Fig 2 - uploaded by Luís Moreira De Sousa
Content may be subject to copyright.
Source publication
SoilGrids maps soil properties for the entire globe at medium spatial resolution (250 m cell side) using state-of-the-art machine learning methods. The expanding pool of input data and the increasing computational demands of predictive models required a prediction framework that could deal with large data. This article describes the mechanisms set...
Context in source publication
Similar publications
Task scheduling algorithms in cloud have come a long way, from simplistic algorithms like first come first serve, to bio-inspired & machine learning algorithms like Q-learning and genetic algorithms. The main objective of any task scheduling algorithm is to minimize the number of execution cycles needed to completely and effectively execute a given...
Citations
Present global maps of soil water retention (SWR) are mostly derived from pedotransfer functions (PTFs) applied to maps of other basic soil properties. As an alternative, ‘point-based’ mapping of soil water content can improve global soil data availability and quality. We developed point-based global maps with estimated uncertainty of the volumetric SWR at 100, 330 and 15 000 cm suction using measured SWR data extracted from the WoSIS Soil Profile Database together with data estimated by a random forest PTF (PTF-RF). The point data was combined with around 200 environmental covariates describing vegetation, terrain morphology, climate, geology, and hydrology using DSM. In total, we used 7292, 33 192 and 42 016 SWR point observations at 100, 330 and 15 000 cm, respectively, and complemented the dataset with 436 108 estimated values at each suction. Tenfold cross-validation yielded a Root Mean Square Error (RMSE) of 6.380, 7.112 and 6.485 10⁻²cm³cm⁻³, and a Model Efficiency Coefficient (MEC) of 0.430, 0.386, and 0.471, respectively, for 100, 330 and 15 000 cm. The results were also compared to three published global maps of SWR to evaluate differences between point-based and map-based mapping approaches. Point-based mapping performed better than the three map-based mapping approaches for 330 and 15 000 cm, while for 100 cm results were similar, possibly due to the limited number of SWR observations for 100 cm. Major sources or uncertainty identified included the geographical clustering of the data and the limitation of the covariates to represent the naturally high variation of SWR.
BACKGROUND
Climate change is the main cause of biotic and abiotic stresses in plants and affects yield. Therefore, we sought to carry out a study on future changes in the agroclimatic conditions of banana cultivation in Brazil. The current agroclimatic zoning was carried out with data obtained from the National Institute of Meteorology related to mean air temperature, annual rainfall, and soil texture data in Brazil. The global climate model BCC‐CSM1.1 (Beijing Climate Center‐Climate System Model, version 1.1), adopted by the Intergovernmental Panel on Climate Change, corresponding to Representative Concentration Pathways (RCPs) 2.6, 4.5, 6.0, and 8.5 for the period 2050 (2041–2060) and 2070 (2061–2080), obtained through the CHELSA V1.2 platform, was chosen for the climate projections of the Coupled Model Intercomparison Project 5. Matrix images at a depth of 5–15 cm, obtained through the product of the SoilGrids system, were used for the texture data. ArcGIS version 10.8 was used to construct the maps.
RESULTS
Areas favorable to the crop plantation were classified as suitable when air temperature TAIR was between 20 and 29 °C, annual rainfall RANNUAL between 1200 and 1900 mm, and soil clay content CSOIL between 30 and 55%. Subsequently, the information was reclassified, summarizing the classes into preferential, recommended, little recommended, and not recommended. The current scenario shows a preferential class of 8.1%, recommended of 44.6%, little recommended of 47.1%, and not recommended of 0.1% for the Brazilian territory.
CONCLUSION
The results show no drastic changes in the total area regarding the classes, but there is a migration from these zones; that is, from tropical to subtropical and temperate regions. RCP 8.5–2070 (2061–2080) showed trends with negative impacts on arable areas for banana cultivation at the end of the century. © 2022 Society of Chemical Industry.
Landscape position co-determines soil formation, and hence soil properties are related to landscape position. In many landscapes, the importance of landscape position is large, and thus relations between soil properties and landscape position are strong. This is particularly true for agricultural landscapes, where tillage has smoothed over local variation in soils. It is less true for some forested landscapes, where local impacts of trees on soil formation have increased local variation. This raises the question whether soil-landscape relations are strong in natural grassland, with neither trees nor plowing. We answer this question for hillslopes in a natural grassland in the American Great Plains by quantifying soil-landscape relations and soil spatial autocorrelation based on a dataset of 100 soil observations. We find that soil-landscape relations are weak, even when accounting for topographic features specific for the study site, and that short-range variation in soil properties is about as large as long-range variation. We conclude that non-tree related natural processes disrupt the formation of clear catenas, and hypothesize that these include small burrowing mammals and infilling of parent-material fractures with finer material.
SoilGrids produces maps of soil properties for the entire globe at medium spatial resolution (250 m cell size) using state-of-the-art machine learning methods to generate the necessary models. It takes as inputs soil observations from about 240 000 locations worldwide and over 400 global environmental covariates describing vegetation, terrain morphology, climate, geology and hydrology. The aim of this work was the production of global maps of soil properties, with cross-validation, hyper-parameter selection and quantification of spatially explicit uncertainty, as implemented in the SoilGrids version 2.0 product incorporating state-of-the-art practices and adapting them for global digital soil mapping with legacy data. The paper presents the evaluation of the global predictions produced for soil organic carbon content, total nitrogen, coarse fragments, pH (water), cation exchange capacity, bulk density and texture fractions at six standard depths (up to 200 cm). The quantitative evaluation showed metrics in line with previous global, continental and large-region studies. The qualitative evaluation showed that coarse-scale patterns are well reproduced. The spatial uncertainty at global scale highlighted the need for more soil observations, especially in high-latitude regions.
