Search

Soil spatial information has traditionally been presented as polygon maps at coarse scales. Solving global and local issues, including food security, water regulation, land degradation, and climate change requires higher quality, more consistent and detailed soil information. Accurate prediction of soil variation over large and complex areas with limited samples remains a challenge, which is especially significant for China due to its vast land area which contains the most diverse soil landscapes in the world. Here, we integrated predictive soil mapping paradigm with adaptive depth function fitting, state-of-the-art ensemble machine learning and high-resolution soil-forming environment characterization in a high-performance parallel computing environment to generate 90-m resolution national gridded maps of nine soil properties (pH, organic carbon, nitrogen, phosphorus, potassium, cation exchange capacity, bulk density, coarse fragments, and thickness) at multiple depths across China. This was based on approximately 5000 representative soil profiles collected in a recent national soil survey and a suite of detailed covariates to characterize soil-forming environments. The predictive accuracy ranged from very good to moderate (Model Efficiency Coefficients from 0.71 to 0.36) at 0–5 cm. The predictive accuracy for most soil properties declined with depth. Compared with previous soil maps, we achieved significantly more detailed and accurate predictions which could well represent soil variations across the territory and are a significant contribution to the GlobalSoilMap.net project. The relative importance of soil-forming factors in the predictions varied by specific soil property and depth, suggesting the complexity and non-stationarity of comprehensive multi-factor interactions in the process of soil development.

Depth patterns of soil microbial distribution have not been well characterized and little is known about the balance between stochastic and deterministic processes in shaping microbial community th...

Organic carbon (OC) plays a pivotal role in earth surface systems. However, current three-dimensional (3D) mapping studies usually focus on a soil depth of 1 m rather than the depth to the bedrock. A top-down method using piecewise depth functions was proposed to fit the OC vertical decline patterns in a subtropical catchment in southern China. The vertical variation in OC was greatly affected by the heterogeneous topsoil due to natural processes and anthropogenic disturbances. Thus, topsoil OC maps were produced and utilized as covariates to indicate the OC decline rates and to benefit the 3D OC simulation. A distribution map of the underground critical zone thickness (UCZT) was applied as a lower boundary for the 3D simulation. Six widely used mapping techniques were performed to predict the spatial distribution of topsoil OC and depth function parameters. The overall cross-validation results showed a root mean squared error (RMSE) of 1.7 g kg–¹ and a ratio of performance to deviation (RPD) of 1.82. Given limited boreholes, validation showed that the depth function performed better in the lower part (>1 m) than in the upper part (

Nitrate accumulation has been reported in the top 1 m and subsurface soil (> 1 m) across arid to semi-humid regions, but not in humid regions. Nitrate inventories through the whole regolith, referred to collectively as soil and saprolite, in humid regions have received little attention to date, likely due to previously assumed low nitrification rates and large nitrogen (N) losses by severe surface runoff and erosion. In order to understand if and how reactive N exists in the below ground (soil and saprolite) in humid environment, the amount of NO3--N, NH4+-N and dissolved organic N (DON) present in the regolith to a depth of 9 m in a typical red soil Critical Zone was investigated under different land uses (upland, woodland and paddy field). The Red Soil Critical Zone Observatory is located in the subtropical Jiangxi Province, China, with a mean annual precipitation of 1795 mm and mean annual potential evapotranspiration of 1229 mm. The examined regoliths were acidic, highly weathered, and mainly clay loam in texture. Results showed that on average 92% (827 ± 97 kg N ha-1) of NO3--N and 82% (521 ± 153 kg N ha-1) of DON were stored at depth (from a depth of 1 m to the bedrock surface) in the upland regolith, while 92% (283 kg N ha-1) of NO3--N and 78% (820 kg N ha-1) of DON were stored at depth in the woodland regolith. Nitrate N significantly accumulated with depth in the upland regolith from the 1- to 4-m depth interval (p < 0.01), while the inventory (632 ± 75 kg N ha-1) in the top 3-m zone accounted for on average 71% of the total. Dissolved organic N significantly accumulated with depth in the upland regolith from the 0- to 3-m depth interval (p < 0.01), while the inventory (408 ± 75 kg N ha-1) in the top 3-m zone accounted for on average 64% of the total. There was no significant accumulation for NH4+-N throughout the upland regolith (p = 0.35). No substantial accumulation of dissolved N was measured at depth in paddy field regoliths with different cultivation ages. The finding that large reservoirs of reactive N can exist in deep regolith rather than in the routinely investigated solum of subtropical regions shows a missing part of the terrestrial N budget and raises concerns about potential 34 groundwater nitrate pollution.

Modern monothetic hierarchical soil classification systems such as the Chinese Soil Taxonomy (CST) are semi‐quantitative and their future development is likely to trend towards a fully quantitative system using the concepts of numerical taxonomy. Previous researchers have calculated the taxonomic distances between individual soils based on soil physiochemical properties, not, however, based on spectra of full soil profiles with different horizons. We hypothesized that numerical taxonomy implemented by cluster analysis of the taxonomic distance matrix based on vis–NIR spectra would accord with some CST Orders assigned by pedologists, and not with others, depending on how closely spectral features represent the diagnostic features used in the classification. Taxonomic distances in spectral space were computed for all pairs of 191 profiles, resulting in a distance matrix on which hierarchical cluster analysis was performed. Different indices were calculated to determine the optimum number of clusters, resulting in four spectral soil classes. These were then compared with CST Orders assigned to the profiles by expert allocation. The numerical classes and CST Orders matched poorly because of the completely different classification philosophies behind numerical taxonomy and the CST, which is based largely on presumed genesis and uses sharp thresholds leading to very similar soils being allocated to different classes. Thus, we consider the numerical classification as information that is complementary to the monothetic hierarchical system. Numerical classification can reveal the taxonomic objective aspects of the relation between the defined classes and can suggest new groupings. HIGHLIGHTS: Taxonomic distances can serve as an objective measure of soil similarity. Soil spectra are a good candidate for numerical classification based on taxonomic distances. The conceptual basis of Chinese Soil Taxonomy (CST) does not match that of taxonomic distance. Numerical classification based on spectra can suggest revisions to the CST.

Soil diagnostic horizons, which each have a set of quantified properties, play a key role in soil classification. However, they are difficult to predict, and few attempts have been made to map their spatial occurrence. We evaluated and compared four machine learning algorithms, namely, the classification and regression tree (CART), random forest (RF), boosted regression trees (BRT), and support vector machine (SVM), to map the occurrence of the soil mattic horizon in the northeastern Qinghai-Tibetan Plateau using readily available ancillary data. The mechanisms of resampling and ensemble techniques significantly improved prediction accuracies (measured based on area under the receiver operator characteristic curve score (AUC)) and produced more stable results for the BRT (AUC of 0.921 ± 0.012, mean ± standard deviation) and RF (0.908 ± 0.013) algorithms compared to the CART algorithm (0.784 ± 0.012), which is the most commonly used machine learning method. Although the SVM algorithm yielded a comparable AUC value (0.906 ± 0.006) to the RF and BRT algorithms, it is sensitive to parameter settings, which are extremely time-consuming. Therefore, we consider it inadequate for occurrence-distribution modeling. Considering the obvious advantages of high prediction accuracy, robustness to parameter settings, the ability to estimate uncertainty in prediction, and easy interpretation of predictor variables, BRT seems to be the most desirable method. These results provide an insight into the use of machine learning algorithms to map the mattic horizon and potentially other soil diagnostic horizons.

Nitrate (NO3−) leaching in agroecosystems has caused much concern worldwide due to its negative environmental and health impacts. To evaluate the effect of fertilization on NO3− leaching and soil N accumulation, a two-year 15N tracing study was conducted in subtropical China with field lysimeters packed with non-destructive sampling red soil. 200 kg N ha−1 yr−1 urea (15N abundance of 10%) was applied for maize crops. Fertilization promoted NO3− leaching by 91.5 ± 6.1 and 57.9 ± 15.2 kg N ha−1 yr−1 at 20 and 100 cm depth, respectively. Soil organic nitrogen (SON) pool was the main NO3− source (>60%), especially at surface soil. Fertilizer contributed 19.4 ± 2.0 and 32.8 ± 1.9% to NO3− leaching. At 20 cm depth, besides NO3− leaching accelerated by fertilization during the crop growth period (51.8%), mineralization of SON also resulted in abundant NO3− leaching during the fallow period (48.2%). At 100 cm depth, NO3− leaching significantly increased with the fertilization year due to the continuous NO3− leaching and the delay of NO3− leaching by soil NO3− adsorption. After two-year fertilization, 55.3 ± 2.2% of the applied N accumulated in the soil, leading to a soil N pool increase of 110.0 ± 10.1 kg N ha−1 yr−1. If the fertilization was maintained, the continuous N accumulation poses a potential threat to groundwater quality or drinking water safety. In subtropical red soil regions, to reduce NO3− leaching, NO3− leaching during both the crop growth and fallow period should be taken seriously, and effective management practices, such as cover crops and intensive tile drainage systems, should be carried out to reduce the fertilizer N accumulation and the contribution of SON.

Soil colour can indicate soil physical, chemical and biological properties and processes, and is an important indicator for soil classification, soil quality evaluation and soil management. It varies in both horizontal and vertical dimensions, and thus regional maps of soil colour can reveal spatial patterns of these properties, processes, and indicators. However, although soil regions are sometimes named for their dominant soil colour, it is directly measured only at “point” support, i.e., during soil profile description, whereas it is desirable to know soil colour over the entire soilscape. To achieve this for China we used predictive soil mapping methods to produce soil colour maps (dry and moist colours) at 1 km2 grid cell size and over multiple depths from a consistent dataset of approximately 4 600 full profile descriptions taken as part of a national survey to define soil series in Chinese Soil Taxonomy, and a set of environmental covariates covering the national territory. The covariates characterized soil forming factors including climate, parent materials, terrain, vegetation, land surface water and thermal conditions. Soil colour descriptions in the Munsell system were extracted from the genetic horizon descriptions at the selected depths and converted to the sRGB and L*a*b* colour spaces. Dry and moist colour separates were not well-correlated in either space ( r 0.76 ). Random forest models were constructed in both spaces, for dry and moist colours separately. Models in sRGB space were moderately successful ( R 2 ≈ 0.43 , RMSE ≈ 26 / 255 ) at 5 cm, with success decreasing with depth. Models smoothed the colour space and thus did not predict the more extreme values or chromas, nor the rarer hues. Models in L*a*b* space were less successful. The fitted sRGB models were used to produce predictive maps over all of China. Regional patterns as well as local detail are clearly shown. Solar radiation, wind exposure, regolith thickness, and Landsat TM bands 7 and 5 contributed most to the predictions, followed by elevation, mean annual precipitation, terrain wetness index, air temperature seasonality, precipitation standard deviation and standard deviation of NDVI. These suggest pedological processes acting on the development of soil colours, including weathering of parent materials, oxidation-reduction chemistry and biochemistry of the decomposing of organic matter. This study shows that predictive methods from points using suitable covariates are an alternative to spatial predictions over map units from their representative profiles.

Visible‐near infrared (vis–NIR) spectroscopy can reveal various soil properties and facilitate soil classification. However, few studies have attempted to classify vertical soil profiles that contain several genetic horizons. Here, we propose the ‘multiple objectives mixed support vector classification’ (MOM–SVC) method to classify soil profiles. A total of 130 soil profiles were collected from genetic horizons in Zhejiang Province, China. After laboratory analysis, soil profiles were classified according to the Chinese Soil Taxonomy system. Vis–NIR spectra were recorded from each genetic horizon of each soil profile and were then pre‐processed. We performed the MOM–SVC method as follows: (i) created a support vector machine (SVM) model (one‐versus‐one approach) using spectral data from all soil horizons in calibration profiles, (ii) applied the SVM model on each horizon of the profile to be predicted, (iii) extracted ‘votes’ from each horizon and mixed (or summarized) them into the votes of each profile to be predicted and (iv) classified each profile by the majority‐voting method. We also investigated whether the additional input of auxiliary soil information (e.g. moist soil colour, soil organic matter and soil texture), which could be measured easily or be well predicted by vis–NIR spectroscopy, could improve the accuracy of soil classification when combined with it. Independent validation results showed that the MOM–SVC method performed better at the soil order level than at the suborder level. Adding auxiliary soil information to the classification model improved the overall accuracy of classification at the soil order level. The proposed MOM–SVC method provides a fast objective diagnostic of soil classes for use in soil surveys and can help to update soil databases when a more objective soil classification system is developed. HIGHLIGHTS: The MOM–SVC method can be used to classify soil profiles objectively with a variety of soil horizons. Stratified random sampling was used to quantify prediction uncertainty in classification MOM–SVC can predict soil orders with greater accuracy than suborders. Adding auxiliary soil information into the classification model improved prediction accuracy.

Up-to-date maps of soil organic carbon (SOC) concentrations can provide vital information for monitoring global or regional soil C changes and soil quality. In this study, a national soil dataset collected in the 2010 s was applied to produce SOC maps of mainland China at soil depths of 0–5 cm, 5–15 cm, 15–30 cm, 30–60 cm, 60–100 cm and 100–200 cm. A stacking ensemble learning framework was utilized to take advantage of the optimal predictions from individual models. A voting-based ensemble learning model (VELM) was proposed with consideration of pedoclimatic zones. In this model, three machine learning models were separately trained for every pedoclimatic zone, and their predictions were selectively merged together. A weighted ensemble learning model (WELM), in which the parameterization considered all zones (i.e., the whole study area) simultaneously, was also trained for comparison. The overall R2 values of these two methods ranged from 0.16 to 0.57 and decreased with depth. Based on the independent validation, the R2 values ranged from 0.41 to 0.57 in the topsoil (0–5 cm, 5–15 cm and 15–30 cm). Overall accuracy metrics implied that the VELM and WELM yielded nearly the same prediction performances. However, model validation in the pedoclimatic zones showed that the VELM obviously outperformed the WELM, with the VELM generally improving the accuracy by 12.6%. Based on the independent validation, we also compared our predictions with other soil map products. Although the spatial patterns were similar, the predicted SOC maps outperformed two other products. The comparison of the two ensemble models should serve as a reminder that if new national or regional soil maps are generated, validation based on pedoclimatic zones or other soil-landscape units may be necessary before applying these maps.

Plants play an important role in Silicon (Si) cycling of terrestrial ecosystems. However, how plant Si is related to soil Si translocation and transformation in subtropical areas with ample weatherable silicates and agricultural activities is largely unknown. Three typical watersheds with different cropland and forest configurations in subtropical China were selected to investigate the biomass of different plant species, including coniferous forest, bamboo and paddy, as well as the geochemical composition of soil, rock, vegetation and water. The main soil types are Lithic Udorthents, Lithic Dystrudepts and Typic Epiaquepts derived from granite. Si concentrations in rock, soil, plant and water, as well as Si cycling fluxes, were measured. The results show that although a large amount of Si is cycled by plants, weathering and runoff contribute similarly to, or even more significantly than, plant cycling. High Si bio-accumulators assimilate and return more Si than low Si bio-accumulators, and also cause more phytolith-Si to be added to the surface soil, which decreases desilification rate of surface soil. Agricultural activities increase silicate weathering. Stream water Si concentrations are not affected by plant Si cycling. Thus, both agricultural activities and plant species affect Si translocation. In contrast to results from strong tropical weathering environments, Si cycling by plants in this subtropical area with abundant weatherable silicates does not dominate Si loss from soil; rather, it promotes silicate weathering and controls Si redistribution in soil profiles.

A paddy soil chronosequence consisting of five profiles derived from calcareous marine sediments with cultivation history from 0 to 1000 years was studied to understand the underlying mechanisms and processes controlling the millennial scale Fe evolution. We evaluated the chronosequencial changes in depth distribution of Fe oxide contents and Fe isotopic compositions. Results showed that paddy soil evolution under the influence of periodic flooding and groundwater fluctuation resulted with time in variations of soil moisture regime and redox condition that control Fe mobilization, translocation and redistribution, leading to enhanced profile differentiation of Fe oxides and measurable Fe isotope fractionation. Total Fe and oxide bound Fe as well as their differentiation between surface and subsurface horizons increased as paddy soils age, leading to the formation of diagnostic horizons and features characterizing Fe distribution and redistribution. Selective extractions showed that the weakly-bound, oxide-bound and silicate bound Fe corresponded to 1–16%, 8–46%, and 52–91% of the total Fe, respectively, and these proportions varied with both time and depth due to the redox-related Fe transformation and translocation. δ56Fe values in the studied paddy soil chronosequence ranged from − 0.01‰ to 0.18‰ and exhibited a strong negative correlation with the logarithm of total Fe concentrations, suggesting mass-dependent Fe isotope fractionation occurred as a result of the preferential removal of lighter Fe isotopes during long-term paddy soil evolution under the predominant reducing conditions. However, the Fe isotopic ratio of a specific paddy soil horizon was a result of a complex interaction of different processes, which were summarized and interpreted in our proposed conceptual model. Comparison of Fe isotopic compositions in the worldwide soils demonstrated that Fe isotopes can evidence Fe transfer and pinpoint the factors and processes that control Fe mobilization and redistribution particularly in soils with changing moisture regimes and redox conditions. Our findings provide new insights into the behavior and geochemical cycle of Fe at the Earth's surface strongly affected by human activities and contributes to an improved understanding of how anthropedogenesis affects Fe evolution in the Earth's Critical Zone.

Accurate estimation of soil carbon is essential for accounting carbon cycling on the background of global environment change. However, previous studies made little contribution to the patterns and stocks of soil inorganic carbon (SIC) in large scales. In this study, we defined the structure of the soil depth function to fit vertical distribution of SIC based on pedogenic knowledge across various landscapes. Soil depth functions were constructed from a dataset of 99 soil profiles in the alpine area of the northeastern Tibetan Plateau. The parameters of depth functions were mapped from environmental covariates using random forest. Finally, SIC stocks at three depth intervals in the upper 1 m depth were mapped across the entire study area by applying predicted soil depth functions at each location. The results showed that the soil depth functions were able to improve accuracy for fitting the vertical distribution of the SIC content, with a mean determination coefficient of R2 = 0.93. Overall accuracy for predicted SIC stocks was assessed on training samples. High Lin's concordance correlation coefficient values (0.84–0.86) indicate that predicted and observed values were in good agreement (RMSE: 1.52–1.67 kg m− 2 and ME: − 0.33 to − 0.29 kg m− 2). Variable importance showed that geographic position predictors (longitude, latitude) were key factors predicting the distribution of SIC. Terrain covariates were important variables influencing the three-dimensional distribution of SIC in mountain areas. By applying the proposed approach, the total SIC stock in this area is estimated at 75.41 Tg in the upper 30 cm, 113.15 Tg in the upper 50 cm and 190.30 Tg in the upper 1 m. We concluded that the methodology would be applicable for further prediction of SIC stocks in the Tibetan Plateau or other similar areas.

Multi-scale soil variations are increasingly employed to improve the accuracy for digital soil mapping (DSM). In this study, we attempted to explore a methodology of wavelet analysis on this topic. The terrain attributes of a study area were decomposed using the wavelet analysis, and the resulted components were applied to map soil organic carbon (SOC) content, pH and clay content using multiple linear regression (MLR) and regression kriging (RK). The results showed that the wavelet components strengthened soil-landscape relationships in terms of correlation coefficients, enhanced soil-landscape modelling in terms of MLR modelling coefficients of determination (R2). Compared with several standard DSM approaches, i.e., ordinary kriging (OK), MLR and RK with the original terrain attributes, the use of wavelet components improved the prediction accuracy at some scales, but not all the scales. Most of the improvements were at the slight to moderate levels, e.g., 3.66–14.24% increases in the accuracy based on mean error, mean absolute error, root mean square error and R2. Maps made with wavelet components were relatively smooth and sometimes contained hotspots due to characteristics of wavelet components, which differed a lot from those made by the standard DSM methods. The potential benefits of using wavelet components as predictors in DSM may be further revealed in the future when more predictor selection approaches and mapping methods are considered.

Soil class maps are useful representations of the landscape distribution of holistic soil functions. However these are often only available as generalized classes at small cartographic scales. One reason is that allocating a soil profile to a class in most current soil classification system requires laboratory determination of many diagnostic soil properties. The advantage of reflectance spectroscopy along with the development of spectral libraries can provide a relatively low-cost solution to this problem. Reflectance spectroscopy has demonstrated its ability to rapidly predict soil physio-chemical properties; however prediction accuracy varies among soil properties. When properties predicted with different accuracies are used to substitute for traditional laboratory determinations in allocating a soil profile to a class, the resulting reliability of the allocation is questionable. The objective of this research is to explore whether the soil properties predicted by reflectance spectroscopy can be used to correctly allocate soil profiles into soil taxa at different hierarchical levels. Two hundred and six soil profiles were allocated to eight Orders, 12 Suborders, 23 Groups and 49 Subgroups according to Chinese Soil Taxonomy, with the help of ten soil properties predicted by spectra using ten-fold cross-validated PLSR modelling. The overall allocation accuracy at Order, Suborder, Group and Subgroup level was 98.5%, 98.5%, 87.7% and 76.0% respectively. These results show that soil reflectance spectroscopy can assist in allocation of profiles. When predicted soil properties with varying accuracy are used for soil allocation, propagation of prediction errors and model uncertainties must be considered. We propose the use of multiple indicators (RPD, confidence intervals, comparison of RMSE and threshold requirements) to evaluate the allocation results.

Changes in soil organic matter (SOM) can affect food security, soil and water conservation, and climate change. However, the drivers of changes in SOM in paddy soils of China are not fully understood because the effects of agricultural management and environmental factors are studied separately. Soil, climate, terrain, and agricultural management data from 6 counties selected based on representative soil types and cropping systems in China were used in correlation analysis, analysis of variance, and cforest modeling to analyze the drivers of changes in SOM in paddy soils in the Middle and Lower Yangtze River Plain from 1980 to 2011. The aims of this study were to identify the main factors driving the changes in SOM and to quantitatively evaluate their individual impacts. Results showed that the paddy SOM stock in the study area increased by 12.5% at an average rate of 0.023 kg m−2 year−1 over the 31-year study period. As a result of long-term rice planting, agricultural management practices had a greater influence than soil properties, climate, and terrain. Among the major drivers, straw incorporation, the most influential driver, together with fertilization and tillage practices, significantly increased the accumulation of SOM, while an increase in temperature significantly influenced SOM decomposition. Therefore, to confront the challenge of rising temperatures, it is important to strengthen the positive effects of agricultural management. Rational fertilizer use for stabilizing grain production and crop straw incorporation are promising measures for potential carbon sequestration in this region.

This paper was oriented to characterise spatio-temporal variation of soil organic carbon (SOC) over a 30-year period and to explore its driving forces in East China. A comparative study was done of data obtained during 1980 and 2010 in Anhui Province. Spatio-temporal changes of SOC density and storage in the surface (0–30 cm) and 0–100 cm layers were analysed using geographic information system (GIS) spatial analysis techniques. Relationships between soil erosion, land use change, agricultural management, and changes in SOC are discussed. Results show: (1) over 30 years, the average SOC density decreased by 0.59 kg C m−2 in the surface layer and 1.63 kg C m−2 in 0–100 cm layer. The average SOC density increased in Calcaric Cambisols and Haplic Fluvisols, and decreased in other soil types.The average SOC density in upland increased but decreased as a result of other land use practices. (2) SOC density in the surface layer increased to the north but decreased to the south in Anhui. SOC density in the 0–100 cm layer increased in the central area and decreased to the south. About 59% of all soil increased in SOC density. (3) SOC storage decreased by 117.91 Tg C in the surface layer and decreased by 237.65 Tg C in the 0–100 cm layer. The SOC storage increased mainly in the northern and central areas. (4) Severe soil erosion was a main reason for the large reduction in SOC storage in western and southern Anhui. Cultivated land use that changed to paddy field from upland improved SOC more than changes to other land use types. SOC storage change in farmland was significantly and positively related to organic matter content in crop roots.

High-precision digital soil organic carbon (SOC) stocks mapping is very important for agricultural production management and global carbon cycle. The spatial heterogeneity of farmland SOC is not only influence by the environmental factors of soil formation but also the management practices of tillage, fertilization, and irrigation. However, the traditional modeling covariates of digital soil mapping, such as terrain factors, land use types and climate factors have weak spatial variations in low-relief agricultural areas, and they cannot reflect the large spatial variation of SOC. Thus the time-series multispectral remote sensing images will be used for mapping soil properties in low relief regions in this study, meanwhile a new collaborative verification strategy was put forward to evaluate the spatial distribution characteristics of soil maps. The current study was performed in a nearly flat agricultural region southeast of Iowa (with an area of approximately 385.45 ha), where 195 surface soil samples (0–15 cm) were collected. A hyperspectral image (Headwall-Hyperspec, 380–1700 nm) and the time-series multispectral remote sensing images of Sentinel 2 and Landsat 8 were used to construct the prediction models of SOC stock and its relevant soil properties of SOC and soil bulk density (SBD) through partial least square regression (PLSR) and extreme learning machine (ELM) models. The collected soil samples and evaluation indexes of root mean square error (RMSE), R2, and ratio of performance to interquartile range (RPIQ) were used to evaluate the model performance. Results are as follows: (1) hyperspectral images were successfully used to predict the SOC stock, SOC, and SBD through PLSR and ELM, while ELM (RPIQ = 2.03, 1.97, 1.64) outperformed PLSR (RPIQ = 1.83, 1.97, 1.53); (2) the time-series multispectral remote sensing images of Sentinel 2 and Landsat 8 can reflect the spatial distribution characteristics of the SOC stock, SOC and SBD by PLSR and ELM, but the combination of Sentinel 2 images and ELM obtained the best prediction results (RPIQ = 1.45, 1.25, 1.26); and (3) the differences of the soil maps predicted by the hyperspectral image and time-series multispectral remote sensing images were small, and the largest percentage errors nearly appeared on the edges of the farmland patches owing to mixed pixels. This study further confirmed the good prediction abilities of the time-series multispectral remote sensing images in low relief farmland regions. Lastly, this mapping strategy can provide additional valuable information for agricultural management and carbon cycle.

Soil properties are often assumed to be static over time in hydrological studies, especially in hydrological modelling. Although it is well appreciated that soil structure and its impact on hydraulic properties are time-variable, particularly on cultivated land, very few studies have focused on quantifying the influence of such changes on soil hydrology, especially at the short term (i.e. seasonal). This study explored the value of incorporating such short-term time-variable soil properties in hydrological models. It is based on soil hydraulic properties from temporal field data under no-till done by direct seeding and under conventional cultivation done by ploughing to 0.2 m and harrowing. It uses a controlled tillage experiment in Scotland, on a soil with very good structural stability that experiences gentle rainfall in a temperate oceanic climate (Koppen Cfb). Water retention data were collected from intact soil cores sampled at 0.025, 0.095 and 0.275 m depth at three times between April and August 2013; (i) immediately following tillage, (ii) at barley crop establishment 1 month later and (iii) after harvest. Soil structure varied over time, with no-till soils gaining porosity and ploughed soils losing porosity. We hypothesised that no-till soils would have less seasonal temporal variability, but found it to be comparable to ploughed soils, albeit with pore structure changes following different trends. These changes were reflected in Van Genuchten fitting parameters, which if accounted for in 1-D HYDRUS modelling, had a marked impact on modelled soil water content over time if contrasted to predictions assuming a static pore structure. Using data from multiple sampling events, as opposed to one sampling event, resulted in up to a 44 % difference in soil water content predictions and increased the temporal variability by a factor of 1.5. Hence, our results have demonstrated that it is important to account for short-term temporal variability in soil physical properties in soil water modelling studies, and should not be ignored as a default, particularly on cultivated agricultural soils.

Increasing acid deposition and intense nitrogen (N) fertilization have resulted in severe soil acidification in many regions of the world. However, the soil acidification rate, especially quantitative contributions of different proton (H+) sources remain unclear in various cropping systems. In this study, a two-year field lysimeter experiment was conducted with a set of different N fertilization treatments (0, 100, 200 and 400 kg N ha−1 yr−1) for maize crop in typical red soil in subtropical China. The pathways and budgets of different H+ sources were quantified by calculating the input–output balance of major elements. Results showed that fertilization significantly accelerated soil acidification by increasing the leaching of NO3– and base cations (K+, Ca2+, Na+, and Mg2+) as well as increasing the cation removals by plant uptake and harvest. Unbalanced plant uptake of anions and cations was the primary H+ source (64.4–80.5%) to soils. Plants have an important effect on soil acidification by redistributing cations in soils. Due to cation removals by plant uptake and harvest, there were significant decreases of soil exchangeable base cations under high urea-N treatments (200 and 400 kg N ha−1 yr−1) and significant increases of soil exchangeable H+ and Al3+ under three fertilization treatments. Of course, N transformation also plays an important role (12.1% to 38.8%). However, the effect of direct H+ deposition is minor in this area comparatively (5.78% to 7.34%). Quantification of acidification contribution of different sources is useful for the establishment of further soil remediation practices. The field management is considered very important to alleviate soil acidification, including increasing straw return and controlling fertilizer application.

Mattic epipedon (ME), a special diagnostic surface horizon for soils in alpine meadow, plays an important role in carbon storage, soil water retention, and indication of alpine meadow degradation. At the global scale, it mainly distributes in the Qinghai-Tibetan Plateau and other similar alpine environments. However, its spatial patterns and the relations with environmental conditions remain unknown. This study attempts to explore key environmental variables responsible for the occurrence of the special surface soil horizon and based on those factors to predict and map digitally the spatial distribution of ME. By combining a variable selection procedure and the Random Forest (RF) algorithm, the variables extracted from Landsat 5 TM mosaic image, ASTER GDEM, and climate data were optimized and their importance was measured. The classification accuracy was compared with that obtained from binary logistic regression algorithm. In addition, a land use/land cover (LULC) map-based modification was conducted to further improve the classification accuracy. Results showed that the variable selection procedure had little effect in improving prediction accuracy. However, the number of used variables markedly reduced from 26 to 6 with a 77% decrease, which could speed up the training of the RF model (about one-third of the computation time could be saved). Analysis of variable importance showed that band 3 of TM and normalized difference vegetation index were the most important environmental variables influencing the occurrence of ME. The final overall accuracy of the ME map was predicted to be 84%. Our results demonstrate that the proposed procedure, which combined the proposed variables (derived from remote sensing, GDEM, and climate data), the variable selection approach, the RF algorithm, and the LULC map-based modification method can identify key environmental variables influencing the occurrence of ME and map the spatial distribution of ME effectively on the Qinghai-Tibetan Plateau.

Local terrain attributes, which are derived directly from the digital elevation model, have been widely applied in digital soil mapping. This study aimed to evaluate the mapping accuracy of soil organic carbon (SOC) concentration in 2 zones of the Heihe River in China, by combining prediction methods with local terrain attributes derived from different polynomial models. The prediction accuracy was used as a benchmark for those who may be more concerned with how accurately the variability of soil properties is modeled in practice, rather than how morphometric variables and their geomorphologic interpretations are understood and calculated. In this study, 2 neighborhood types (square and circular) and 6 representative algorithms (Evans-Young, Horn, Zevenbergen-Thorne, Shary, Shi, and Florinsky algorithms) were applied. In general, 35 combinations of first- and second-order derivatives were produced as candidate predictors for soil mapping using two mapping methods (i.e., kriging with an external drift and geographically weighted regression). The results showed that appropriate local terrain attribute algorithms could better capture the spatial variation of SOC concentration in a region where soil properties are strongly influenced by the topography. Among the different combinations of first- and second-order derivatives used, there was a best combination with a more accurate estimate. For different prediction methods, the relative improvement in the two zones varied between 0.30% and 9.68%. The SOC maps resulting from the higher-order algorithms (Zevenbergen-Thorne and Florinsky) yielded less interpolation errors. Therefore, it was concluded that the performance of predictive methods, which incorporated auxiliary variables, could be improved by attempting different terrain analysis algorithms.

Calcareous soils are those soils containing layers of carbonate accumulation formed by either secondary accumulation or inheritance of CaCO 3 from calcareous parent materials. These soils are widely distributed in arid, semi-arid and even some humid climate environments. However, in the Chinese Soil Taxonomy (CST, 3rd version), the soils that contain large amounts of CaCO 3 but do not meet requirements of Aridosols are classified as Hapli-Ustic Cambosols. This group also includes many non-calcareous soils and those soils strongly affected by secondary carbonate accumulation together with quite dissimilar soils in terms of their morphology and properties. This study was conducted to determine calcification patterns and calcic processes occurring in representative pedons of calcareous soils in northwestern Henan Province, China and to classify these soils in CST. A Calcic subgroup was proposed to add within the Hapli-Ustic Cambosols. A diagnostic key was also provided to separate this subgroup from the Typic subgroup. In this way the different calcification degrees of soils were better reflected in the classification, which should lead to more uniform interpretive groups for better soil management.

Summary Holocene loess has been recognized as the primary source of the silty topsoil in the northeast Qinghai-Tibetan Plateau. The processes through which these uniform loess sediments develop into diverse types of soil remain unclear. In this research, we examined 23 loess-derived soil samples from the Qilian Mountains with varying amounts of pedogenic modification. Soil particle-size distribution and non-calcareous mineralogy were changed only slightly because of the weak intensity of chemical weathering. Accumulation of soil organic carbon (SOC) and leaching of carbonate were both identified as predominant pedogenic responses to soil forming processes. Principal component analysis and structural analysis revealed the strong correlations between soil carbon (SOC and carbonate) and several soil properties related to soil functions. Accretion of SOC effectively decreased soil bulk density (R2 = 0.81) and increased cation exchange capacity (R2 = 0.96), soil water retention at saturation (R2 = 0.77), field capacity (R2 = 0.49) and wilting point (R2 = 0.56). These results indicate that soil ecological functions are strengthened during pedogenic modification of such loess sediments. Soil C/N ratio was constant at small SOC contents, but after reaching a threshold of approximately 35 g kg−1 SOC, soil C/N increased linearly with SOC. This indicates a change from a carbon-limited loess ecosystem in arid regions to a nitrogen-limited one in alpine settings. This research suggests that loess sequences within environmental gradients offer great potential as natural experiments to explore intrinsic soil behaviour and ecosystem evolution because the effect of parent material is well constrained. Highlights We examined pedogenic modifications of loess with uniform origin from contrasting environments. Accumulation of SOC and depletion of carbonate coincide during pedogenesis of loess-derived soil. Pedogenesis underpins functional evolution of loess-derived soil across the Qilian Mountains. Loess sequences provide ideal natural experiments to study soil and ecosystem evolution.

Soil organic phosphorus transformation during ecosystem development exerts a crucial influence on soil fertility and ecosystem properties. This paper reviews the use of solution 31P NMR spectroscopy for characterizing organic phosphorus speciation in soil chronosequence and long-term field experiments in order to improve our understanding of the temporal changes, fundamental processes, and associated natural and anthropogenic controls of organic phosphorus transformation during long-term ecosystem evolution. Published soil chronosequence studies show that organic phosphorus compounds under aerobic conditions are dominated by phosphate monoesters (occurred mainly as inositol phosphates) followed by phosphate diesters (occurred mainly as DNA) and phosphonates, irrespective of the different parent materials, vegetation covers and climatic conditions. This contrasted markedly with wetland soils in which phosphate monoesters and diesters maintained approximately equal proportions, which is attributed to the limited reactive clay surfaces for stabilization and/or decomposition of myo-inositol hexakisphosphate under frequent anaerobic conditions. Most organic phosphorus compounds in soil chronosequences increase with age to reach a maximum and then decline with time, although the apex varies significantly among different organic phosphorus compounds and chronosequences. Variations of the potential for phosphorus stabilization resulting from mineralogical transformation, changes in phosphorus sources due to shifts in plant and microbial communities, and differences in the biological utilization of various phosphorus compounds have been suggested as three main mechanisms controlling the temporal changes in organic phosphorus species, abundance and availability during natural ecosystem development. In agricultural soils, the amounts, forms, and dynamics of organic phosphorus are determined by internal soil properties, external environmental conditions and managements, including the history and intensity of land use, different tillage practices and fertilizer treatments. These mechanisms are interlinked and more research is required to isolate both internal and external factors that regulate organic phosphorus transformation in agricultural ecosystems. Given the universal dependence on organic phosphorus for life and its critical roles in biogeochemical cycling, we put forward several open questions that need to be resolved in the future studies by emphasizing the multidisciplinary collaborations, the use of multiple analytical techniques and the establishment of quantitative organic phosphorus transformation models.

Evaluating the accuracy of allocation to classes in monothetic hierarchical soil classification systems, including the World Reference Base for Soil Classification, US Soil Taxonomy, and Chinese Soil Taxonomy, is poorly-served by binomial methods (correct/incorrect allocation per evaluation observation), since some errors are more serious than others in terms of soil properties, map use, pedogenesis, and ease of mapping. Instead, evaluations should account for the taxonomic distance between classes, expressed as class similarities, giving partial credit to some incorrect allocations. These can then be used in weighted accuracy measures, either direct measures of agreement or measures that account for chance agreement, such as the tau index. Similarities can be determined in one of four ways: (1) by the expert opinion of a soil classification specialist; (2) by the distance between classes in a numerical taxonomy assessment; (3) by distance within a taxonomic hierarchy; or (4) by an error loss function. Expert opinion can be from the point of view of the map user, to assess map utility, or map producer, to assess mapping skill. Examples are given of determining similarity between a subset of Chinese Soil Taxonomy classes by expert opinion and by numerical taxonomy from soil spectra, and then using these for weighted accuracy assessment. A method for assessing the accuracy of probabilistic predictions of several classes at a location is also proposed.

High-resolution and detailed regional soil spatial distribution information is increasingly needed for ecological modeling and land resource management. For areas with no point data, regional soil mapping includes two steps: soil sampling and soil mapping. Because sampling over a large area is costly, efficient sampling strategies are required. A multi-grade representative sampling strategy, which designs a small number of representative samples with different representative grades to depict soil spatial variations at different scales, could be a potentially efficient sampling strategy for regional soil mapping. Additionally, a suitable soil mapping approach is needed to map regional soil variations based on a small number of samples. In this study, the multi-grade representative sampling strategy was applied and a fuzzy membership-weighted soil mapping approach was developed to map soil sand percentage and soil organic carbon (SOC) at 0–20 and 20–40 cm depths in a study area of 5 900 km 2 in Anhui Province of China. First, geographical sub-areas were delineated using a parent lithology data layer. Next, fuzzy c -means clustering was applied to two climate and four terrain variables in each stratum. The clustering results (environmental cluster chains) were used to locate representative samples. Evaluations based on an independent validation sample set showed that the addition of samples with lower representativeness generally led to a decrease of root mean square error (RMSE). The declining rates of RMSE with the addition of samples slowed down for 20–40 cm depth, but fluctuated for 0–20 cm depth. The predicted SOC maps based on the representative samples exhibited higher accuracy, especially for soil depth 20–40 cm, as compared to those based on legacy soil data. Multi-grade representative sampling could be an effective sampling strategy at a regional scale. This sampling strategy, combined with the fuzzy membership-based mapping approach, could be an optional effective framework for regional soil property mapping. A more detailed and accurate soil parent material map and the addition of environmental variables representing human activities would improve mapping accuracy.

The parental material of soils in the Qilian Mountains of northwest China is mainly aeolian loess containing CaCO3 which may remain in soils under the semiarid-arid climate. To disclose the CaCO3 characteristics change with the altitude and the terrain attributes, we surveyed 18 soil profiles in an altitude sequence from 3076 m to 4510 m in the Hulugou Watershed in the Qilian Mountains, measured CaCO3 contents of all genetic horizon samples, analyzed the densities, illuviation modes and depths of CaCO3 in the profiles, extracted values of the terrain attributes of the profiles including altitude slope, aspect, plane curvature, profile curvature and terrain wetness index (TWI) from the 90 m resolution SRTM3 DEM data on ArcGIS 9.3 platform. We found that CaCO3 weighted content of the profiles ranged from 1.30 g·kg-1 to 93.09 g·kg-1, CaCO3 density from 0.05 kg/m2 to 75.69 kg/m2, CaCO3 illuviation depth from 12 cm to 54 cm. CaCO3 illuviation modes could be divided into three types, i.e., no illuviation mode in which the profile has only A horizon or CaCO3 content -1, middle illuviation mode in which CaCO3 accumulated in a middle horizon, and down illuviation mode in which CaCO3 content increases with the depth. CaCO3 weighted content, density and illuviation depth had significant correlation with certain terrain attributes. In general, the altitude sequence is an effective way to study CaCO3 characteristics in the alpine region, and the data of terrain attributes which can influence the precipitation and its redistribution in soil are potential in predicting soil CaCO3 characteristics in the alpine region.

Plant growth contributes to mineral weathering, but this contribution remains poorly understood. Weathering rates in an aggrading forested watershed in subtropical China were studied by means of geochemical mass balance. Rainfall, dry deposition, and streamwater were monitored from March 2007 to February 2012. Samples of vegetative components, rainfall, dry deposition, streamwater, representative soils, and parent rock were collected and determined for mass balance calculation and clarifying plant-driven weathering mechanisms stoichiometrically. Ignoring biomass, weathering rates of Ca 2+ , Mg 2+ , Na + , and Si were 25.6, 10.7, 2.8, and 51.0 kg ha −1 year −1 , respectively. Taking biomass into consideration, weathering rates of Ca 2+ , Mg 2+ , and Si and the sum of weathering rates of Ca 2+ , Mg 2+ , Na + , K + , and Si were 2.6, 1.8, 1.2, and 1.5-fold higher than those ignoring biomass, respectively. This is attributed to plant-driven weathering due to the nutrient ( e.g ., Ca 2+ , Mg 2+ , and K + ) absorption by vegetation and substantial proton production during assimilation of these nutrients, with the former acting as a pump for removing weathering products and the latter being a source of weathering agents solubilizing mineral components. The same pattern of weathering, i.e ., higher rates of weathering with than without including biomass in mass balance calculation, was reported in previous studies; however, the extent to which plants drive weathering rates varied with vegetation types and climatic zones. The documented biological weathering driven by plants is expected to play a critical role in regulating nutrient cycling and material flows within the Earth's Critical Zone.

Visible near-infrared (VNIR) spectra can provide rich information on soil physical and chemical properties, which implies the possibility of using soil spectra to aid in the discrimination of soil types. Pedological soil classification systems use a selected set of soil properties to identify diagnostic horizons and features, and to build a classification key. This research explored the application of VNIR spectra to classify typical soil profiles collected in Anhui province, China. The 279 soil profiles used are classified into five orders (Cambosols, Vertosols, Argosols, Primosols and Anthrosols), six suborders and 21 groups according to Chinese Soil Taxonomy. Soil spectra were collected within 350–2500 nm and principal component analysis (PCA) was applied to reduce data dimension. These principal components were used as independent variables in multinomial logistic regression for soil classification. Topsoil spectra, subsoil spectra and their combination were compared for prediction accuracy. Accuracy achieved at the level of suborder using spectra of topsoil, subsoil and combined horizons were 76.3%, 71.3% and 70.3% respectively, while the results for the level of soil group using the topsoil horizon was 40.5%. Since topsoil spectra alone achieved a prediction accuracy of more than 75%, reflectance spectroscopy can be judged a promising tool for soil classification. Taxonomic distances between classes calculated on the basis of physio-chemical properties and spectra were quite different, showing that the concept of distance between classes in feature space depends on the features chosen for evaluation. Taxonomic distances can serve as a supplement for better selection and evaluation of prediction models.

The Qilian Mountains are generally capped by a productive silty soil layer wherever the environment allows. As Holocene loess is prevalent across the Qilian Mountains, we assume that at least some of these fine sediments are derived from loess deposition and that soils across the region may be genetically linked. To test these hypotheses, nine pedons in different landscapes within a typical alpine watershed of the Qilian Mountains were sampled. We also collected fine earth samples from glacier surfaces and crevices of glacial debris in the moraine zone. The particle-size distribution (PSD) is used as a proxy for identifying aeolian fractions in soils. The PSD curves of all of the fine-earth fractions examined are polymodal, although three modal sizes of roughly 16 μm, 35 μm and 80 μm are found in almost all sediments in varying proportions. These modal sizes have been identified as three sources of aeolian sediments in different transport systems. The composition and thickness of loess sediments are related to local site conditions and may be related to the paleoenvironmental history of the site. Clastic debris and vegetation cover serve as dust traps during different stages of soil formation in this alpine environment, and soils grow upward with accumulating loess sediment. Our study demonstrates that PSD curves can be used to determine the origins (and especially aeolian origins) of soils. Furthermore, this study provides insight into the pedology, ecology and paleoenvironmental history of loess-affected ecosystems in alpine areas of the Qinghai-Tibetan Plateau.

Due to the almost homogeneous topography in low relief areas, it is usually difficult to make accurate predictions of soil properties using topographic covariates. In this study, we examined how time series of field soil moisture observations can be used to estimate soil texture in an oasis agricultural area with low relief in the semi-arid region of northwest China. Time series of field-observed soil moisture variations were recorded for 132 h beginning at the end of an irrigation event during which the surface soil was saturated. Spatial correlation between two time-adjacent soil moisture conditions was used to select the factors for fuzzy c-means clustering. In each of the ten generated clusters, soil texture of the soil sample with the maximum fuzzy membership value was taken as the cluster centroid. Finally, a linearly weighted average was used to predict soil texture from the centroids. The results showed that soil moisture increased with the increase of clay and silt contents, but decreased with the increase of sand content. The spatial patterns of soil moisture changed during the entire soil drying phase. We assumed that these changes were mainly caused by spatial heterogeneity of soil texture. A total of 64 independent samples were used to evaluate the prediction accuracy. The root mean square error (RMSE) values of clay, silt and sand were 1.63, 2.81 and 3.71, respectively. The mean relative error (RE) values were 9.57% for clay, 3.77% for silt and 12.83% for sand. It could be concluded that the method used in this study was effective for soil texture mapping in the low-relief oasis agricultural area and could be applicable in other similar irrigation agricultural areas.

Water storage and flow in shallow subsurface drives runoff generation, vegetation water use and nutrient cycling. Modelling these processes under non-steady state conditions is challenging, particularly in regions like the subtropics that experience extreme wet and dry periods. At the catchment-scale, physically-based equations (e.g., Richards equation) are impractical due to their complexity, while conceptual models typically rely on steady state assumptions not found in daily hydrological dynamics. We addressed this by developing a simple modelling framework for shallow subsurface water dynamics based on physical relationships and a proxy parameter for the fluxes induced by non-unit hydraulic gradients. We demonstrate its applicability for six generic soil textures and for an Acrisol in subtropical China. Results showed that our new approach represents top soil daily fluxes and storage better than, and as fast as, standard conceptual approaches. Moreover, it was less complex and up to two orders of magnitude faster than simulating Richards equation, making it easy to include in existing hydrological models.

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.

The growing human population and demand for food have significantly impacted soil resources. Understanding the spatiotemporal change of soil conditions is important to support food production, environmental sustainability, and climate change adaptation. Nevertheless, spatiotemporal prediction of soil properties could be seriously influenced by the uncertainties of the data and model. Integrated Nested Laplace Approximation (INLA) with the Stochastic Partial Differential Equation (SPDE) was proposed as a general model that can account for the uncertainties in spatiotemporal soil modelling and prediction. INLA-SPDE has significant advantages in computation efficiency over commonly-used geostatistical methods with Markov Chain Monte Carlo. However, until now, only few pedometrics studies used it for soil spatial modelling. This study demonstrates an application of INLA-SPDE within a hierarchical spatiotemporal model for soil organic matter based on soil survey data collected in Jiangsu, China, during three periods, i.e., 1979–1982, 2000 and 2006–2007. Compared with updating digital soil maps using the Bayesian Maximum Entropy approach, the prediction generated using INLA-SPDE is more accurate. For example, the root mean square error using INLA-SPDE (i.e., 6.57 g kg−1) was reduced by 20% compared to the updating approach (i.e., 8.39 g kg−1). Moreover, accounting for sources of uncertainties made the prediction using INLA-SPDE more certain. Nevertheless, the uncertainty in the temporal prediction of soil change is still large due to the scarcity of data across the sampling periods. The INLA-SPDE model predicts much detailed spatiotemporal changes along the sampling periods. Therefore, this study recommends the use of INLA-SPDE within a hierarchical model as an effective method for studying spatiotemporal soil change.

The stability of soil colloids influences soil physicochemical properties, soil development, and transfer of nutrients and contaminants to surface and ground waters. A better understanding of soil colloids stability dynamics during soil evolution is important for the evaluation of soil's capacity to retain nutrients and/or accommodate toxic contaminants. This study was aimed to determine changes in the stability of water-dispersible soil colloids that accompany mineral transformation and surface charge evolution during pedogenesis using a well characterized chronosequence derived from basalt in the humid tropical region of Hainan Island, South China. The results demonstrated that the pH-dependent colloid stability decreased significantly with tropical soil development, which we attribute to the substantial changes in clay mineral compositions and colloid surface charge properties. Clay minerals in the studied chronosequence were characterized by an increase of kaolinite, gibbsite and Fe oxides and a decrease of quartz and halloysite towards more advanced stages of weathering, which resulted in the decline of permanent negative charges in the older soils. The point of zero charge (pHPZC) increased while ∆ pH decreased across the tropical soil chronosequence, being in good agreement with the observed lower colloid stability in aged soils dominated by kaolinitic minerals. Our study of colloid stability at long-term pedogenic time scale suggests young tropical soils (

Atmospheric deposition, a major pathway of metals entering into soils, plays an important role in soil environment, especially in urban regions where a large amount of pollutants are emitted into atmosphere through various sources. In order to understand the characteristics of atmospheric deposition in urban area and its relation with natural and anthropogenic sources, a three-year study of atmospheric deposition at three typical sites, industrial zone (IN), urban residential area (RZ) and suburban forested scenic area (FA), was carried out in Nanjing, a metropolitan city in eastern China from 2005 to 2007. The bulk deposition rate and element composition of atmospheric deposition varied spatio-temporally in the urban zones of Nanjing. The concentrations of Cu, Zn, Pb and Ca in the atmospheric deposits were strongly enriched in the whole Nanjing region; however, anthropogenic pollutants in atmospheric deposits were diluted by the input of external mineral dust transported from northwestern China. Source apportionment through principal component analysis (PCA) showed that the background atmospheric deposition at the FA site was the combination of external aerosol and local emission sources. The input of long-range transported Asian dust had an important influence on the urban background deposition, especially in spring when the continental dust from the northwestern China prevailed. Marine aerosol source was observed in summer and autumn, the seasons dominated by summer monsoon in Nanjing. In contrast, the contribution of local anthropogenic emission source was constant regardless of seasons. At the RZ and IN sites, the atmospheric deposition was more significantly affected by the nearby human activities than at the FA site. In addition, different urban activities and both the winter and summer Asian monsoons had substantial impacts on the characteristics of dust deposition in urban Nanjing.

Conventional soil maps generally contain one or more soil types within a single soil polygon. But their geographic locations within the polygon are not specified. This restricts current applications of the maps in site-specific agricultural management and environmental modelling. We examined the utility of legacy pedon data for disaggregating soil polygons and the effectiveness of similarity-based prediction for making use of the under- or over-sampled legacy pedon data for the disaggregation. The method consisted of three steps. First, environmental similarities between the pedon sites and each location were computed based on soil formative environmental factors. Second, according to soil types of the pedon sites, the similarities were aggregated to derive similarity distribution for each soil type. Third, a hardening process was performed on the maps to allocate candidate soil types within the polygons. The study was conducted at the soil subgroup level in a semi-arid area situated in Manitoba, Canada. Based on 186 independent pedon sites, the evaluation of the disaggregated map of soil subgroups showed an overall accuracy of 67% and a Kappa statistic of 0.62. The map represented a better spatial pattern of soil subgroups in both detail and accuracy compared to a dominant soil subgroup map, which was commonly used in practice. Incorrect predictions mainly occurred in the agricultural plain area and the soil subgroups that are very similar in taxonomy, indicating that new environmental covariates need to be developed. We concluded that the combination of legacy pedon data with similarity-based prediction is an effective solution for soil polygon disaggregation.

The accuracy and precision of predictions of soil texture based on visible–near-infrared (Vis-NIR) spectra are affected by the soil moisture content at the time of measurement. This study aimed to quantify these effects. We also developed a method to improve the accuracy of soil texture prediction when the difference in moisture content is large and unknown, as is usually the case in field measurements. Reflection spectra (380–2400 nm) of 89 soil samples were obtained in the laboratory under nine moisture conditions. Prediction models for each moisture condition were built separately using partial least-squares (PLS) regression. Each model was applied to independent validation sets under the same moisture condition as the calibration sets, and then the model based on air-dried soils (dry model) was applied to the other eight moisture conditions to quantify the effect of soil moisture on the prediction accuracy. Finally, the perpendicular drought index (PDI) was used as an indicator of soil moisture, and the nine moisture conditions were regrouped to four groups using the PDI. Prediction models for each group with the same PDI were built and evaluated. The results show that Vis-NIR spectra can be directly applied to predict the soil texture of soils in different known moisture states. When the moisture state is unknown, the models based on PDI grouping markedly improve prediction accuracy. The RMSEₚ of prediction from PDI-grouped models of prediction of the content of clay and sand separates were between 1 and 2% and between 8 and11%, respectively, for the different moisture classes. This method has potential for direct application in the field with on-the-go sensors.

This paper presents an approach to predicting three-dimensional (3D) variation of soil organic matter (SOM) concentration by integrating a similarity-based method with depth functions. It was tested in a small hilly landscape. A depth function model was constructed to fit SOM profile distribution using a linear relation in the topsoil and a power function in the subsoil. Then, under the assumption that similar environmental conditions at two sites would lead to the development of similar profile morphologies and thus similar depth function parameters, the similarity-based method was used to spatially interpolate the depth function parameters based on their relationships with environmental variables. With the values of the parameters for every location, a 3D map of SOM distribution was generated. The predicted SOM pattern well reproduced the statistical distribution of the pedon dataset used in this study. The overall mean error (ME) was 0.06 g kg− 1 and ratio of performance to deviation (RPD) was 2.34. We conclude that the proposed approach is effective and accurate for 3D SOM prediction. It overcomes two drawbacks of the frequently used pseudo 3D soil mapping approach: (1) the neglect of vertical soil pattern when performing horizontal soil predictions, and (2) the repeated applications of depth function fittings in the mapping process, both of which may lead to prediction errors. Moreover, the similarity-based method is a transparent and traceable prediction process, allowing for easy interpretation of its results. This is useful for understanding soil–environmental relationships and processes. The method thus is an attractive alternative to the commonly used non-linear “black-box” techniques such as artificial neural networks.

Soil organic carbon (SOC) plays an important role in soil fertility and carbon sequestration, and a better understanding of the spatial patterns of SOC is essential for soil resource management. In this study, we used boosted regression tree (BRT) and random forest (RF) models to map the distribution of topsoil organic carbon content at the northeastern edge of the Tibetan Plateau in China. A set of 105 soil samples and 12 environmental variables (including topography, climate and vegetation) were analyzed. The performance of the models was evaluated using a 10-fold cross-validation procedure. Maps of the mean values and standard deviations of SOC were generated to illustrate model variability and uncertainty. The results indicate that the BRT and RF models exhibited very similar performance and yielded similar predicted distributions of SOC. The two models explained approximately 70% of the total SOC variability. The BRT and RF models robustly predicted the SOC at low observed SOC values, whereas they underestimated high observed SOC values. This underestimation may have been caused by biased distributions of soil samples in the SOC space. Vegetation-related variables were assigned the highest importance in both models, followed by climate and topography. Both models produced spatial distribution maps of SOC that were closely related to vegetation cover. The SOC content predicted by the BRT model was clearly higher than that of the RF model in areas with greater vegetation cover because the contributions of vegetation-related variables in the two models (65% and 43%, respectively) differed significantly. The predicted SOC content increased from the northwestern to the southeastern part of the study area, average values produced by the BRT and RF models were 27.3 g kg−1 and 26.6 g kg−1, respectively. We conclude that the BRT and RF methods should be calibrated and compared to obtain the best prediction of SOC spatial distribution in similar regions. In addition, vegetation variables, including those obtained from remote sensing imagery, should be taken as the main environmental indicators and explicitly included when generating SOC maps in Alpine environments.

Although Hainan Island is located in one of the relative least dusty regions in China, Asian dust has been identified in Hainan soils. We used Sr and Nd isotopes of soils, saprolite, bedrocks and Asian dust to not only identify long-range transported dust but also quantify the dust contribution to Hainan soils. Sr and Nd isotopic compositions of the soils are clearly affected by dust deposition. The soils in near-surface horizons have 87Sr/86Sr values as high as 0.7193 that are more radiogenic than the basalts, indicating involvement of Asian dust; whereas deep horizons show a dominantly basaltic signature (87Sr/86Sr = 0.7038). The eNd(0) values for Hainan soils range from − 6.46 to 1.85. The higher value closely approximates the values measured for the Hainan lava, and the lower value indicates the fingerprint of the Asian dust. The 87Sr/86Sr and eNd(0) values for soils also show a progressive trend with time, which implies that dust accumulates with soil age. We calculated the dust accretion rates for Sr-based and Nd-based to quantify the dust contributions to soils. The estimate of dust accreted shows an approximate increasing trend with the ages of soils for Sr-based and Nd-based methods. The soils in the old sites are likely to have undergone long-term dust input and accumulated high dust levels. We also find excellent agreement in the amount of dust present and dust accretion rates based on the two tracers at the 300 ka site and the 1120 ka site, respectively. Besides lower dust deposition in the Pliocene epoch than the late Quaternary over eastern Eurasian, the loss of mineralogical and geochemical tracers over million year time scales also may lead to underestimates of time-averaged dust deposition rates in the old sites. Therefore, the young sites (180 ka and 300 ka) should provide more realistic estimates of average dust accretion rates. The importance of dust in the tropical Hainan soil chronosequence highlights the significance of dust accretion to soils globally.

Numerous pedotransfer functions (PTFs) have been developed to predict the soil properties of interest from other soil properties and, less commonly, from environmental variables. However, only a few PTFs have been developed to predict soil nutrients using environmental variables and to extrapolate them to characterize spatial soil variations at a regional scale. In this study, we attempted to develop PTFs for the total nitrogen (TN), total phosphorus (TP) and total potassium (TK) concentrations in three typical pedo-climatic areas of China (Fujian Province, Jiangsu Province and Qilian Mountains) with diverse climate, terrain and soil types. A series of linear PTFs were developed to quantify the effect of terrain and climate on the predictive relations between the soil nutrients and other measured soil properties and environmental variables. In addition, digital soil mapping (DSM) based on the random forest (RF) technique was performed to test the hypothesis that the best-fit PTFs could be extrapolated, based on soil maps and environmental variables, to describe regional soil variations in the soil nutrients. The root mean square errors (RMSEs) of the best-fit PTFs for TN, TP and TK ranged from 0.21 to 0.79 g kg−1, 0.20 to 0.58 g kg−1, and 3.68 to 5.00 g kg−1, respectively. Different RMSEs were produced by DSM, namely 0.37-1.89 g kg−1, 0.19−0.56 g kg−1 and 3.79-4.83 g kg−1 for TN, TP and TK, respectively. PTFs provided a sound basis for database compilation if the soil properties were highly correlated. However, the extrapolation of best-fit PTFs to regional scales yielded greater errors than those produced by DSM. The comparison results reveal the limitations of PTFs and suggest that their performance could be improved by using environmental covariates or by fitting data in areas with relatively homogeneous soil landscapes. The DSM techniques may provide satisfactory alternatives to predict soil data at both regional and plot scales.

There is evidence that high levels of nitrate are stored in Earth’s Critical Zone between the land surface and impermeable bedrock. While nitrate in the upper layers of the soil ( 1 m). To study controls on nitrate accumulation in a subtropical monsoon area, a total of 728 regolith samples, taken from land surface to fresh bedrock, were collected from 21 drillings in a typical red soil Critical Zone in southern China. We characterized the variation of nitrate in these deep horizons and investigated the effect of land use, slope position and regolith physiochemical properties. The results showed that the amount of nitrate at 240-380 cm depth was more stable than at 100-240 cm or 380-500 cm depth, and accumulated at that depth in the upland and orchard regoliths. Nitrate contents in the deep horizons of upland and orchard regoliths were significantly larger than that in the paddy fields or woodland (p

Ground-penetrating radar (GPR) is a convenient tool for volumetric soil water content (VSWC) estimation in hydrological and agricultural studies. Although case studies have been widely carried out, little attention has been paid to subsoil moisture estimates. In this research, we investigated three-dimensional soil moisture variation down to a depth of 1 m and the effect of rainfall events on spatial soil moisture dynamics. GPR surveying lines were conducted both before and after a heavy rainfall event to map the VSWC. Soil sampling and time domain reflectometry (TDR) probe data at different depths (20, 40, 60, 80, and 100 cm) were acquired. Our results demonstrated that there was a significant correlation between the dielectric constants and VSWCs at all depths. The established relationships for the different depth ranges had a low VSWC discrepancy when the dielectric constants ranged from 10 to 15. The effective range of each variogram was larger than 20 m, except for that of the 0–100 cm VSWC map after rainfall. In addition, the validation diagrams using corrected TDR values demonstrated relatively reliable VSWC maps. Approximately 89% of the variation in VSWC could be explained by the dielectric constants in the depth range of 0–40 cm, and VSWC predictions at this soil depth outperformed those at other depth ranges, with an overall RMSE of 0.027 m3 m−3 and R2 of 0.725. Furthermore, we also monitored the effect of precipitation on the accuracy of the VSWC prediction on shallow surfaces. Our study shows that three-dimensional soil moisture dynamics can be accurately estimated at the field scale by integrating GPR interpretation and spatial extrapolation methods.