%0 Journal Article
%9 ACL : Articles dans des revues avec comité de lecture répertoriées par l'AERES
%A Zenero, M. D. O.
%A Grimaldi, Michel
%A Cooper, M.
%T Variability in soil shrinkage along forest and pasture toposequences in Amazonia
%D 2019
%L fdi:010075128
%G ENG
%J Geoderma
%@ 0016-7061
%K Shrinkage curve analysis ; Soil structure ; Soil porosity ; Soil hydro-structural properties ; Soil water holding capacities ; Land use
%K BRESIL ; AMAZONIE
%K PARA
%M ISI:000457657000030
%P 291-301
%R 10.1016/j.geoderma.2018.12.013
%U https://www.documentation.ird.fr/hor/fdi:010075128
%> https://www.documentation.ird.fr/intranet/publi/2019/02/010075128.pdf
%V 338
%W Horizon (IRD)
%X Physically fragile soils are subjected to rapid and abrupt land-use changes in many tropical countries. In this respect, it is important to characterize the changes in soil structure as an essential driver of soil functioning. Soil structure and hydric behavior can be analyzed together using soil shrinkage curves (SSC). The purpose of this study was to highlight the relevance of SSC analysis, even for soils with low shrink-swell capacity, to investigate the influence of three factors of vertical and lateral differentiation in the soil: soil horizon type, topographic position and land use. We analyzed SSC variability along two neighboring toposequences, one under forest and the other under pasture, in the southeast of the state of Para in the Brazilian Amazon. Quasi-continuous SSC were obtained using a Retractometer(C), and the pedostructure model was adjusted well to experimental data. Despite the overall weakness of the shrinkage and the high SSC variability, partial Redundancy Analysis demonstrated significant (P < 0.001) influence of each of the three factors addressed in our study. The influence of soil vertical differentiation was modest, explaining 30% of the total variance in parameters of the pedostructure model. Topographic position and land use had a weaker influence on the SSC, explaining only 5% and 4%, respectively, of total variance. These relatively weak influences, in the particular case of our study, were explained by few soil differences along the toposequences (i.e. Ultisols from uphill to downhill, extensive pasture management). However, statistical analysis of hydro-structural variables deduced from the SSC revealed significant differences among soil horizons regarding distribution of porosity between plasma and structural pores, and dynamics of these two compartments as a function of soil moisture, which is specific information provided by SSC analysis.
%$ 068 ; 082