Characterising wide spatial variation in population size structure of a keystone African savanna tree

Elephant impact on trees is of conservation concern in African savannas, but it's effect on population size structure of heavily utilised species has received little attention. Although studies have shown that the population structure and density of Sclerocarya birrea subsp. caffra (marula), a heavily utilised savanna tree species, varies widely across its range, the causes of this variation are poorly understood. This study aimed to describe and compare these divergent structures and develop an understanding of the relationship between population structure and site characteristics, specifically disturbance. Four distinct groups of population structure were identified: (1) adult dominated populations consisting of only adults and no juveniles or seedlings. The size class distributions (SCDs) of this group were bell-shaped, indicating an unstable population which may have come about through circumstances that no longer occur in these areas, for example human settlement or a massive episodic recruitment event. High impala utilisation of seedlings or rodent seed predation are proposed to be the main cause of the lack of recruitment; (2) rotated sigmoid (or negative J shaped) curves, representing stable population structures, when protected from elephants, as well as at high rainfall sites with elephants, indicating that without elephants and in suitable habitats with elephants, marulas are able to attain a relatively stable size structure and instability is not inherent; (3) juvenile dominated populations with almost no adults. Adults are relegated to a few refugia, where elephants are unable to reach them; (4) juvenile dominated populations consisting of mostly juveniles and few adults. These populations, all exposed to fire and elephants, have missing size classes varying between 5 and 40 cm in stem diameter and 2-8 m in height. This unstable structure was most likely caused by fire suppression of juveniles combined with elephant utilisation of individuals in the missing size classes. Consistent application of this pressure is postulated to result in the progression from stable to missing size class structure, to juvenile or adult dominated populations, and finally to local extirpation, with increasing elephant densities and concomitant utilisation, interacting with fire, rainfall and soil type. (C) 2011 Elsevier B.V. All rights reserved.