Estimating Soil Organic Carbon Research Articles

Agricultural soil organic carbon stocks in the north-eastern Iberian Peninsula: Drivers and spatial variability

Estimating soil organic carbon (SOC) stocks under agriculture, assessing the importance of their drivers and understanding the spatial distribution of SOC stocks are crucial to predicting possible future SOC stocks scenarios under climate change conditions and to designing appropriate mitigation and adaptation strategies. This study characterized and modelled SOC stocks at two soil depth intervals, topsoil (0–30 cm) and subsoil (30–100 cm), based on both legacy and recent data from 7245 agricultural soil profiles and using environmental drivers (climate, agricultural practices and soil properties) for agricultural soils in Catalonia (NE Spain). Generalized Least Square (GLS) and Geographical Weighted Regression (GWR) were used as modelling approaches to: (i) assess the main SOC stock drivers and their effects on SOC stocks; (ii) analyse spatial variability of SOC stocks and their relationships with the main drivers; and (iii) predict and map SOC stocks at the regional scale. While topsoil variation of SOC stocks depended mainly on climate, soil texture and agricultural variables, subsoil SOC stocks changes depended mainly on soil attributes such us soil texture, clay content, soil type or depth to bedrock. The GWR model revealed that the relationship between SOC stocks and drivers varied spatially. Finally, the study was only able to predict and map topsoil SOC stocks at the regional scale, because controlling factors of SOC stocks at the subsoil level were largely unavailable for digital mapping. According to the resulting map, the mean SOC stock value for Catalan agriculture at the topsoil level was 4.88 ± 0.89 kg/m2 and the total magnitude of the carbon pool in agricultural soils of Catalonia up to 30 cm reached 47.9 Tg. The present study findings are useful for defining carbon sequestration strategies at the regional scale related with agricultural land use changes and agricultural management practices in a context of climate change.

Effect of cultivation history on soil organic carbon status of arable land in northeastern China

Estimation of soil organic carbon (SOC) in arable land can help with the better understanding of the impact of human activities on soils and the environment. Knowledge of cultivation history can assist with studies on soil carbon balance and CO2 emissions. This study used historical cultivation data for the last 300 years and quantified the influence on the spatial distribution of SOC in arable land from northeastern China. More than 230 SOC observations were used to train the prediction model derived from boosted regression trees (BRT) where topography, temperature, precipitation, and length of cultivation were employed as SOC predictors. Two BRT models, one with all predictors except length of cultivation (MA) and the other with all predictors (MB) were generated and the models were compared for SOC prediction performance. The MB model accounted for 76% of SOC variability in the study area and showed a better prediction performance compared to the MA model which exhibited lower values for Lin's concordance correlation coefficient (LCCC) and coefficient of determination (R2) and the higher mean absolute prediction error (MAE) and root mean square error (RMSE) indices. The model assigned the highest variable importance to length of cultivation indicating that this variable was key to determining the status of SOC spatial distribution. Our analysis showed that the study area lost about 45% SOC during the past 300 years of cultivation in Northeastern China. The predicted maps from both models suggested a decreasing trend in SOC content from north to south in the study area. We conclude that the historical data in cultivation history were a key variable in SOC predictions on the arable lands and it should be considered as a major variable in future SOC mapping studies, especially in agricultural areas with a long history of cultivation.

The role of topography in shaping the spatial patterns of soil organic carbon

In regions with a relatively complex topography, the soil organic carbon (SOC) is characterized by a high spatial heterogeneity, which is likely to contribute to large uncertainties in the regional SOC estimation. Generally, the elevation and aspect are expected to be the two main topographic controls of the spatial patterns of SOC on steep hillslopes. However, the manner in which the aspect-induced variability in SOC varies with elevation zone is still less understood. In this study, SOC in three elevation-dependent vegetation zones (~2450 m: the shrub-steppe zone, ~2900 m: the forest-steppe zone, and ~3350 m: the shrub-meadow zone) was examined in the Qilian Mountains. Soil samples at 0–10, 10–20, 20–40, and 40–60 cm were collected on the south-, southwest-, west-, northwest-, and north-facing slopes with similar slope gradients (~32°) in each of the three vegetation zones. The results showed that the SOC density tends to increase following the elevation zone, with an average density of about 24.30 kg m−2 at the 0–60 cm. In addition, the SOC concentration appeared to have increased with the cosine of the aspect following an exponential function. The average rate of increase in terms of the aspect at 0–10 cm in the forest-steppe zone was approximately 0.46 g kg−1 degree−1, which amounted to 1.39 and 2.19 fold greater than in the shrub-steppe and shrub-meadow zones, respectively. Path analysis performed using structural equation modelling (SEM) demonstrated that the elevation and the aspect had an indirect influence on the SOC density, through a direct effect on the soil temperature, vegetation type, and soil moisture, and this could possibly explain 88% of the total variation in the SOC density. The present results suggest that the forest-steppe zone is characterized by the strongest aspect effect on SOC in comparison with the shrub-steppe and shrub-meadow zones, and confirm the importance of elevation and aspect in shaping the spatial patterns of SOC.

Large, climate-sensitive soil carbon stocks mapped with pedology-informed machine learning in the North Pacific coastal temperate rainforest

Accurate soil organic carbon (SOC) maps are needed to predict the terrestrial SOC feedback to climate change, one of the largest remaining uncertainties in Earth system modeling. Over the last decade, global scale models have produced varied predictions of the size and distribution of SOC stocks, ranging from 1000 to >3000 Pg of C within the top 1 m. Regional assessments may help validate or improve global maps because they can examine landscape controls on SOC stocks and offer a tractable means to retain regionally-specific information, such as soil taxonomy, during database creation and modeling. We compile a new transboundary SOC stock database for coastal watersheds of the North Pacific coastal temperate rainforest, using soil classification data to guide gap-filling and machine learning approaches to explore spatial controls on SOC and predict regional stocks. Precipitation and topographic attributes controlling soil wetness were found to be the dominant controls of SOC, underscoring the dependence of C accumulation on high soil moisture. The random forest model predicted stocks of 4.5 Pg C (to 1 m) for the study region, 22% of which was stored in organic soil layers. Calculated stocks of 228 ± 111 Mg C ha−1 fell within ranges of several past regional studies and indicate 11–33 Pg C may be stored across temperate rainforest soils globally. Predictions compared very favorably to regionalized estimates from two spatially-explicit global products (Pearson’s correlation: ρ = 0.73 versus 0.34). Notably, SoilGrids 250 m was an outlier for estimates of total SOC, predicting 4-fold higher stocks (18 Pg C) and indicating bias in this global product for the soils of the temperate rainforest. In sum our study demonstrates that CTR ecosystems represent a moisture-dependent hotspot for SOC storage at mid-latitudes.

Dynamics of soil organic carbon (SOC) content in stands of Norway spruce (Picea abies) in central Europe

Abstract: Norway spruce is the main forest tree species in the Czech Republic. Until now, little attention has been given in the literature to the dynamics of soil organic carbon (SOC) content under Norway spruce stands as a function of stand characteristics. The aim of this study is to estimate soil organic carbon (SOC) content and stock changes in organic and surface mineral soil horizons on forest sites with a dominant representation of Norway spruce. In the study area, a significantly higher content of SOC was found in the surface mineral soil horizon than in the organic soil horizon. In both soil horizons, there was evidence of an increasing trend of SOC with the increasing age of forest stands, a decreasing trend of SOC with increasing density of stocking and an increasing trend of SOC with increasing altitude. The relationship of SOC content with soil group (Podzol vs. non-Podzol) has also been demonstrated. The greatest potential for long-term carbon sequestration in soils was shown in older stands (101-190 years) dominated by Norway spruce with lower density of stocking, located in forest vegetation zones (altitude range: 1010-1225 m a.s.l.) where natural mountain Norway spruce forests currently occur. According to our results, Norway spruce stands may perform a stable function of carbon sequestration in the soil at these sites, especially in the mineral soil horizon.

Effects of topography on soil organic carbon stocks in grasslands of a semiarid alpine region, northwestern China

Soil organic carbon (SOC) in mountainous regions is characterized by strong topography-induced heterogeneity, which may contribute to large uncertainties in regional SOC stock estimation. However, the quantitative effects of topography on SOC stocks in semiarid alpine grasslands are currently not well understood. Therefore, the purpose of this research study is to determine the role of topography in shaping the spatial patterns of SOC stocks. Soils from the summit, shoulder, backslope, footslope, and toeslope positions along nine toposequences within three elevation-dependent grassland types (i.e., montane desert steppe at ~ 2450 m, montane steppe at ~ 2900 m, and subalpine meadow at ~ 3350 m) are sampled at four depths (0–10, 10–20, 20–40, and 40–60 cm). SOC content, bulk density, soil texture, soil water content, and grassland biomass are determined. The general linear model (GLM) is employed to quantify the effects of topography on the SOC stocks. Ordinary least squares regressions are performed to explore the underlying relationships between SOC stocks and the other edaphic factors. In accordance with the present results, the SOC stocks at 0–60 cm show an increasing trend in respect to the elevation zone, with the highest stock being approximately 37.70 g m−2 in the subalpine meadow, about 2.07 and 3.41 times larger than that in the montane steppe and montane desert steppe, respectively. Along the toposequences, it is revealed the SOC stocks are maximal at toeslope, reaching to 14.98, 31.76, and 49.52 kg m−2, which are also significantly larger than those at the shoulder by a factor of 1.38, 2.31, and 1.44, in montane desert steppe, montane steppe, and subalpine meadow, respectively. Topography totally is seen to explain about 84% of the overall variation in SOC stocks, of which 70.61 and 9.74% are attributed to elevation zone and slope position, while the slope aspect and slope gradient are seen to plausibly explain only about 1.84 and 0.01%, respectively. The elevation zone and the slope position are seen to markedly shape the spatial patterns of the SOC stocks, and thus, they may be considered as key indicating factors in constructing the optimal SOC estimation model in such semiarid alpine grasslands.

Uncertainty in CENTURY-modelled changes in soil organic carbon stock in the uplands of Northeast China, 1980–2050

Process-based models have been successfully applied to predict long-term changes in soil organic carbon (SOC) at plot scales, but considerable uncertainties are still introduced into regional or national extrapolations due to the lack of spatially explicit information on the model input parameters. Using the CENTURY model we predicted SOC changes in the uplands of Northeast China during the period from 1980 to 2050 and provided 95% confidence intervals regarding the uncertainties associated with variability in the key input parameters. Regional SOC estimation predicted by CENTURY was reliable for the uplands of Northeast China when considering the uncertainty associated with heterogeneous key input parameters. SOC stocks were estimated to be 0.99, 0.88 and 0.87 Pg C in 1980, 2010 and 2050, with 95% confidence intervals ranging from 0.69 to 1.31, 0.66 to 1.11, and 0.69 to 1.07 Pg C, respectively. Overall, the upland soils of Northeast China functioned as a carbon source from 1980 to 2010, with a net decrease of 106 (9–207) Tg C. The SOC losses mainly occurred where SOC contents were high (Heilongjiang Province and eastern Jilin Province). However, assuming unchanged management, whether the uplands of Northeast China will serve as a carbon sink/source over the next 40 years remains uncertain. Information collection on the most influential input parameters (the initial SOC content and clay content) is critical to reduce uncertainty and to provide meaningful information for decision makers.

Application of Spectrally Derived Soil Type as Ancillary Data to Improve the Estimation of Soil Organic Carbon by Using the Chinese Soil Vis-NIR Spectral Library

Ancillary data, such as soil type, may improve the visible and near-infrared (vis-NIR) estimation of soil organic carbon (SOC); however, they require data collection or expert knowledge. The application of a national soil spectral library to local SOC estimations usually requires soil type information, because the relationships between vis-NIR spectra and SOC from different populations may vary. Using 515 samples of five soil types (genetic soil classification of China, GSCC) from the Chinese soil spectral library (CSSL), we compared three strategies in the vis-NIR estimation of SOC. Different regression models were calibrated using the entire dataset (Strategy I, without using soil type as ancillary data) and the subsets stratified by soil type from CSSL as ancillary data (strategies II and III). In Strategy II, the subsets were stratified by soil type from the CSSL for validation. In Strategy III, the subsets were stratified by spectrally derived soil type for validation. The results showed that 86.72% of the samples were successfully discriminated for the soil types by using the vis-NIR spectra. The coefficients of determination in the prediction ( R p 2 ) of SOC estimation by strategies I, II, and III were 0.74, 0.83, and 0.82, respectively. The stratified calibration strategies (strategies II and III) improved the vis-NIR estimation of SOC. The misclassification of the soil type in the application of Strategy III slightly affected the SOC estimations. Nevertheless, this strategy is inexpensive and beneficial when expert knowledge on soil classification is lacking. We concluded that vis-NIR spectroscopy could be applied to distinguish some soil types in terms of GSCC, which further provided essential and easily accessible ancillary data for the application of stratified calibration strategies in the vis-NIR estimation of SOC.

Airborne Laser Scanning Cartography of On-Site Carbon Stocks as a Basis for the Silviculture of Pinus Halepensis Plantations

Forest managers are interested in forest-monitoring strategies using low density Airborne Laser Scanning (ALS). However, little research has used ALS to estimate soil organic carbon (SOC) as a criterion for operational thinning. Our objective was to compare three different thinning intensities in terms of the on-site C stock after 13 years (2004–2017) and to develop models of biomass (Wt, Mg ha−1) and SOC (Mg ha−1) in Pinus halepensis forest, based on low density ALS in southern Spain. ALS was performed for the area and stand metrics were measured within 83 plots. Non-parametric kNN models were developed to estimate Wt and SOC. The overall C stock was significantly higher in plots subjected to heavy or moderate thinning (101.17 Mg ha−1 and 100.94 Mg ha−1, respectively) than in the control plots (91.83 Mg ha−1). The best Wt and SOC models provided R2 values of 0.82 (Wt, MSNPP) and 0.82 (SOC-S10, RAW). The study area will be able to stock 134,850 Mg of C under a non-intervention scenario and 157,958 Mg of C under the heavy thinning scenario. High-resolution cartography of the predicted C stock is useful for silvicultural planning and may be used for proper management to increase C sequestration in dry P. halepensis forests.

Machine-learning-based quantitative estimation of soil organic carbon content by VIS/NIR spectroscopy.

Soil organic carbon (SOC) is an important soil property that has profound impact on soil quality and plant growth. With 140 soil samples collected from Ebinur Lake Wetland National Nature Reserve, Xinjiang Uyghur Autonomous Region of China, this research evaluated the feasibility of visible/near infrared (VIS/NIR) spectroscopy data (350–2,500 nm) and simulated EO-1 Hyperion data to estimate SOC in arid wetland regions. Three machine learning algorithms including Ant Colony Optimization-interval Partial Least Squares (ACO-iPLS), Recursive Feature Elimination-Support Vector Machine (RF-SVM), and Random Forest (RF) were employed to select spectral features and further estimate SOC. Results indicated that the feature wavelengths pertaining to SOC were mainly within the ranges of 745–910 nm and 1,911–2,254 nm. The combination of RF-SVM and first derivative pre-processing produced the highest estimation accuracy with the optimal values of Rt (correlation coefficient of testing set), RMSEt and RPD of 0.91, 0.27% and 2.41, respectively. The simulated EO-1 Hyperion data combined with Support Vector Machine (SVM) based recursive feature elimination algorithm produced the most accurate estimate of SOC content. For the testing set, Rt was 0.79, RMSEt was 0.19%, and RPD was 1.61. This practice provides an efficient, low-cost approach with potentially high accuracy to estimate SOC contents and hence supports better management and protection strategies for desert wetland ecosystems.

Predicting Soil Organic Carbon and Total Nitrogen at the Farm Scale Using Quantitative Color Sensor Measurements

Sensor technology can be a reliable and inexpensive means of gathering soils data for soil health assessment at the farm scale. This study demonstrates the use of color system readings from the Nix ProTM color sensor (Nix Sensor Ltd., Hamilton, ON, Canada) to predict soil organic carbon (SOC) as well as total nitrogen (TN) in variable, glacial till soils at the 147 ha Cornell University Willsboro Research Farm, located in Upstate New York, USA. Regression analysis was conducted using the natural log of SOC (lnSOC) and the natural log of TN (lnTN) as dependent variables, and sample depth and color data were used as predictors for 155 air dried soil samples. Analysis was conducted for combined samples, Alfisols, and Entisols as separate sample sets and separate models were developed using depth and color variables, and color variables only. Depth and L* were significant predictors of lnSOC and lnTN for all sample sets. The color variable b* was not a significant predictor of lnSOC for any soil sample set, but it was for lnTN for all sample sets. The lnSOC prediction model for Alfisols, which included depth, had the highest R2 value (0.81, p-value < 0.001). The lnSOC model for Entisols, which contained only color variables, had the lowest R2 (0.62, p-value < 0.001). The results suggest that the Nix ProTM color sensor is an effective tool for the rapid assessment of SOC and TN content for these soils. With the accuracy and low cost of this sensor technology, it will be possible to greatly increase the spatial and temporal density of SOC and TN estimates, which is critical for soil management.

Prediction of soil organic carbon stock by laboratory spectral data and airborne hyperspectral images

Soil organic carbon (SOC) plays an important role in controlling the function and quality of soil and offsetting the emissions of greenhouse gases. However, the dynamic monitoring and estimation of SOC are very difficult due to the complex traditional methods and the changing environmental variables. For instance, the calculation of SOC stock requires measurement of a few relevant soil attributes, such as soil organic matter (SOM), soil bulk density (SBD), soil moisture, and soil weight, in the laboratory. Many studies have suggested that visible and near-infrared (vis–NIR) spectra are a practical and affordable alternative to accurately and rapidly estimate the soil attributes relevant to SOC stock, and airborne hyperspectral images can be used as a valuable data source to perform digital soil mapping with high spatial resolution. The objective of this research was to check the predicted capability of SOC stock through laboratory and airborne vis–NIR spectral data. A total of 50 topsoil samples (0–15 cm) from the farmland of Iowa City were used as the study object. The partial least squares regression model was used to predict SOC stock through the direct and indirect methods. In the direct method, the SOC stock was predicted using the spectral data. In the indirect method, the relevant soil properties (SOM and SBD) of the SOC stock were predicted using the spectral data, and then the SOC stock was calculated. The mechanism of the prediction methods and the potential influencing factors of the model performance were discussed from the aspect of electromagnetic theory and empirical statistics. Results showed the following: (i) SOC stock can be successfully predicted using the laboratory spectra and the airborne hyperspectral image through the direct and indirect methods; (ii) the SOC stock and its relevant soil properties (SOM and SBD) showed evident spectral absorption characteristics in the vis–NIR spectral band; (iii) the atmospheric environment and soil surface conditions were the main influencing factors of the prediction accuracy between the airborne and laboratory spectra. This research might be useful for the dynamic monitoring and modeling of SOC in agricultural and environmental fields.

National soil organic carbon estimates can improve global estimates

Global inventories of spatial and vertical distribution of soil organic carbon (SOC) stocks are being used in national and global initiatives targeted to mitigate climate change and land degradation impacts. Yet, national level high-resolution estimates of SOC stocks can be useful for improving the accuracy of global SOC inventories. We estimated spatially resolved SOC stocks of surface 0–30 cm and subsurface 30–100 cm layers at a spatial resolution of 30 m in tropical Island, Sri Lanka using a legacy harmonized soil database of 122 soil profiles. The national estimates were compared with two global estimates derived from WISE30sec and SoilGrids250m. The tropical Island (land area = 64,610 km2) occupying 0.03% of global land area showed a considerable heterogeneity in SOC stocks ranging from 2.0–342.5 Mg ha−1 and 2.7–391.7 Mg ha−1 in the surface and subsurface soil layers, respectively. We found, elevation, precipitation and slope angle as main environmental controllers of the spatial distribution of SOC stocks under tropical climate. Incorporating the pedogenic information (derived from soil series level legacy map, soil orders and suborders) with environmental controllers resulted in better regression models of predicting surface (R2 = 0.61) and subsurface (R2 = 0.81) SOC stocks. Geographically weighted regression kriging derived maps of SOC stocks revealed that 0–100 cm soil layer of the tropical Island stored 500 Tg C contributing for 0.04% of the global SOC stocks. The validation results of our estimates showed low Mean Estimation Error (MEE: surface −1.6 and subsurface −1.6 Mg ha−1) and Root Mean Square Error (RMSE: surface 29.5 and subsurface 24.9 Mg ha−1) indicating a low bias and satisfactory predictions. The relative improvement of the prediction accuracy of the SOC stocks of our geospatial estimates in the 0–30 cm layer in comparison to SoilGrids250m and WISE30sec data derived SOC stocks were 51.7% and 35.2%, respectively. The SOC stocks predictions of the 30–100 cm soil layer showed even better relative improvement compared to SoilGrids250m (78.4%) and WISE30sec (57.4%) SOC estimates. Compared to estimates of total SOC stocks resulted in this study, WISE30sec data derived SOC stock maps showed 30% over estimation of the C stock in surface 0–30 cm (332 Tg C) and 41% overestimation in 30–100 cm layer (343 Tg C). The over estimation of total SOC stocks by the SoilGrids250 SOC stocks map for the surface 0–30 cm layer was 122% (567 Tg C) and for the 30–100 cm layer it was 209% (750 Tg C). We conclude that the fusion of legacy soil information of SOC stocks with appropriate environmental covariates and pedogenic information derived from legacy area-class soil maps at national level can produce more accurate inventories of spatial and vertical distribution of SOC stocks. These national inventories have a great potential of upgrading global inventories of SOC stocks.

The use of pedo-transfer functions for estimating soil organic carbon contents in maize cropland ecosystem in the Coastal Plains of Tanzania

Soil organic carbon (OC) plays a vital role on physico-chemical and biological properties of soils and on climate change regulation. The use of pedo-transfer functions from easily available soil properties for estimating soil OC could be fast and cheap when considering field and laboratory work implications especially in Sub Saharan Africa including Tanzania. This paper attempts to develop a model for estimating soil OC contents under maize croplands ecosystem using pedo-transfer functions from soil texture. A total of 100 epipedon data entries were randomly collected from the previous soil sampling works that were conducted under maize croplands in coastal plains of Tanzania. Eighty percent of the collected data were used for training the model by using multiple regression analysis while the remained 20% were used to validate the model. The results indicated that, clay and silt had significant (p < 0.001) positive correlation with soil OC while sand contents in soils had negative (p < 0.001) correlation with soil OC. All together clay, sand and silt were revealed powerful predictors (p < 0.001, R2 = 0.82) of OC content in soils. On validation, the soil OC predicted agreed by 81.2% with the soil OC determined by laboratory test. These results imply that pedo-transfer functions for predicting soil OC based on soil texture is not only fast and cheap but also is an effective option for estimating OC content in soils.

Using present and past climosequences to estimate soil organic carbon and related physical quality indicators under future climatic conditions

This research aimed at testing the use of present and past climosequences to estimate soil organic carbon (SOC) and related physical quality indicators under future climatic conditions. The influence of climate on soil features was studied for four combinations of typical Mediterranean soil types and cropping systems, placed along climosequences of the past (P1: 1961–1990), present (P2: 1981–2010) and future (P3: 2021–2050). The four test areas were located in Italy, each one characterized by the same soil typology and cropping system, placed on similar morphological position and parent material, wide enough to cross climatic boundaries. Legacy soil profiles that were sampled in the P1 time-period were re-sampled in 2010–2011, to check for possible variations in soil characteristics. Besides SOC content and stock (Cstock), we examined some physical quality indicators for which the existence of relations with SOC is well-known, namely soil compaction and soil crusting susceptibility, soil erodibility, and soil loss by water erosion.Among several climatic indexes, the de Martonne index (IDM) resulted the most correlated with SOC. The IDM vs. SOC relationship was significant and not different in both P1 and P2 climosequences, highlighting the temporal stability of the relation between climate and SOC content. In the Vertisols of Sicily, cultivated under cereals, the P3 climosequence predicted a SOC reduction of more than 11%. This will lead to an increase of soil erodibility, susceptibility to compaction, and surface crust formation. On the contrary, in the Luvisols under forage crops of the Po Plain, a substantial Cstock increase (28.8%) is expected, with a consequent improvement in soil physical indicators. For the Luvisols under meadows of Sardinia, an increase in erosion of 13.5% is expected, because of increased precipitation volume (7.4%) and aggressiveness. In the Andosols under olive trees of Campania there is a predicted reduction in the Cstock (−6.3%) and an associated increase in soil loss (4.6%), while no marked variation is expected for the other soil physical indicators.We can conclude that climosequences are a useful tool to predict the future dynamics of some soil physical characteristics affected by climate change. Cstock and soil loss by water erosion are expected to change significantly under future climatic conditions, while minor changes are observed for erodibility, compaction and crusting susceptibility, even when SOC variations are significant. In the climosequences, the considered soil physical quality indicators resulted proportionally more affected by the cropping system than by the climate and, within the same cropping system, more variable according to the climate than the time. This outcome confirms the fundamental role of soil physics in controlling the resilience of the soil system to climate change.

No silver bullet for digital soil mapping: country-specific soil organic carbon estimates across Latin America

Abstract. Country-specific soil organic carbon (SOC) estimates are the baseline for the Global SOC Map of the Global Soil Partnership (GSOCmap-GSP). This endeavor is key to explaining the uncertainty of global SOC estimates but requires harmonizing heterogeneous datasets and building country-specific capacities for digital soil mapping (DSM). We identified country-specific predictors for SOC and tested the performance of five predictive algorithms for mapping SOC across Latin America. The algorithms included support vector machines (SVMs), random forest (RF), kernel-weighted nearest neighbors (KK), partial least squares regression (PL), and regression kriging based on stepwise multiple linear models (RK). Country-specific training data and SOC predictors (5 × 5 km pixel resolution) were obtained from ISRIC – World Soil Information. Temperature, soil type, vegetation indices, and topographic constraints were the best predictors for SOC, but country-specific predictors and their respective weights varied across Latin America. We compared a large diversity of country-specific datasets and models, and were able to explain SOC variability in a range between ∼ 1 and ∼ 60 %, with no universal predictive algorithm among countries. A regional (n = 11 268 SOC estimates) ensemble of these five algorithms was able to explain ∼ 39 % of SOC variability from repeated 5-fold cross-validation. We report a combined SOC stock of 77.8 ± 43.6 Pg (uncertainty represented by the full conditional response of independent model residuals) across Latin America. SOC stocks were higher in tropical forests (30 ± 16.5 Pg) and croplands (13 ± 8.1 Pg). Country-specific and regional ensembles revealed spatial discrepancies across geopolitical borders, higher elevations, and coastal plains, but provided similar regional stocks (77.8 ± 42.2 and 76.8 ± 45.1 Pg, respectively). These results are conservative compared to global estimates (e.g., SoilGrids250m 185.8 Pg, the Harmonized World Soil Database 138.4 Pg, or the GSOCmap-GSP 99.7 Pg). Countries with large area (i.e., Brazil, Bolivia, Mexico, Peru) and large spatial SOC heterogeneity had lower SOC stocks per unit area and larger uncertainty in their predictions. We highlight that expert opinion is needed to set boundary prediction limits to avoid unrealistically high modeling estimates. For maximizing explained variance while minimizing prediction bias, the selection of predictive algorithms for SOC mapping should consider density of available data and variability of country-specific environmental gradients. This study highlights the large degree of spatial uncertainty in SOC estimates across Latin America. We provide a framework for improving country-specific mapping efforts and reducing current discrepancy of global, regional, and country-specific SOC estimates.

Use of different chemometric approaches for an estimation of soil properties at field scale with near infrared spectroscopy

AbstractVisible and near infrared spectroscopy (vis‐NIRS) may be useful for an estimation of soil properties in arable fields, but the quality of results are often variable depending on the applied chemometric approach. Partial least squares regression (PLSR) may be replaced by approaches which employ supervised learning methods or variable selection procedures in order to increase the proportion of informative wavelengths used in the estimation procedure, to reduce the noise of the spectra and to find the best fitting solution. Objectives were (1) to compare the usefulness of PLSR with either PLSR combined with a genetic algorithm (GA‐PLSR) or support vector machine regression (SVMR) for an estimation of soil organic carbon (SOC), total nitrogen (N), pH, cation exchange capacity (CEC) and soil texture for surface soils (0–5 cm, n = 144) of an arable field in Bangalore (India) and (2) to test and optimize different calibration strategies for GA‐PLSR for an improved estimation of soil properties. PLSR was useful for an estimation of SOC, N, sand and clay. In the cross‐validation (n = 96), accuracies of estimated soil properties generally decreased in the order GA‐PLSR &gt; SVMR &gt; PLSR. However, the order of estimation accuracies for the random validation sample (n = 48) changed to SVMR &gt; GA‐PLSR &gt; PLSR for SOC, N, pH, and CEC, whereas for clay the order changed to SVMR &gt; PLSR &gt; GA‐PLSR. A sequential procedure, which used the most frequently selected wavelengths of the GA‐PLSR runs, proved to be useful for an improved estimation of SOC and N. Overall, SVMR especially improved estimations of SOC and clay, whereas GA‐PLSR was particularly useful for SOC and N and it was the only approach which successfully estimated CEC in cross‐validation and validation.

Transferability of Vis‐NIR models for Soil Organic Carbon Estimation between Two Study Areas by using Spiking

Core Ideas Vis‐NIR spectroscopy was applied to estimate soil organic carbon in two different study areas. Model developed in one study area couldn't be transferred in the other area. We used spiking strategy to adapt the calibration models to target area characteristics. |s2–σ2| could be used to identify the optimal spiking samples to improve model transferability. Visible and near infrared (Vis‐NIR) spectroscopy technique has been shown to be a cost‐effective alternative for rapidly analyzing soil organic carbon (SOC). However, great challenges remain when applying a Vis‐NIR model for SOC estimation developed in one study area to other study areas without further calibration. The scope of this study was to use spiking strategy to improve the transferability of Vis‐NIR models between two study areas. Specifically, we explored the optimal spiking subset by adding different quantities of spiking samples to construct different‐sized models, and the strategy of spiking with extra‐weighting was used for comparison. Soil data was acquired in two independent study areas (WH area and HH area) in Hubei Province, Central China. The reflectance spectra and SOC contents were measured in the laboratory. Partial least squares regression (PLSR) was used for model calibration. The representativeness of the spiking samples was assessed through the absolute difference between the selected sample variance (s2) and the original variance (σ2) in the principal component space derived from soil spectra. Results showed that the initial models yielded successful SOC predictions for the soil samples from the same area as the calibration samples, but failed in those samples from the other area. Spiking improved the model transferability between these two study areas. Approximately 33%/48% of the HH/WH calibration set was required as spiking samples in model calibrations and applications in the other area. Spiking with extra‐weighting was of limited use in small‐sized spectral libraries. The use of |s2–σ2| is potentially effective in identifying the optimal spiking samples to improve model transferability between different small‐sized study areas in the Vis‐NIR assessments of SOC.

Predicting Soil Organic Carbon in Agroecosystems under Climate Change

Predicting Soil Organic Carbon in Agroecosystems under Climate Change

Soil Carbon Response to Projected Climate Change in the US Western Corn Belt.

The western US Corn Belt is projected to experience major changes in growing conditions due to climate change over the next 50 to 100 yr. Projected changes include increases in growing season length, number of high temperature stress days and warm nights, and precipitation, with more heavy rainfall events. The impact these changes will have on soil organic carbon (SOC) needs to be estimated and adaptive changes in management developed to sustain soil health and system services. The process-based model CQESTR was used to model changes in SOC stocks (0-30 cm) of continuous corn ( L.) and a corn-soybean [ (L.) Merr.] rotation under disk, chisel, ridge, and no-tillage using projected growing season conditions for the next 50 yr. Input for the model was based on management and harvest records from a long-term tillage study (1986-2015) in eastern Nebraska, and model output was validated using measured changes in SOC from 1999 to 2011 in the study. The validated model was used to estimate changes in SOC over 17 yr under climatic conditions projected for 2065 under two scenarios: (i) crop yields increasing at the observed rate from 1971 to 2016 or (ii) crop yields reduced due to negative effects of increasing temperature. CQESTR estimates of SOC agreed well with measured SOC ( = 0.70, < 0.0001). Validated model simulated changes in SOC under projected climate change differed among the three soil depths (0-7.5, 7.5-15, and 15-30 cm). Summed over the 0- to 30-cm depth, there were significant three-way interactions of year × rotation × yield ( = 0.014) and year × tillage × yield ( < 0.001). As yield increased, SOC increased under no-tillage continuous corn but was unchanged under no-tillage corn-soybean and ridge tillage regardless of cropping system. Under chisel and disk tillage, SOC declined regardless of cropping system. With declining yields SOC decreased regardless of tillage or cropping system. These results highlight the interaction between genetics and management in maintaining yield trends and soil C.