Showing papers on "Digital soil mapping published in 2017"

SoilGrids250m: Global gridded soil information based on machine learning

Abstract: This paper describes the technical development and accuracy assessment of the most recent and improved version of the SoilGrids system at 250m resolution (June 2016 update). SoilGrids provides global predictions for standard numeric soil properties (organic carbon, bulk density, Cation Exchange Capacity (CEC), pH, soil texture fractions and coarse fragments) at seven standard depths (0, 5, 15, 30, 60, 100 and 200 cm), in addition to predictions of depth to bedrock and distribution of soil classes based on the World Reference Base (WRB) and USDA classification systems (ca. 280 raster layers in total). Predictions were based on ca. 150,000 soil profiles used for training and a stack of 158 remote sensing-based soil covariates (primarily derived from MODIS land products, SRTM DEM derivatives, climatic images and global landform and lithology maps), which were used to fit an ensemble of machine learning methods-random forest and gradient boosting and/or multinomial logistic regression-as implemented in the R packages ranger, xgboost, nnet and caret. The results of 10-fold cross-validation show that the ensemble models explain between 56% (coarse fragments) and 83% (pH) of variation with an overall average of 61%. Improvements in the relative accuracy considering the amount of variation explained, in comparison to the previous version of SoilGrids at 1 km spatial resolution, range from 60 to 230%. Improvements can be attributed to: (1) the use of machine learning instead of linear regression, (2) to considerable investments in preparing finer resolution covariate layers and (3) to insertion of additional soil profiles. Further development of SoilGrids could include refinement of methods to incorporate input uncertainties and derivation of posterior probability distributions (per pixel), and further automation of spatial modeling so that soil maps can be generated for potentially hundreds of soil variables. Another area of future research is the development of methods for multiscale merging of SoilGrids predictions with local and/or national gridded soil products (e.g. up to 50 m spatial resolution) so that increasingly more accurate, complete and consistent global soil information can be produced. SoilGrids are available under the Open Data Base License.

High Resolution Mapping of Soil Properties Using Remote Sensing Variables in South-Western Burkina Faso: A Comparison of Machine Learning and Multiple Linear Regression Models.

Abstract: Accurate and detailed spatial soil information is essential for environmental modelling, risk assessment and decision making. The use of Remote Sensing data as secondary sources of information in digital soil mapping has been found to be cost effective and less time consuming compared to traditional soil mapping approaches. But the potentials of Remote Sensing data in improving knowledge of local scale soil information in West Africa have not been fully explored. This study investigated the use of high spatial resolution satellite data (RapidEye and Landsat), terrain/climatic data and laboratory analysed soil samples to map the spatial distribution of six soil properties–sand, silt, clay, cation exchange capacity (CEC), soil organic carbon (SOC) and nitrogen–in a 580 km2 agricultural watershed in south-western Burkina Faso. Four statistical prediction models–multiple linear regression (MLR), random forest regression (RFR), support vector machine (SVM), stochastic gradient boosting (SGB)–were tested and compared. Internal validation was conducted by cross validation while the predictions were validated against an independent set of soil samples considering the modelling area and an extrapolation area. Model performance statistics revealed that the machine learning techniques performed marginally better than the MLR, with the RFR providing in most cases the highest accuracy. The inability of MLR to handle non-linear relationships between dependent and independent variables was found to be a limitation in accurately predicting soil properties at unsampled locations. Satellite data acquired during ploughing or early crop development stages (e.g. May, June) were found to be the most important spectral predictors while elevation, temperature and precipitation came up as prominent terrain/climatic variables in predicting soil properties. The results further showed that shortwave infrared and near infrared channels of Landsat8 as well as soil specific indices of redness, coloration and saturation were prominent predictors in digital soil mapping. Considering the increased availability of freely available Remote Sensing data (e.g. Landsat, SRTM, Sentinels), soil information at local and regional scales in data poor regions such as West Africa can be improved with relatively little financial and human resources.

Soil Moisture Remote Sensing: State-of-the-Science

Abstract: This is an update to the special section “Remote Sensing for Vadose Zone Hydrology—A Synthesis from the Vantage Point” [Vadose Zone Journal 12(3)]. Satellites (e.g., Soil Moisture Active Passive [SMAP] and Soil Moisture and Ocean Salinity [SMOS]) using passive microwave techniques, in particular at L-band frequency, have shown good promise for global mapping of near-surface (0–5-cm) soil moisture at a spatial resolution of 25 to 40 km and temporal resolution of 2 to 3 d. C- and X-band soil moisture records date back to 1978, making available an invaluable data set for long-term climate research. Near-surface soil moisture is further extended to the root zone (top 1 m) using process-based models and data assimilation schemes. Validation of remotely sensed soil moisture products has been ongoing using core monitoring sites, sparse monitoring networks, intensive field campaigns, as well as multi-satellite comparison studies. To transfer empirical observations across space and time scales and to develop improved retrieval algorithms at various resolutions, several efforts are underway to associate soil moisture variability dynamics with land surface attributes in various energy- and water-rich environments. We describe the most recent scientific and technological advances in soil moisture remote sensing. We anticipate that remotely sensed soil moisture will find many applications in vadose zone hydrology in the coming decades.

Evaluation of digital soil mapping approaches with large sets of environmental covariates

Abstract: . The spatial assessment of soil functions requires maps of basic soil properties. Unfortunately, these are either missing for many regions or are not available at the desired spatial resolution or down to the required soil depth. The field-based generation of large soil datasets and conventional soil maps remains costly. Meanwhile, legacy soil data and comprehensive sets of spatial environmental data are available for many regions. Digital soil mapping (DSM) approaches relating soil data (responses) to environmental data (covariates) face the challenge of building statistical models from large sets of covariates originating, for example, from airborne imaging spectroscopy or multi-scale terrain analysis. We evaluated six approaches for DSM in three study regions in Switzerland (Berne, Greifensee, ZH forest) by mapping the effective soil depth available to plants (SD), pH, soil organic matter (SOM), effective cation exchange capacity (ECEC), clay, silt, gravel content and fine fraction bulk density for four soil depths (totalling 48 responses). Models were built from 300–500 environmental covariates by selecting linear models through (1) grouped lasso and (2) an ad hoc stepwise procedure for robust external-drift kriging (georob). For (3) geoadditive models we selected penalized smoothing spline terms by component-wise gradient boosting (geoGAM). We further used two tree-based methods: (4) boosted regression trees (BRTs) and (5) random forest (RF). Lastly, we computed (6) weighted model averages (MAs) from the predictions obtained from methods 1–5. Lasso, georob and geoGAM successfully selected strongly reduced sets of covariates (subsets of 3–6 % of all covariates). Differences in predictive performance, tested on independent validation data, were mostly small and did not reveal a single best method for 48 responses. Nevertheless, RF was often the best among methods 1–5 (28 of 48 responses), but was outcompeted by MA for 14 of these 28 responses. RF tended to over-fit the data. The performance of BRT was slightly worse than RF. GeoGAM performed poorly on some responses and was the best only for 7 of 48 responses. The prediction accuracy of lasso was intermediate. All models generally had small bias. Only the computationally very efficient lasso had slightly larger bias because it tended to under-fit the data. Summarizing, although differences were small, the frequencies of the best and worst performance clearly favoured RF if a single method is applied and MA if multiple prediction models can be developed.

Barest Pixel Composite for agricultural areas using Landsat time series

Abstract: Many soil remote sensing applications rely on narrow-band observations to exploit molecular absorption features. However, broadband sensors are invaluable for soil surveying, agriculture, land management and mineral exploration, amongst others. These sensors provide denser time series compared to high-resolution airborne imaging spectrometers and hold the potential of increasing the observable bare-soil area at the cost of spectral detail. The wealth of data coming along with these applications can be handled using cloud-based processing platforms such as Earth Engine. We present a method for identifying the least-vegetated observation, or so called barest pixel, in a dense time series between January 1985 and March 2017, based on Landsat 5, 7 and 8 observations. We derived a Barest Pixel Composite and Bare Soil Composite for the agricultural area of the Swiss Plateau. We analysed the available data over time and concluded that about five years of Landsat data were needed for a full-coverage composite (90% of the maximum bare soil area). Using the Swiss harmonised soil data, we derived soil properties (sand, silt, clay, and soil organic matter percentages) and discuss the contribution of these soil property maps to existing conventional and digital soil maps. Both products demonstrate the substantial potential of Landsat time series for digital soil mapping, as well as for land management applications and policy making.

Exploring the spatial variability of soil properties in an Alfisol soil catena

Abstract: Detailed digital soil maps showing the spatial heterogeneity of soil properties consistent with the landscape are required for site-specific management of plant nutrients, land use planning and process-based environmental modeling. We characterized the short-scale spatial heterogeneity of soil properties in an Alfisol catena in a tropical landscape of Sri Lanka. The impact of different land-uses (paddy, vegetable and un-cultivated) was examined to assess the impact of anthropogenic activities on the variability of soil properties at the catenary level. Conditioned Latin hypercube sampling was used to collect 58 geo-referenced topsoil samples (0–30 cm) from the study area. Soil samples were analyzed for pH, electrical conductivity (EC), organic carbon (OC), cation exchange capacity (CEC) and texture. The spatial correlation between soil properties was analyzed by computing cross-variograms and subsequent fitting of theoretical model. Spatial distribution maps were developed using ordinary kriging. The range of soil properties, pH: 4.3–7.9; EC: 0.01–0.18 dS m− 1; OC: 0.1–1.37%; CEC: 0.44–11.51 cmol (+) kg− 1; clay: 1.5–25% and sand: 59.1–84.4% and their coefficient of variations indicated a large variability in the study area. Electrical conductivity and pH showed a strong spatial correlation which was reflected by the cross-variogram close to the hull of the perfect correlation. Moreover, cross-variograms calculated for EC and Clay, CEC and OC, CEC and clay and CEC and pH indicated weak positive spatial correlation between these properties. Relative nugget effect (RNE) calculated from variograms showed strongly structured spatial variability for pH, EC and sand content (RNE < 25%) while CEC, organic carbon and clay content showed moderately structured spatial variability (25% < RNE < 75%). Spatial dependencies for examined soil properties ranged from 48 to 984 m. The mixed effects model fitting followed by Tukey's post-hoc test showed significant effect of land use on the spatial variability of EC. Our study revealed a structured variability of topsoil properties in the selected tropical Alfisol catena. Except for EC, observed variability was not modified by the land uses. Investigated soil properties showed distinct spatial structures at different scales and magnitudes of strength. Our results will be useful for digital soil mapping, site specific management of soil properties, developing appropriate land use plans and quantifying anthropogenic impacts on the soil system.

The Application of EM38: Determination of Soil Parameters, Selection of Soil Sampling Points and Use in Agriculture and Archaeology.

Abstract: Fast and accurate assessment of within-field variation is essential for detecting field-wide heterogeneity and contributing to improvements in the management of agricultural lands. The goal of this paper is to provide an overview of field scale characterization by electromagnetic induction, firstly with a focus on the applications of EM38 to salinity, soil texture, water content and soil water turnover, soil types and boundaries, nutrients and N-turnover and soil sampling designs. Furthermore, results concerning special applications in agriculture, horticulture and archaeology are included. In addition to these investigations, this survey also presents a wide range of practical methods for use. Secondly, the effectiveness of conductivity readings for a specific target in a specific locality is determined by the intensity at which soil factors influence these values in relationship to the desired information. The interpretation and utility of apparent electrical conductivity (ECa) readings are highly location- and soil-specific, so soil properties influencing the measurement of ECa must be clearly understood. From the various calibration results, it appears that regression constants for the relationships between ECa, electrical conductivity of aqueous soil extracts (ECe), texture, yield, etc., are not necessarily transferable from one region to another. The modelling of ECa, soil properties, climate and yield are important for identifying the location to which specific utilizations of ECa technology (e.g., ECa−texture relationships) can be appropriately applied. In general, the determination of absolute levels of ECa is frequently not possible, but it appears to be quite a robust method to detect relative differences, both spatially and temporally. Often, the use of ECa is restricted to its application as a covariate or the use of the readings in a relative sense rather than as absolute terms.

Multiple linear regression and random forest to predict and map soil properties using data from portable X-ray fluorescence spectrometer (pXRF).

Abstract: Determination of soil properties ​​helps in the correct management of soil fertility. The portable X-ray fluorescence spectrometer (pXRF) has been recently adopted to determine total chemical element contents in soils, allowing soil property inferences. However, these studies are still scarce in Brazil and other countries. The objectives of this work were to predict soil properties using pXRF data, comparing stepwise multiple linear regression (SMLR) and random forest (RF) methods, as well as mapping and validating soil properties. 120 soil samples were collected at three depths and submitted to laboratory analyses. pXRF was used in the samples and total element contents were determined. From pXRF data, SMLR and RF were used to predict soil laboratory results, reflecting soil properties, and the models were validated. The best method was used to spatialize soil properties. Using SMLR, models had high values of R² (≥0.8), however the highest accuracy was obtained in RF modeling. Exchangeable Ca, Al, Mg, potential and effective cation exchange capacity, soil organic matter, pH, and base saturation had adequate adjustment and accurate predictions with RF. Eight out of the 10 soil properties predicted by RF using pXRF data had CaO as the most important variable helping predictions, followed by P2O5, Zn and Cr. Maps generated using RF from pXRF data had high accuracy for six soil properties, reaching R2 up to 0.83. pXRF in association with RF can be used to predict soil properties with high accuracy at low cost and time, besides providing variables aiding digital soil mapping.

S‐World: A Global Soil Map for Environmental Modelling

Abstract: The research community increasingly analyses global environmental problems like climate change and desertification with models These global environmental modelling studies require global, high resolution, spatially exhaustive, and quantitative data describing the soil profile This study aimed to develop a pedological approach that takes stock of available legacy and auxiliary data to create a global, 30 arc second soil property database for modelling The methodology uses the Harmonized World Soil Database and the ISRIC-WISE 31 soil profile database Auxiliary information at 30 arc second resolution for various landscape properties is used to describe the variation in landscape properties (temperature and precipitation, topography, elevation, land use, and land cover) Complex mapping units of the HWSD were first disaggregated using a digital elevation model and toposequences to generate delineated areas described by a single soil type Secondly, ranges of soil properties per soil type were determined using the soil profile data Then a meta-analysis on a broad literature survey was used to develop a simple model that, based on landscape properties at a particular location, determines the position within these ranges and thus provides an estimation of the local soil properties Finally, the model was implemented at the global level to determine the distribution of soil properties The methodology, denominated S-World (Soils of the world) resulted in readily available, high resolution, global soil property maps that are now available for environmental modelling

Spatial soil nutrients prediction using three supervised learning methods for assessment of land potentials in complex terrain

Abstract: Mountain soils play an essential role in ecosystem management. Assessment of land potentials can provide detailed spatial information particularly concerning nutrient availability. Spatial distributions of topsoil carbon, nitrogen and available phosphorus in mountain regions were identified using supervised learning methods, and a functional landscape analysis was performed in order to determine the spatial soil fertility pattern for the Soyang Lake watershed in South Korea. Specific research aims were (1) to identify important predictors; (2) to develop digital soil maps; (3) to assess land potentials using digital soil maps. Soil profiles and samples were collected by conditioned Latin Hypercube Sampling considering operational field constraints such as accessibility and no-go areas contaminated by landmines as well as budget limitations. Terrain parameters and different vegetation indices were derived for the covariates. We compared a generalized additive model (GAM) to random forest (RF) and support vector regression (SVR). For the predictor selection, we used the recursive feature elimination (RFE). A land potential assessment for soil nutrients was conducted using trimmed k-mean cluster analysis. Results suggested that vegetation indices have powerful abilities to predict soil nutrients. Using selected predictors via RFE improved prediction results. RF showed the best performance. Cluster analysis identified four land potential classes: fertile, medium and low fertile with an additional class dominated by high phosphorus and low carbon and nitrogen contents due to human impact. This study provides an effective approach to map land potentials for mountain ecosystem management.

Short-range and regional spatial variability of soil chemical properties in an agro-ecosystem in eastern Croatia

Abstract: Spatial and temporal characterization of soil properties in agro-ecosystems is crucial for monitoring the evolution of soil functions and for understanding the main influential processes. Moreover, the objective mapping of soil properties in agro-ecosystems is urgently needed for regional planning purposes and the proper choice of land management practices. In this work, the geostatistical analysis of a dataset of soil properties, derived from 2411 soil samples collected in Vukovar-Srijem County (Croatia), highlighted the multiple benefits of a spatial-statistical approach. The main aim of this paper is to jointly examine short-range (i.e., within-field) and regional spatial variability of several soil chemical properties: soil pH, organic matter (OM), plant available phosphorus (AP) and potassium (AK). The available sampling network, characterized by a set of 2411 (0–30 cm depth) irregularly and field-clustered soil samples, allowed to derivate of two typologies of soil nutrient maps by means of ordinary block kriging: within-field high-resolution maps (block size 250 m) and regional low-resolution maps (block size 2000 m). Soil pH and OM had lower variability compared to AP and AK. The OM content and pH ranged from 1.24% to 5.25% and from 3.69 to 7.84, respectively. Almost 94% of all samples had an OM content below 3%, indicating the need for future adoption of environmentally friendly soil management in this county. The mean values of AP and AK were 173 mg kg − 1 and 238 mg kg − 1 , respectively, indicating a moderate supply of these nutrients. Geostatistical analysis revealed that the best-fit models were spherical for pH and AP, with moderate spatial dependency, and exponential for OM and AK, with strong spatial dependency. The within-field high-resolution soil property maps can be used as guidance for site-specific fertilization and liming. In addition, the regional maps derived for larger interpolation support provide quantitative information for regional planning and environmental monitoring and protection purposes. Consequently, the multi-resolution mapping of soil properties and the analysis of their spatial variability highlighted possible connections with influential factors and processes, including the relationships with different soil types. Finally, quantification of the spatial variability of soil properties by means of variogram models constitutes a basis for optimizing soil sample spacing for mapping purposes in the studied region.

Assessment of spatial variability and mapping of soil properties for sustainable agricultural production using geographic information system techniques (GIS)

Abstract: Mapping of soil properties is an important operation as it plays an important role in the knowledge about soil properties and how it can be used sustainably. The study was carried out in a local government area in Oyo state in order to map out some soil characteristics and assess their variability within the area. Soil sampling was carried out in three different locations in the local government area using the cluster sampling technique. Ten samples were collected in each location within a 10 by 10 km area, soil was sampled at two depths (0–20 and 20–40 cm) respectively. The soil samples were air-dried, crushed and passed through a 2 mm sieve before analyzing it for Nitrogen, phosphorus, Potassium, Organic carbon, pH and exchangeable bases in the laboratory while the SAR, ESP% and cation exchange capacity (CEC) were calculated. After the normalization of data classical statistics was used to describe the soil properties and geo-statistical analysis was used to illustrate the spatial variability of...