Temporal changes in species diversity and composition in abandoned fields in a trans-Himalayan landscape, Nepal

Abstract

Secondary succession is an increasing phenomenon due to global changes in agriculture policies and practices. The empirical findings are biased towards the temperate zone. Abandonment of agriculture fields is less frequent in the subtropical and tropical zones where agriculture areas are, in general, expanding. But there are exceptions; a rapid rate of abandonment of agricultural fields have taken place in the arid trans-Himalayan region, due to today’s globalization of economy. We analysed agriculture fields that were abandoned between 1950 and 2003 in a large u-valley in central Nepal (3400 m a.s.l.). The potential forest vegetation is dominated by Pinus wallichina and shrubs of junipers and cotoneaster species. We tested the intermediate richness hypothesis in relation to vegetation cover, soil development and whether old-field succession is convergent or divergent with species data from 242 1 m2 plots in 5 age-classes. The main species compositional turnover expressed by Detrended Correspondence Analyses (DCA) correlated, as expected, with time after abandonment. Fields that were abandoned a long time ago are closer to forest at the periphery of the agricultural landscape. Moisture of the soil significantly increased with age of abandonment, but total vegetation cover and pH were negatively related to age. Beta diversity expressed in DCA SD-units showed an increasing trend with age of abandonment, supporting the divergence pattern in old-field succession. The reason why the succession is not converging may be due to browsing by domestic animals that prevent a closed canopy of pines and juniper to develop. There was a significant hump-shaped pattern in species richness along the temporal gradient, which agrees with the intermediate species-richness hypothesis. There was a rapid increase in species richness in plots close to the villages that were used for haymaking which increased the seed input significantly.