Linking historical land use to present vegetation and soil characteristics under slash-and-burn cultivation in Madagascar

Aim: In this paper, we link historical land use of cultivated parcels formerly slashed and burnt, with current secondary vegetation and soil characteristics, in the region of evergreen rain forest in Madagascar. Aim is to understand the legacy effect of former cultivation practices in order to optimize slash-and-burn cultivation. Location: Forest corridor of Fianarantsoa, southeast Madagascar. Methods: We examined three study sites to capture the ecological range in the forest corridor of Fianarantsoa. At each site, we selected formerly hand-cultivated parcels since first forest clearing so as to represent up to four classes of age of abandonment, encompassing forest fallow periods after cultivation of 0-5, 6-10, 11-20 and 21-30 yrs. These parcels served for selection of 50 plots. Interviews were conducted with each farmer to document former cultivation practices of the parcels, and from these, land-use metrics were calculated. Vegetation was surveyed in the plots and soil samples were taken and analysed for chemical and physical properties of the organo-mineral layer (horizon A). Results: We found that soil properties, such as pH, have higher values with an increasing number of cultivation cycles and duration of cultivation, and that this trend is inversely related to exchangeable P. C/N ratio is correlated to age of abandonment, and inversely related to soil N content. With increasing age of abandonment, species richness and diversity increase due mainly to an increase in ligneous species. The number of herbaceous and fern species increases with duration of cultivation and number of cultivation cycles. Conclusion: Historical land use has had a lasting effect on vegetation and even longer-term effect on soils. Different land-use variables explain soil and vegetation characteristics of the current secondary forest. The ability of soils to recover P appears unrelated to abandonment duration, suggesting that longer periods of time are required to recover P fertility, or that secondary vegetation is shifting the pool of nutrients into biomass. Our results point to the need for a management regime that would maintain both scattered trees and patches of a grass layer, since this might be key to promoting both biodiversity and soil restoration.