Soil organic carbon responses to forest expansion on mountain grasslands

Grassland abandonment followed by progressive forest expansion is the dominant land-use change in the European Alps. Contrasting trends in soil organic carbon (SOC) stocks have been reported for mountainous regions following forest expansion on grasslands. Moreover, its effects on SOC properties involved into long-term stability are largely unknown. The aim of this PhD thesis was to explore changes in: (i) SOC stocks; (ii) physical SOC fractions and their sensitivity to SOC changes; and (iii) labile carbon (C) in soil and fractions following forest expansion on mountain grasslands. A land-use gradient located in the Southern Alps (Italy) was examined, comprising managed grassland, two transitional phases in which grassland abandonment led to colonization by Picea abies (L.) Karst., and old forest dominated by Fagus sylvatica L. and P. abies. Organic and mineral soil layers were collected within three plots for each successional stage. Soil samples were fractionated according to: (i) aggregate size fractionation, separating aggregates through wet-sieving, and (ii) size-density fractionation, separating stable aggregates from particulate organic matter (POM) non-occluded within aggregates. Changes in labile soil C were assessed by analysis of carbohydrate monomers and by thermal analysis of soil and size-density fractions. Our study showed that mineral SOC stocks were lower in early-stage and old forest (-28% in the top 10 cm of mineral soil) than in managed and abandoned grassland. The SOC accumulation within the organic layers following forest establishment could not fully compensate the mineral SOC stock difference between forest and grassland successional stages. The dimension of aggregates assessed by aggregate size fractionation tended to increase, while SOC allocation to stable aggregates assessed by size-density fractionation decreased following conversion of grassland to forest (e.g from 81 to 59% in the 0-5 cm layer). Higher SOC allocation to POM fraction in the mineral soil, together with changed carbon distribution between mineral and organic soil layers, suggest an overall decrease in physical SOC protection and a concomitant shift to unprotected SOC fractions. The size-density fractions showed higher sensitivity to SOC changes than aggregate size fractions and therefore may be better suited for the detection of SOC changes. Carbohydrate and thermal analyses indicated higher shares of labile soil C in grassland abandoned since 10 years than in managed grassland and forest successional stages. Ratios of 6 microbially to plant-derived carbohydrates decreased from grassland to forest successional stages, and corresponded to decreased SOC allocation to stable aggregates. Forest expansion on mountain grasslands caused a decrease in SOC stocks within the mineral soil and in physically protected SOC fractions, which can be explained by lower accumulation of binding agents of microbial origin. This can have implications for the accumulation of atmospheric CO2 in soil and for the susceptibility of SOC to external disturbances such as management and environmental changes.