HR: 08:45h
AN: B11G-04 INVITED    [Abstracts]
TI: Do iron and aluminium oxides stabilise organic matter in soil? A multi-scale statistical analysis, from field to horizon
AU: * Moni, C
EM: moni@grignon.inra.fr
AF: Laboratoire de biogéochimie et écologie des milieux continentaux, Centre INRA Versailles- Grignon; Bātiment EGER, Thiverval-Grignon, 78850, France
AU: Chabbi, A
EM: abad.chabbi@lusignan.inra.fr
AF: Laboratoire de biogéochimie et écologie des milieux continentaux, Centre INRA Versailles- Grignon; Bātiment EGER, Thiverval-Grignon, 78850, France
AU: Nunan, N
EM: nunan@grignon.inra.fr
AF: Laboratoire de biogéochimie et écologie des milieux continentaux, Centre INRA Versailles- Grignon; Bātiment EGER, Thiverval-Grignon, 78850, France
AU: Rumpel, C
EM: cornelia.rumpel@grignon.inra.fr
AF: Laboratoire de biogéochimie et écologie des milieux continentaux, Centre INRA Versailles- Grignon; Bātiment EGER, Thiverval-Grignon, 78850, France
AU: Chenu, C
EM: chenu@grignon.inra.fr
AF: Laboratoire de biogéochimie et écologie des milieux continentaux, Centre INRA Versailles- Grignon; Bātiment EGER, Thiverval-Grignon, 78850, France
AB: Although many "in vitro" experiments demonstrated that Fe and Al oxides stabilise organic carbon (OC) in soil, very few field studies validated that result, and none of them took into account the influence of scale and that of depth on the relationships linking OC to metal oxides in soil. In the present study, we characterised the latter as a function of depth and horizon at two different sampling scales: the field scale and the pedological pit scale. Twenty seven pits, 2 m deep, were opened over a 22 ha area. At large scale, samples were collected per horizons, whereas at fine scale, 150 positioned samples were collected over the surface of a selected profile. All samples were analysed for their content in OC, in Fe and Al oxides (amorphous Feo, Alo and crystalline Fed-o, Ald-o), and in both forms of Si associated either with amorphous or crystalline forms of metal oxides (Sio, Sid-o). For both sampling scales, relationships between OC and Feo, Alo, Sio, Fed-o, Ald-o, Sid-o were statistically investigated per whole profiles and per horizons. For whole profile at large and fine scale, results showed that all variables were correlated to OC. However, all variables, except Sio at fine scale, were also correlated with depth, suggesting that most of these correlations might only be due to a simultaneous depth dependency of correlated variables. On the contrary, within single horizons much less correlations were observed suggesting that this mode of investigation limits the confounding effects of depth. At fine scale per horizons, seven correlations independent from depth were observed in deep horizons only. Four of them were positive and implied amorphous form of Fe and Al suggesting that these oxides physically interact with OM in the subsoil, whereas the three last were negative and implied both forms of Si, suggesting that Si compete with OC for the fixation on sorption sites. In the three deepest horizons, slope coefficients of linear correlations between OC and Feo decreased with depth, suggesting that Feo surfaces were less saturated with depth. In the deepest horizon, partial least square regression showed that Feo, Alo, Sio explained 78% of the OC variability while it represents only 0.36 % of the soil mass. At large scale, the quasi absence of correlations proves that the fine scale of investigation is adapted to the minimisation of the lateral variability, which improves the results of correlations. As a conclusion, to lessen the confounding effect of depth and lateral variability, correlations between pedological variables should be performed per single horizons and at fine scale. Results, obtained with sampling procedure, show that even a small amounts of pedogenetic oxides do have an effects on the soil organic carbon content that might lead to its stabilisation.
DE: 4806 Carbon cycling (0428)
SC: Biogeosciences [B]
MN: 2007 Fall Meeting