2017-01-01 ACTA OECOLOGICA-INTERNATIONAL JOURNAL OF ECOLOGY 2017 78(卷), null(期), (61-70页)
The hypothesis that African elephants may cause the local extirpation of selected woody species was evaluated in a medium-sized, semi-arid reserve following their reintroduction at low density. Mortality, state-change, and regeneration of 25 tree and 17 shrub species were studied between 1997 and 2010. Annual mortality of shrub species ranged from 0.2 to 8.0%, with six species experiencing 6-8%. Eight shrub species lost more than half their adult population (range 10-94%). Annual tree mortality ranged from 0.2 to 10.5%. The two dominant dryland tree species experienced <1% annual mortality, 18 species lost more than half their tree population, and one was eliminated. Elephants accounted for >63% and stress-related agents >20% of tree deaths. The manner in which elephants induced tree death depended on species. The proportion of individuals of a species killed by pollarding or uprooting ranged from 0 to 74%, and by debarking from 0 to 100%. Complete uprooting was a common cause of death for three shrub species. Regeneration ranged from zero for six tree and one shrub species to a seedling every 7 m(2) for Colophospermum mopane and 23 m(2) for Dichrostachys cinerea in riparian habitat. Three shrub and eight tree species were identified as vulnerable to local extirpation owing to a combination of high mortality and poor regeneration that is likely to result in a considerably simplified system. Reintroduction of elephants into medium-sized reserves without regulation of their numbers may not be a desirable action. (C) 2017 Elsevier Masson SAS. All rights reserved.