Generalizing soil properties in geographic space: Approaches used and ways forward.
Détails
Etat: Public
Version: de l'auteur⸱e
Licence: CC BY 4.0
ID Serval
serval:BIB_6F566DF18F92
Type
Article: article d'un périodique ou d'un magazine.
Collection
Publications
Institution
Titre
Generalizing soil properties in geographic space: Approaches used and ways forward.
Périodique
PloS one
ISSN
1932-6203 (Electronic)
ISSN-L
1932-6203
Statut éditorial
Publié
Date de publication
2018
Peer-reviewed
Oui
Volume
13
Numéro
12
Pages
e0208823
Langue
anglais
Notes
Publication types: Journal Article ; Research Support, Non-U.S. Gov't
Publication Status: epublish
Publication Status: epublish
Résumé
Soil is one of the most complex systems on Earth, functioning at the interface between the lithosphere, biosphere, hydrosphere, and atmosphere and generating a multitude of functions. Moreover, soil constitutes the belowground environment from which plants capture water and nutrients. Despite their great importance, soil properties are often not sufficiently considered in other disciplines, especially in spatial studies of plant distributions. Most soil properties are available as point data and, to be used in spatial analyses, need to be generalised over entire regions (i.e. digital soil mapping). Three categories of statistical approaches can be used for such purpose: geostatistical approaches (GSA), predictive-statistical approaches (PSA), and hybrid approaches (HA) that combine the two previous ones. How then to choose the best approach in a given soil study context? Does it depend on the soil properties to be spatialized, the study area's characteristics, and/or the availability of soil data? The main aims of this study was to review the use of these three approaches to derive maps of soil properties in relation to the soil parameters, the study area characteristics, and the number of soil samples. We evidenced that the approaches that tend to show the best performance for spatializing soil properties were not necessarily the ones most used in practice. Although PSA was the most widely used, it tended to be outperformed by HA in many cases, but the latter was far less used. However, as the study settings were not always properly described and not all situations were represented in the set of papers analysed, more comparative studies would be needed across a wider range of regions, soil properties, and spatial scales to provide robust conclusions on the best spatialization methods in a specific context.
Mots-clé
Geography, Models, Theoretical, Soil
Pubmed
Web of science
Open Access
Oui
Création de la notice
27/11/2018 22:38
Dernière modification de la notice
21/11/2022 9:19