Effects of elephant browsing on woodlands in a Tanzanian National Park : measurements, models and management.

Abstract

1. Two woodland populations were surveyed in Ruaha National Park to assess the effects of elephant browsing. Forty per cent of the Acacia albida trees were dead, and 67% of the Commiphora ugogensis trees had been killed during the previous 6 years. The size distribution of both species showed a lack of regeneration and both populations are therefore in decline.
2. The proportion of A. albida trees killed by elephants increased with tree density, but the proportion of C. ugogensis trees killed was independent of tree density. Evidence from the literature suggested two more mortality patterns caused by elephant browsing: (a) the proportion of trees killed decreased with tree density: (b) elephants killed a fixed number of trees rather than a proportion of the population.
3. Simple deterministic models showed that very large numbers of elephants need to be killed in order to stabilize or reverse the woodland decline. The four browsing mortality patterns mean that different tree species behave differently under the same elephant population and therefore the benefit/cost ratio of culling (number of trees saved per elephant shot) is higher for some species than for others. The benefit/cost ratio declines rapidly with time: to be effective culling must start early, but Park managers probably realize they have an elephant problem when it is too late for culling to be effective.