N.H. Batjes

SoilGrids1km — Global Soil Information Based on Automated Mapping

Background Soils are widely recognized as a non-renewable natural resource and as biophysical carbon sinks. As such, there is a growing requirement for global soil information. Although several global soil information systems already exist, these tend to suffer from inconsistencies and limited spatial detail. Methodology/Principal Findings We present SoilGrids1km — a global 3D soil information system at 1 km resolution — containing spatial predictions for a selection of soil properties (at six standard depths): soil organic carbon (g kg−1), soil pH, sand, silt and clay fractions (%), bulk density (kg m−3), cation-exchange capacity (cmol+/kg), coarse fragments (%), soil organic carbon stock (t ha−1), depth to bedrock (cm), World Reference Base soil groups, and USDA Soil Taxonomy suborders. Our predictions are based on global spatial prediction models which we fitted, per soil variable, using a compilation of major international soil profile databases (ca. 110,000 soil profiles), and a selection of ca. 75 global environmental covariates representing soil forming factors. Results of regression modeling indicate that the most useful covariates for modeling soils at the global scale are climatic and biomass indices (based on MODIS images), lithology, and taxonomic mapping units derived from conventional soil survey (Harmonized World Soil Database). Prediction accuracies assessed using 5–fold cross-validation were between 23–51%. Conclusions/Significance SoilGrids1km provide an initial set of examples of soil spatial data for input into global models at a resolution and consistency not previously available. Some of the main limitations of the current version of SoilGrids1km are: (1) weak relationships between soil properties/classes and explanatory variables due to scale mismatches, (2) difficulty to obtain covariates that capture soil forming factors, (3) low sampling density and spatial clustering of soil profile locations. However, as the SoilGrids system is highly automated and flexible, increasingly accurate predictions can be generated as new input data become available. SoilGrids1km are available for download via http://soilgrids.org under a Creative Commons Non Commercial license.

Mitigation of atmospheric CO2 concentrations by increased carbon sequestration in the soil

Abstract The International Panel on Climate Change distinguished three main options for the mitigation of atmospheric CO2 concentrations by the agricultural sector: (1) reduction of agriculture-related emissions, (2) creation and strengthening of C sinks in the soil, and (3) production of biofuels to replace fossil fuels. Options for sustained sequestration of C in the soil through adapted management of land resources are reviewed in the context of the ongoing discussion on the need to reduce greenhouse gas concentrations in the atmosphere. Enhanced sequestration of atmospheric CO2 in the soil, ultimately as stable humus, may well prove a more lasting solution than (temporarily) sequestering CO2 in the standing biomass through reforestation and afforestation. Such actions will also help to reverse processes of land degradation, thus contributing to sustained food productivity and security for the people in the regions concerned.

SoilGrids250m: Global gridded soil information based on machine learning

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.

Development of a world data set of soil water retention properties using pedotransfer rules

The World Inventory of Soil Emission Potentials (WISE) database is used to compile a standardized and spatially explicit data set of soil water retention properties. WISE holds 4353 globally distributed profiles considered to be representative of the soil units shown on a 1/2° latitude by 1/2° longitude version of the corrected and digitized 1:5 M FAO-Unesco Soil Map of the World. Pedotransfer functions (PTFs) are presented for the prediction of (a) volumetric soil-water content (θ) at 9 pre-selected soil-water potentials (h), and (b) available water capacity (AWC) from measured silt, clay and organic matter content. All regressions for θh are significant (0.88 < r2 < 0.94; P < 0.001). The predictive capability of the regression for the calculation of AWC, however, was relatively low and this PTF systematically underestimated AWC for the independent data set. An alternative approach, which uses pedotransfer rules (PTRs) and functional grouping for estimating AWC from FAO-Unesco soil unit type, horizon textural class and organic matter class, was developed and tested for its predictive capability. For the independent data set, the PTR-derived AWC values showed a better correlation (r2 = 0.80) with measured AWC values than was the case for the PTF-derived AWC values. In addition to this, the median of the relative difference between predicted and measured AWC values was smaller (−3%) than for the PTF-predicted AWC values (−13%). Nonetheless, both methods showed a fairly large scatter between the predicted and measured AWC values. The PTRs were used for the estimation of profile available AWC to a depth of 100 cm, except for shallow Lithosols, Rankers and Rendzinas. Median AWC to a depth of 1 m is 42 mm for coarse-textured Arenosols, 80 mm for strongly weathered Ferralsols, 130 mm for Vertisols, 187 mm for Andosols, and 480 mm for Histosols. Correction factors were introduced to account for effects of coarse fragments and presence of groundwater at shallow depth. The “corrected” values were used to generate a 1/2° latitude by 1/2° longitude world data set of AWC properties. This raster image file uses the general format of the Global Ecosystems Database (Kineman, 1993). The resolution of the spatial data set is considered appropriate for water balance studies in global assessments of crop production potentials, soil vulnerability to pollution, and soil gaseous emissions.

Principal land use changes anticipated in Europe

Abstract Major changes in land use may be anticipated in Europe in the decades to come as a result of technological, socio-economic and political developments as well as global environmental change. The type and effects of these changes will strongly depend on policy decisions which are governed, amongst others, by: (i) an increasing agricultural productivity; (ii) an increasing realization of the need to conserve bio-diversity and environmental quality for current and future generations; (iii) pressure from an increasingly urban population to emphasize non-agricultural forms of land use in terms of nature and landscape conservation; (iv) increasing market-driven demand for high quality produce made with environmentally friendly forms of management; and (v) increasing food demand on the world market as the world population doubles and purchasing powers increases, particularly in Asia. These developments occur now both in Eastern and Western Europe, even though historical developments during the last 50 years have been strikingly different in these two regions. Collectivization in Eastern Europe after World War II was associated with higher yields but also with unfavourable changes in land use and cropping patterns causing acidification, soil erosion, salinization and chemical pollution. The change to democracy in the late 1980s implied a change from an essentially quantity- to a more quality-oriented type of agriculture, like in Western Europe where an industrialized agriculture had also caused environmental problems. Emphasis is now being placed on rational land use, which includes optimization of farm size and the development and implementation of economically viable crop production techniques which result in high quality produce as well as limited adverse environmental side effects. However, these ideals are far from being realized. According to some studies, the technical possibilities of modern agriculture can theoretically in future provide an adequate volume of produce from only 30 to 50% of the current agricultural area in western Europe. This implies potential for other forms of land use in the remaining land area. However, other studies emphasizing low external input agriculture or extrapolation of historical trends, indicate that the current land area may be needed to produce adequate food, certainly when considering the future world market demand. This conclusion is also supported by the fact that theoretical production levels are often much higher than real levels, because of various agronomic and socio-economic factors. A major challenge for the decades ahead is to avoid uncontrolled developments of land use with possibly adverse socio-economic and environmental effects; the latter include the time-delayed release of harmful chemicals, currently held in some soils, into the environment and the enhanced emission of radiatively active trace gases from soils to atmosphere. Controlled developments yielding sustainable forms of agricultural land use in some areas and nature development in others, is to be preferred and should ideally be based on eco-regional approaches. The authors advocate initiation of comprehensive, exploratory studies for Europe in which sustainable production of major land units is defined as a function of different types of land management and in which Europe is seen as part of the world economy. Six exploratory studies are reviewed in this paper; however, none of these scenario studies has the necessary comprehensive character and none is based on adequate land data.

Carbon and nitrogen stocks in the soils of the Amazon Region

Soil nitrogen and organic carbon stocks, to a depth of 0.3 m and 1 m respectively, were determined for the Amazon Region using the soil and terrain (SOTER-LAC) database for Latin America and the Caribbean. Mean carbon densities, to a depth of 1 m, range from 4.0 kg m−2 for coarse textured Arenosols to 72.4 kg m−2 for the poorly drained Histosols. Mean carbon density for the mineral soils, excluding Arenosols and Andosols (30.5 kg C m−2), is 9.8 kg m−2. In total, the top 1 m holds 66.9 Pg C and 6.9 Pg N. About 52% of this carbon pool is held in the top 0.3 m of the soil, the layer which is most prone to changes upon land use conversion and deforestation.

Toward more realistic projections of soil carbon dynamics by Earth system models

Soil carbon (C) is a critical component of Earth system models (ESMs), and its diverse representations are a major source of the large spread across models in the terrestrial C sink from the third to fifth assessment reports of the Intergovernmental Panel on Climate Change (IPCC). Improving soil C projections is of a high priority for Earth system modeling in the future IPCC and other assessments. To achieve this goal, we suggest that (1) model structures should reflect real-world processes, (2) parameters should be calibrated to match model outputs with observations, and (3) external forcing variables should accurately prescribe the environmental conditions that soils experience. First, most soil C cycle models simulate C input from litter production and C release through decomposition. The latter process has traditionally been represented by first-order decay functions, regulated primarily by temperature, moisture, litter quality, and soil texture. While this formulation well captures macroscopic soil organic C (SOC) dynamics, better understanding is needed of their underlying mechanisms as related to microbial processes, depth-dependent environmental controls, and other processes that strongly affect soil C dynamics. Second, incomplete use of observations in model parameterization is a major cause of bias in soil C projections from ESMs. Optimal parameter calibration with both pool- and flux-based data sets through data assimilation is among the highest priorities for near-term research to reduce biases among ESMs. Third, external variables are represented inconsistently among ESMs, leading to differences in modeled soil C dynamics. We recommend the implementation of traceability analyses to identify how external variables and model parameterizations influence SOC dynamics in different ESMs. Overall, projections of the terrestrial C sink can be substantially improved when reliable data sets are available to select the most representative model structure, constrain parameters, and prescribe forcing fields.

Potential emissions of radiatively active gases from soil to atmosphere with special reference to methane: Development of a global database (WISE)

The role of soil in controlling production and fluxes of biotic greenhouse gases is the focus of research in progress at the International Soil Reference and Information Centre (ISRIC). There are two main goals in this project on World Inventory of Soil Emission Potentials (WISE). The first is to assemble a global soil database in association with the Land and Water Division of the Food and Agricultural Organization (FAO), using a ½° × ½° grid of geographic soil data (1:5 M scale). This “area” data will be linked to a database of soil profile “attribute” data using a geographical information system. The foundations for this work have now been put in place and, providing the soil profile collection programme proceeds satisfactorily, it is anticipated that a preliminary database should begin to emerge by the end of 1993. When the soil database is complete, the second thrust will be to make an inventory of the worlds poorly drained soils, providing the geographical framework for an improved estimate of methane production potentials. To do this, controlled long-term field experiments are required and modeling techniques must be developed and tested. ISRIC is cooperating with the International Rice Research Institute (IRRI) in the Philippines for these aspects of the work. An important corollary to the development of a global soil database is that many requests are being received for soil information relevant for studies of “global change.” At present, much of this information does not exist in an adequate format, so ISRIC is proceeding as rapidly as possible to implement the WISE digital database in a format which is compatible and user-friendly, for ultimate distribution in the public domain.

Exploring land quality effects on world food supply

In a previous study, simulations of agricultural production potentials were made for different exploratory scenarios considering population growth, type of diet and low and high input agriculture. Results indicated that future world populations can be fed, but problems are likely in South Asia. The simulations involved gross generalizations for soil conditions. For example, possible effects of soil degradation were not expressed. The current study was made to explore the effects of the different types of soil degradation on agricultural production, using major soil groupings of the Humid Tropics and Seasonally Dry (Sub)Tropics as examples. Degradation (compaction, erosion and acidification) is expressed in terms of soil quality indicators relating potential to actual production. Results are characteristically different for different soil units (genoforms), and the suggestion is made to present such differences in future soil databases for phenoforms that express effects of different forms of degradation or improvement, allowing better assessments for exploratory land use scenarios. Field studies should be initiated to describe realistic phenoforms for any given genoform.

Effects of mapped variation in soil conditions on estimates of soil carbon and nitrogen stocks for South America

Abstract Organic carbon and total nitrogen stocks for South America are computed using four 1:5,000,000 scale soil data sets of different spatial resolution. These are the 60′ by 60′ resolution Zobler soil data file, the 30′ by 30′ resolution World Inventory of Soil Emission Potentials (WISE) database, a 5′ by 5′ raster version of the Digital Soil Map of the World (DSMW), and a vector-based Soil and Terrain Database (SOTER). Estimates for total, regional stocks of organic carbon (SOC) range from 146.1 Pg C (Zobler) to 159.7 Pg C (SOTER) for the first 1 m of soil. More pronounced differences are observed among the four data sets when values for SOC stocks are compared at the level of the major soil grouping and their component soil units. In the case of Ferralsols, for example, the estimates for SOC stocks are 31.1 Pg C (SOTER), 41.4 Pg C (DSMW), 41.6 Pg C (WISE), and 58.3 Pg C (Zobler), respectively. In the case of Xanthic Ferralsols (Fx), the SOC stocks estimated with the Zobler and DSMW sets differ by about 14 Pg C (≈157%). By and large, the observed differences are due to varying areal estimates for the various soil units for the data sets under consideration, and the selection of soil profile data used. This study illustrates (a) the importance of the aggregation procedure used to account for the mapped spatial variation in soil conditions and (b) the need for updating the soil geographic and attribute data for the world in SOTER, so as to provide modelers with up-to-date soil data for studying interactions of human and natural systems at the regional and continental level within the context of global change.