters) was reported using the coefficient of determination (R²) and root mean square error (RMSE) values of the validation data set. Results revealed that least square regression model performed better in predicting cation exchange capacity-CEC (R² = 0.88 and RMSE = 8.98), soil organic carbon-OC (R² = 0.88, RMSE = 0.55), and total nitrogen-TN (R² = 0.91 and RMSE = 0.04). The first five principal components (PC) accounted for 78.17% of the total variance (PC1 = 25.75%, PC2 = 18.06%, PC3 = 13.85%, PC4 = 11.12%, and PC5 = 9.39%) and represented most of the variation within the data set. The coefficient of variation ranged from 6.73% for soil pH to 57.02% for available phosphorus (av. P). The soil pH values ranged from 4.21 to 6.57. The mean soil OC density was 2.14 kg m⁻² within 50 cm soil depth. Nearly 96–97% of the soils contained av. P and sulfur (SO42−-S) below the critical levels. The overall results revealed that soil properties varied spatially. Hence, we suggest that mapping the spatial variability of soils across a field is a cost-effective solution for soil management.

Tracking changes in soil organic carbon across the heterogeneous agricultural landscape of the Lower Fraser Valley of British Columbia

Increasing soil organic carbon (SOC) can improve the capacity of agricultural systems to both adapt to and mitigate climate change. Despite its importance, the current understanding of the magnitude or even the direction of SOC change in agricultural landscapes is limited. While changes in land use/land cover (LULC) and climate are among the main drivers of changes in SOC, their relative importance for the spatiotemporal assessment of SOC is unclear. This study evaluated LULC and SOC dynamics using archived and recent soil samples, remote sensing, and digital soil mapping in the Lower Fraser Valley of British Columbia, Canada. We combined both pixel- and object-based analysis of Landsat satellite imagery to assess LULC changes from 1984 to 2018. We achieved an overall accuracy of 81% and kappa coefficient of 0.77 for LULC classification using a random forest model. For predicting SOC for the same time period, we applied soil and vegetation indices derived from Landsat images, topographic indices, historic soil survey variables, and climate data in a random forest model. The SOC prediction of 2018 resulted in a coefficient of determination (R²) of 0.67, concordance correlation coefficient (CCC) of 0.76, and normalized root mean square error (nRMSE) of 0.12. For 1984, the SOC prediction accuracies were 0.46, 0.58, and 0.18 for R², CCC, and nRMSE, respectively. We detected SOC loss in 61%, gain in 12%, while 27% remained unchanged across the study area. Although we detected large losses of SOC due to LULC change, the majority of the SOC losses across the landscape were attributed to areas that were remained in the same type of agricultural production since 1984. Climate variability did not, however, have a strong effect on SOC changes. These results can inform decision making in the study area to support sustainable LULC management for enhancing SOC sequestration.

High-resolution digital mapping of soil organic carbon and soil total nitrogen using DEM derivatives, Sentinel-1 and Sentinel-2 data based on machine learning algorithms

Soil organic carbon (SOC) and soil total nitrogen (STN) are important indicators of soil health and play a key role in the global carbon and nitrogen cycles. High-resolution radar Sentinel-1 and multispectral Sentinel-2 images have the potential to investigate soil spatial distribution information over a large area, although Sentinel-1 and Sentinel-2 data have rarely been combined to map either SOC or STN content. In this study, we applied machine learning techniques to map both SOC and STN content in the southern part of Central Europe using digital elevation model (DEM) derivatives, multi-temporal Sentinel-1 and Sentinel-2 data, and evaluated the potential of different remote sensing sensors (Sentinel-1 and Sentinel-2) to predict SOC and STN content. Four machine-learners including random forest (RF), boosted regression trees (BRT), support vector machine (SVM) and Bagged CART were used to construct predictive models of SOC and STN contents based on 179 soil samples and different combinations of environmental covariates. The performance of these models was evaluated based on a 10-fold cross-validation method by three statistical indicators. Overall, the BRT model performed better than RF, SVM and Bagged CART, and these models yielded similar spatial distribution patterns of SOC and STN. Our results showed that multi-source sensor methods provided more accurate predictions of SOC and STN contents than individual sensors. The application of radar Sentinel-1 and multispectral Sentinel-2 images proved useful for predicting SOC and STN. A combination of Sentinel-1/2-derived predictors and DEM derivatives yielded the highest prediction accuracy. The prediction accuracy changed with and without the Sentinel-1/2-derived predictors, with the R² for estimating both SOC and STN content using the BRT model increasing by 12.8% and 18.8%, respectively. Topographic variables were the main explanatory variables for SOC and STN predictions, where elevation was assigned as the variable with the most importance by the models. The results of this study illustrate the potential of free high-resolution radar Sentinel-1 and multispectral Sentinel-2 data as input when developing SOC and STN prediction models.

Improved digital soil mapping with multitemporal remotely sensed satellite data fusion: A case study in Iran

Modeling and mapping of soil properties are critical in many environmental, climatic, ecological and hydrological applications. Digital soil mapping (DSM) techniques are now commonly applied to predict soil properties with limited data by developing predictive relationships with environmental covariates. Most studies derive covariates from a digital elevation model (named static covariates). Many works also include single-day remotely sensed satellite imagery. However, multitemporal satellite images can capture information about soil properties over time and bring additional information in predicting soil properties in DSM. We refer to covariates derived from multitemporal satellite images as dynamic covariates. The objective of this study was to assess the performance of DSM when using terrain derivatives (static covariates), single-date remotely sensed satellite indices (limited dynamic covariates), multitemporal satellite indices (dynamic covariates), and combinations of terrain derivatives and satellite indices (covariate fusion) as covariates in predicting soil properties and estimating uncertainty. Three soil properties are considered in this study: organic carbon (OC), sand content, and calcium carbonate equivalent (CCE). Inclusion of single and/or multitemporal remotely sensed satellite indices improved the prediction of soil properties over traditionally used terrain indices. Significant improvements were observed in the prediction of soil properties using two models, Cubist and random forest (RF). The increase in the R² values for Cubist and RF were 126% and 78% for OC, 110% and 54% for sand, and 87% and 32% for CCE. The RMSE decreased by 34% and 27% for OC, 25% and 12% for sand, and 39% and 19% for CCE, when compared to the terrain indices only model. This also reduced the uncertainty of estimation and mapping. These clearly showed the advantage of using multitemporal satellite data fusion rather than simply using static terrain indices for DSM of soil properties to deliver a great potential in improving soil modeling and mapping for many applications.

An improved similarity-based approach to predicting and mapping soil organic carbon and soil total nitrogen in a coastal region of northeastern China

Soil organic carbon (SOC) and soil total nitrogen (STN) are major soil indicators for soil quality and fertility. Accurate mapping SOC and STN in soils would help both managed and natural soils and ecosystem management. This study developed an improved similarity-based approach (ISA) to predicting and mapping topsoil (0-20 cm soil depth) SOC and STN in a coastal region of northeastern China. Six environmental variables including elevation, slope gradient, topographic wetness index, the mean annual temperature, the mean annual temperature, and normalized difference vegetation index were used as predictors. Soil survey data in 2012 was designed based on the clustering of the study area into six climatic vegetation landscape units. In each landscape unit, 20-25 sampling points were determined at different landform positions considering local climate, soil type, elevation and other environmental factors, and finally 126 sampling points were obtained. Soil sampling from the depth of 0-20 cm were used for model prediction and validation. The ISA model performance was compared with the geographically weighted regression (GWR), regression kriging (RK), boosted regression trees (BRT) considering mean absolute prediction error (MAE), root mean square error (RMSE), coefficient of determination (R ²), and maximum relative difference (RD) indices. We found that the ISA method performed best with the highest R² and lowest MAE, RMSE compared to GWR, RK, and BRT methods. The ISA method could explain 76% and 83% of the total SOC and STN variability, respectively, 12-40% higher than other models in the study area. Elevation had the largest influence on SOC and STN distribution. We conclude that the developed ISA model is robust and effective in mapping SOC and STN, particularly in the areas with complex vegetation-landscape when limited samples are available. The method needs to be tested for other regions in our future research.

Updated information on soil salinity in a typical oasis agroecosystem and desert-oasis ecotone: Case study conducted along the Tarim River, China

Precise and spatially explicit regional estimates of soil salinity are necessary to efficiently management and utilise limited land and water resources. Despite advances achieved in remote sensing over the past century, knowledge about the distribution and severity of soil salinization in economically important areas, such as oasis agroecosystems and desert-oasis ecotones (OADoE), is currently limited. An example of an area is southern Xinjiang, where the OADoE has a high anthropogenic influence. This study was conducted with the aim of mapping soil salinity in typical OADoE using remote sensing and machine learning techniques (Cubist and Random Forest, RF). A range of covariates was obtained from the multi-temporal Landsat-8 operational land imager (OLI) satellite for the period from 2013 to 2018. The values of coefficients of determination (R²), Lin's concordance correlation coefficient, root mean square error, and relative root mean squared error values, were 0.78, 0.87, 9.59, and 0.76, respectively, for the Cubist and 0.78, 0.86, 9.79, and 0.78, respectively, for RF models. The slope of the linear fitting equation was higher for the Cubist model (0.75) than for RF (0.69). The explanatory power of Cubist and RF for soil salinity variation were 33.22% and 31.41% in the agroecosystem, and 72.25% and 71.66% in desert-oasis ecotone, respectively. For the agroecosystem, the range of the predicted values for 89.13% (Cubist) and 84.78% (RF) of sample was controlled within the same observational range at an interval of 0-5 dS m^-1. Compared to single-year data (from 2013 to 2018), the ability to account for model spatial variability in soil salinity based on multi-year Landsat images was increased by 16%-35%. According to the variable importance evaluation, soil-related indices are the most important predictor variables, followed by vegetation, topography, landform, and land use, with relative importance values of 60%, 21%, 16%, and 3%, respectively. The predicted map was also broadly consistent with those obtained for Xinjiang in the Harmonized World Soil Database (HWSD) from the second national soil survey of China conducted from 1984 to 1997. The results also showed that the average value of the study area is 8.10 dS m^-1 based on the Cubist-based map whereas that of the HWSD is 10.60 dS m^-1, this implied that the overall salinity level has reduced by 23.58%. The methodological framework presented covers all prediction process steps and has considerable potential to be used in future soil salinity mapping at large scales for other similar region as OADoEs. The map derived from the Cubist/RF model revealed more detailed variation information about spatial distribution of the soil salinity compared to HWSD, and can further assist with decision-making when planning and utilising on existing soil and water resources in OADoEs.

High-resolution three-dimensional mapping of soil organic carbon in China: Effects of SoilGrids products on national modeling

Soil organic carbon (SOC) is a key factor in soil fertility and structure and plays an important role in the global carbon cycle. However, SOC causes a large uncertainty in Earth System Models for predicting future climate change. The GlobalSoilMap (GSM) project aims to provide global digital soil maps of primary functional soil properties at six standard depth intervals (0-5, 5-15, 15-30, 30-60, 60-100, and 100-200 cm) with a grid resolution of 90 × 90 m. Currently, few SOC national products that meet the GSM specifications are available. This study describes the three-dimensional spatial modeling of SOC maps according to GSM specifications. We used 5982 soil profiles collected during the Second National Soil Survey of China, along with 16 environmental covariates related to soil formation. The results were obtained by parallel computing over tiles of 100 × 100 km, and the predictions for the tiles were subsequently merged into a single SOC map for the whole of China per standard GSM depth interval. For each standard GSM depth interval, SOC contents and their uncertainties were predicted and mapped at a spatial resolution of approximately 90 m using bootstrapping. Southwestern and northeastern China had higher SOC contents than the rest of China did, whereas northwestern China had a lower SOC content. The range of the coefficient of determination for the six depth intervals ranged from 0.35 to 0.02, and the mean SOC content was 17.86-8.67 g kg^-1. Both these values decreased strongly with increasing soil depth. Cropland SOC content was lower than that of forest and grassland. The results of variable importance show that SoilGrids data were the best predictors for defining the soil-landscape relationship during regression modeling for SOC. These SOC maps can provide a data source for environmental modeling, a benchmark against which to evaluate and monitor SOC dynamics, and a guide for the design of future soil surveys.

Characterising dryland salinity in three dimensions

Due to frequent salt migration and large spatial variability within soil profiles, salinity characterisation by traditional drilling sampling methods is time-consuming and labour-intensive. Thus, it is necessary to develop monitoring technology and three-dimensional (3D) characterisation methods for rapid, non-invasive, and accurate soil salinity measurement. This study presents a new framework combining sensor technology and an inversion algorithm to characterise 3D soil salinity. Four typical land-use types (natural desert, natural vegetation, apple orchard, and winter wheat farmland) in the Aksu region of southern Xinjiang were surveyed and apparent conductivity (ECa) data were recorded at depths of 0.75 m and 1.50 m. ECa data were converted to electrical conductivity and salinity characterisation was conducted following U.S. Salinity Laboratory recommendations. Ordinary Kriging interpolation was used to map the spatial distribution and an iterative inversion model was used to map the vertical distribution of soil salinity. Model parameters were adjusted several times and the accuracy of different inversion algorithms was compared to obtain the best inversion effect. As a result, the Multilevel Orthogonal Inversion model was developed to characterise 3D soil salinity for different land-use types. Due to crop activities including irrigation, managed land use types (apple orchard and winter wheat plots) typically exhibited weaker salinity than natural systems (desert and vegetation plots) but greater spatial variability overall. The proposed framework combining EM (electromagnetic) sensing and the 3D inversion algorithm can effectively characterise and visualise soil salinity for the entire soil profile, which is important for land evaluation and improvement.

Soil class map of the Rio Jardim watershed in Central Brazil at 30 meter spatial resolution based on proximal and remote sensed data and MESMA method

Geospatial soil information is critical for agricultural policy formulation and decision making, land-use suitability analysis, sustainable soil management, environmental assessment, and other research topics that are of vital importance to agriculture and economy. Proximal and Remote sensing technologies enables us to collect, process, and analyze spectral data and to retrieve, synthesize, visualize valuable geospatial information for multidisciplinary uses. We obtained the soil class map provided in this article by processing and analyzing proximal and remote sensed data from soil samples collected in toposequences based on pedomorphogeological relashionships. The soils were classified up to the second categorical level (suborder) of the Brazilian Soil Classification System (SiBCS), as well as in the World Reference Base (WRB) and United States Soil Taxonomy (ST) systems. The raster map has 30 m resolution and its accuracy is 73% (Kappa coefficient of 0.73). The soil legend represents a soil class followed by its topsoil color.

Mapping dynamics of soil organic matter in croplands with MODIS data and machine learning algorithms

As an important indicator of soil quality, soil organic matter (SOM) significantly contributes to land productivity and ecosystem health. Accurately mapping SOM at regional scales is of critical importance for sustainable agriculture and soil utilization management and remains a grand challenge. Many studies used soil sampling data and machine learning algorithms to predict SOM at regional scales for a given year, while few studies mapped SOM for multiple years and examined its temporal dynamics. We compared the performance of four machine learning algorithms: decision tree (DT), bagging decision tree (BDT), random forest (RF), and gradient boosting regression trees (GBRT) in mapping SOM in Hubei province, China over the 18-year period from 2000 to 2017. Our results showed that RF and DT had the highest coefficient of determination (R²) (0.61) and the lowest potential bias (9.48 g/kg), respectively, while GBRT had the lowest mean error (ME) (1.26 g/kg), root mean squared error (RMSE) (5.41 g/kg) and Lin's concordance correlation coefficient (LCCC) (0.72). The SOM map based on GBRT better captured the distribution of the soil sample data than that based on RF. The trained GBRT model and the spatially explicitly data on explanatory variables (e.g., climate, terrain, remote sensing) were used to predict SOM for each 500 m × 500 m grid cell in Hubei for the period from 2000 to 2017. Our results showed that the SOM content of cropland was relatively high in the southeast and relatively low in the north. The SOM content in the topsoil varied from 0.89 to 58.86 g/kg and was averaged at 20.52 g/kg. The mean cropland SOM content of the province exhibited an increasing trend from 2000 to 2017 with an increase of 0. 26 g/kg and a growth rate of 1.28%. Spatially, the SOM content increased in southern Hubei and decreased in central and northern parts of the province. A large portion of the areas with decreasing SOM content in northern Hubei was reclaimed cropland, while a large part of the high-quality cropland with rising SOM content in the east (~0.45 × 10⁴ ha) was lost due to land use change (e.g., urbanization).

Assessing soil organic carbon stock of Wisconsin, USA and its fate under future land use and climate change

Carbon stored in soils contributes to a variety of soil functions, including biomass production, water storage and filtering, biodiversity maintenance, and many other ecosystem services. Understanding soil organic carbon (SOC) spatial distribution and projection of its future condition is essential for future CO₂ emission estimates and management options for storing carbon. However, modeling SOC spatiotemporal dynamics is challenging due to the inherent spatial heterogeneity and data limitation. The present study developed a spatially explicit prediction model in which the spatial relationship between SOC observation and seventeen environmental variables was established using the Cubist regression tree algorithm. The model was used to compile a baseline SOC stock map for the top 30 cm soil depth in the State of Wisconsin (WI) at a 90 m × 90 m grid resolution. Temporal SOC trend was assessed by comparing baseline and future SOC stock maps based on the space-for-time substitution model. SOC prediction for future considers land use, precipitation and temperature for the year 2050 at medium (A1B) CO₂ emissions scenario of the Intergovernmental Panel on Climate Change. Field soil observations were related to factors that are known to influence SOC distribution using the digital soil mapping framework. The model was validated on 25% test profiles (R²: 0.38; RMSE: 0.64; ME: -0.03) that were not used during model training that used the remaining 75% of the data (R²: 0.76; RMSE: 0.40; ME: -0.006). In addition, maps of the model error, and areal extent of Cubist prediction rules were reported. The model identified soil parent material and land use as key drivers of SOC distribution including temperature and precipitation. Among the terrain attributes, elevation, mass-balance index, mid-slope position, slope-length factor and wind effect were important. Results showed that Wisconsin soils had an average baseline SOC stock of 90 Mg ha^-1 and the distribution was highly variable (CV: 64%). It was estimated that WI soils would have an additional 20 Mg ha^-1 SOC by the year 2050 under changing land use and climate. Histosols and Spodosols were expected to lose 19 Mg ha^-1 and 4 Mg ha^-1, respectively, while Mollisols were expected to accumulate the largest SOC stock (62 Mg ha^-1). All land-use types would be accumulating SOC by 2050 except for wetlands (-34 Mg C ha^-1). This study found that Wisconsin soils will continue to sequester more carbon in the coming decades and most of the Driftless Area will be sequestering the greatest SOC (+63 Mg C ha^-1). Most of the SOC would be lost from the Northern Lakes and Forests ecological zone (-12 Mg C ha^-1). The study highlighted areas of potential C sequestration and areas under threat of C loss. The maps generated in this study would be highly useful in farm management and environmental policy decisions at different spatial levels in Wisconsin.

A high-resolution map of soil pH in China made by hybrid modelling of sparse soil data and environmental covariates and its implications for pollution

The soil's pH is the single most important indicator of the soil's quality, whether for agriculture, pollution control or environmental health and ecosystem functioning. Well documented data on soil pH are sparse for the whole of China - data for only 4700 soil profiles were available from China's Second National Soil Inventory. By combining those data, standardized for the topsoil (0-20 cm), with 17 environmental covariates at a fine resolution (3 arc-second or 90 m) we have predicted the soil's pH at that resolution, that is at more than 10⁹ points. We did so by parallel computing over tiles, each 100 km × 100 km, with two machine learning techniques, namely Random Forest and XGBoost. The predictions for the tiles were then merged into a single map of soil pH for the whole of China. The quality of the predictions were assessed by cross-validation. The root mean squared error (RMSE) was an acceptable 0.71 pH units per point, and Lin's Concordance Correlation Coefficient was 0.84. The hybrid model revealed that climate (mean annual precipitation and mean annual temperature) and soil type were the main factors determining the soil's pH. The pH map showed acid soil mainly in southern and north-eastern China, and alkaline soil dominant in northern and western China. This map can provide a benchmark against which to evaluate the impacts of changes in land use and climate on the soil's pH, and it can guide advisors and agencies who make decisions on remediation and prevention of soil acidification, salinization and pollution by heavy metals, for which we provide examples for cadmium and mercury.

Some methods to improve the utility of conditioned Latin hypercube sampling

The conditioned Latin hypercube sampling (cLHS) algorithm is popularly used for planning field sampling surveys in order to understand the spatial behavior of natural phenomena such as soils. This technical note collates, summarizes, and extends existing solutions to problems that field scientists face when using cLHS. These problems include optimizing the sample size, re-locating sites when an original site is deemed inaccessible, and how to account for existing sample data, so that under-sampled areas can be prioritized for sampling. These solutions, which we also share as individual R scripts, will facilitate much wider application of what has been a very useful sampling algorithm for scientific investigation of soil spatial variation.

Land use and climate change effects on soil organic carbon in North and Northeast China

Soil is recognized as the largest carbon reservoir in the terrestrial ecosystem. Soil organic carbon (SOC) is vulnerable to changes in land use and climate. For a better understanding of the SOC dynamics and its driving factors, we collected data of the 1980s and 2000s in the North and Northeast China and conducted the digital soil mapping for spatial variation of SOC for the respective period. In the 1980s, 585 soils were sampled and the area was resampled in 2003 and 2004 (1062 samples) in a 30-km grid. The main land use in the area was cropland, forest and grassland. The random forest was used to predict the SOC concentration and its temporal change using land use, terrain factors, vegetation index, vis-NIR spectra and climate factors as predictors. The average SOC concentration in 1985 was 10.0 g kg^-1 compared to 12.5 g kg^-1 in 2004. The SOC variation was similar over the two periods, and levels increased from south to north. The estimated SOC stock was 1.68 Pg in 1985 and 1.66 Pg in 2004, but the SOC changes were different under different land uses. Over the twenty-year period, average temperatures increased and large areas of forests and grassland were converted to cropland. SOC under cropland was increased by 0.094 Pg (+9%) whereas 0.089 Pg SOC was lost under forests (-25%) and 0.037 Pg in the soils under grassland (-25%). It is concluded that land use is the main drivers for SOC changes in this area while climate change had different contributions in different regions. SOC loss was remarkable under the land use conversion while cropland has considerable potential to sequester SOC.

High resolution mapping of soil organic carbon stocks using remote sensing variables in the semi-arid rangelands of eastern Australia

Efficient and effective modelling methods to assess soil organic carbon (SOC) stock are central in understanding the global carbon cycle and informing related land management decisions. However, mapping SOC stocks in semi-arid rangelands is challenging due to the lack of data and poor spatial coverage. The use of remote sensing data to provide an indirect measurement of SOC to inform digital soil mapping has the potential to provide more reliable and cost-effective estimates of SOC compared with field-based, direct measurement. Despite this potential, the role of remote sensing data in improving the knowledge of soil information in semi-arid rangelands has not been fully explored. This study firstly investigated the use of high spatial resolution satellite data (seasonal fractional cover data; SFC) together with elevation, lithology, climatic data and observed soil data to map the spatial distribution of SOC at two soil depths (0-5cm and 0-30cm) in semi-arid rangelands of eastern Australia. Overall, model performance statistics showed that random forest (RF) and boosted regression trees (BRT) models performed better than support vector machine (SVM). The models obtained moderate results with R² of 0.32 for SOC stock at 0-5cm and 0.44 at 0-30cm, RMSE of 3.51MgCha^-1 at 0-5cm and 9.16MgCha^-1 at 0-30cm without considering SFC covariates. In contrast, by including SFC, the model accuracy for predicting SOC stock improved by 7.4-12.7% at 0-5cm, and by 2.8-5.9% at 0-30cm, highlighting the importance of including SFC to enhance the performance of the three modelling techniques. Furthermore, our models produced a more accurate and higher resolution digital SOC stock map compared with other available mapping products for the region. The data and high-resolution maps from this study can be used for future soil carbon assessment and monitoring.

Fine resolution map of top- and subsoil carbon sequestration potential in France

Although soils have a high potential to offset CO₂ emissions through its conversion into soil organic carbon (SOC) with long turnover time, it is widely accepted that there is an upper limit of soil stable C storage, which is referred to SOC saturation. In this study we estimate SOC saturation in French topsoil (0-30cm) and subsoil (30-50cm), using the Hassink equation and calculate the additional SOC sequestration potential (SOC_sp) by the difference between SOC saturation and fine fraction C on an unbiased sampling set of sites covering whole mainland France. We then map with fine resolution the geographical distribution of SOC_sp over the French territory using a regression Kriging approach with environmental covariates. Results show that the controlling factors of SOC_sp differ from topsoil and subsoil. The main controlling factor of SOCsp in topsoils is land use. Nearly half of forest topsoils are over-saturated with a SOC_sp close to 0 (mean and standard error at 0.19±0.12) whereas cropland, vineyard and orchard soils are largely unsaturated with degrees of C saturation deficit at 36.45±0.68% and 57.10±1.64%, respectively. The determinant of C sequestration potential in subsoils is related to parent material. There is a large additional SOC_sp in subsoil for all land uses with degrees of C saturation deficit between 48.52±4.83% and 68.68±0.42%. Overall the SOCsp for French soils appears to be very large (1008Mt C for topsoil and 1360Mt C for subsoil) when compared to previous total SOC stocks estimates of about 3.5Gt in French topsoil. Our results also show that overall, 176Mt C exceed C saturation in French topsoil and might thus be very sensitive to land use change.

Description and spatial inference of soil drainage using matrix soil colours in the Lower Hunter Valley, New South Wales, Australia

Soil colour is often used as a general purpose indicator of internal soil drainage. In this study we developed a necessarily simple model of soil drainage which combines the tacit knowledge of the soil surveyor with observed matrix soil colour descriptions. From built up knowledge of the soils in our Lower Hunter Valley, New South Wales study area, the sequence of well-draining → imperfectly draining → poorly draining soils generally follows the colour sequence of red → brown → yellow → grey → black soil matrix colours. For each soil profile, soil drainage is estimated somewhere on a continuous index of between 5 (very well drained) and 1 (very poorly drained) based on the proximity or similarity to reference soil colours of the soil drainage colour sequence. The estimation of drainage index at each profile incorporates the whole-profile descriptions of soil colour where necessary, and is weighted such that observation of soil colour at depth and/or dominantly observed horizons are given more preference than observations near the soil surface. The soil drainage index, by definition disregards surficial soil horizons and consolidated and semi-consolidated parent materials. With the view to understanding the spatial distribution of soil drainage we digitally mapped the index across our study area. Spatial inference of the drainage index was made using Cubist regression tree model combined with residual kriging. Environmental covariates for deterministic inference were principally terrain variables derived from a digital elevation model. Pearson's correlation coefficients indicated the variables most strongly correlated with soil drainage were topographic wetness index (-0.34), mid-slope position (-0.29), multi-resolution valley bottom flatness index (-0.29) and vertical distance to channel network (VDCN) (0.26). From the regression tree modelling, two linear models of soil drainage were derived. The partitioning of models was based upon threshold criteria of VDCN. Validation of the regression kriging model using a withheld dataset resulted in a root mean square error of 0.90 soil drainage index units. Concordance between observations and predictions was 0.49. Given the scale of mapping, and inherent subjectivity of soil colour description, these results are acceptable. Furthermore, the spatial distribution of soil drainage predicted in our study area is attuned with our mental model developed over successive field surveys. Our approach, while exclusively calibrated for the conditions observed in our study area, can be generalised once the unique soil colour and soil drainage relationship is expertly defined for an area or region in question. With such rules established, the quantitative components of the method would remain unchanged.