Fourteen hillslope soil profiles were sampled under natural vegetation (ie, grassland or forest) and plantations in the Nilgiri highlands, Southern India. Delta 13C ratios were measured at different depths and 14C ages determined for six profiles. In these highland soils where the turnover rate of organic matter is extremely low, the δ13C ratios of entire soil profiles have recorded signatures of past land cover. By correlating the data with results previously obtained from peat bogs and with knowledge concerning the history of human settlement, we distinguish three contrasting trajectories of palaeoenvironmental history and landscape change since the last glacial maximum. In the Central Nilgiris, between 18 and 10 ka BP, forest expansion occurred through the conjunction of a wetter climate (the maximum of southwest monsoon-related humidity occurring at c. 11 ka BP) and higher temperatures; since 10 ka BP, the reversal towards grassland vegetation is attributed to drier conditions. In the Western Nilgiris, where strong southwest monsoon winds permanently restrict forest patches to sheltered valley sites, steady but limited expansion of forest from 18 ka BP to the present is recorded and attributed to rising temperatures. The Southern and Eastern Nilgiris, where the northeast monsoon contributes 20% of the annual rainfall, are the less sensitive to fluctuations in the southwest monsoon. In these areas, rapid and extensive expansion of forest occurred mainly as a consequence of higher temperatures from 18 ka BP to the present. Massive deforestation by Badaga cultivators and Europeans planters followed after the sixteenth century AD.As a result, and in contrast with the Western Nilgiris where the land cover mosaic has remained remarkably stable in the last 18 ka BP, the current landscape differs sharply from the land cover pattern detected by the soil record.

No abstract

No abstract

No abstract

International audienc

No abstract

A specific methodology was developed to refine the complex clay mineralogy commonly encountered in soil environments. The soil examined was a Cambisol developed into a ferralitic paleosol. The sample was split into four sub-fractions of different particle sizes (<0.05, 0.05-0.1, 0.1-0.2, and 0.2-2 µm), and their respective mass contributions to the overall <2 µm clay fraction were determined. For each sub-fraction, X-ray diffraction (XRD) patterns were modeled using a trial-and-error approach based on the direct comparison of experimental and calculated profiles. Quantitative information derived from the fitting procedure for the different sub-fractions allowed for the determination of the complex mineralogy of the <2 µm clay fraction through the identification and quantification of eight clay phases. The results show that the finest and most reactive clay fraction (<0.05 µm) was totally hidden in the XRD pattern of the <2 µm fraction, the fraction commonly considered in soil mineralogical analyses. Similarly, this procedure revealed the presence of illite-smectite-chlorite and kaolinite-illite mixed-layer minerals seldom described in soil literature using classical methods. The use of this methodology improved our understanding of the pedogenesis of this soil through the identification and quantification of clay phases structural properties. The analysis of the evolution of structural parameters with particle size allowed for the detection of local modifications in the interlayer composition of expandable and hydroxy-interlayered vermiculite layers. Following this approach, key information can be derived to determine subtle changes in clay mineralogical composition that are related to microorganism and/or plant activity.

No abstract