Mass harvested per trunkload as a constraint to forage consumption by the African savanna elephant (

Background: African elephants can convert woodland to shrubland or grassland. Moderate conversion observed at low elephant densities may improve conditions for other animals, while extensive transformation at high densities may reduce plant and animal diversity. The threshold density separating facilitation from habitat destruction varies spatially and is partly determined by food choice, which differs between adult bulls and members of breeding herds. When elephants consume herbaceous forage, woodland damage is low but this increases when woody plants are the primary food source. Consequently, an understanding of diet selection by elephants is important for forecasting the degree of vegetation conversion. One hypothesis is that elephants select forage that provides the highest rate of intake. The mass harvested per trunkload is a constraint to intake and therefore this study sought to determine if trunkload mass changes seasonally; varies across common forage types utilised by elephants; and differs between adult bulls and members of breeding herds. Methods: Mechanistic models were used to estimate the mass harvested per trunkload of green grass, mixed green and dry grass, forbs, and leaves and bark from woody plants across a heterogenous, semi-arid savanna at a daily time step for one annual cycle. Separate models were constructed for adult bulls and members of breeding herds. Results: Harvestable mass changed seasonally for herbaceous forage and for leaves from woody plants but was constant for canopy bark. The maximum average trunkload mass of green grass was >75 times heavier than the bite mass reported for other grazers while trunkloads of leaves from woody plants were only eight times heavier than the bite mass reported for other browsers. This is attributed to the advantage provided by the trunk, which increases harvestable mass beyond the constraint of mouth volume, particularly when feeding on grass. Herbaceous forage yielded heavier trunkloads than leaves and bark from woody plants during the wet season, but this was reversed in the dry season. Adult bulls harvested heavier trunkloads than members of breeding herds for all forage types except forbs; and adult bulls harvested disproportionately large trunkloads of grass and bark. Conclusion: The strong correlation between the model outputs and well-established trends in the seasonal changes in elephants' diet suggests that elephants are preferential foragers of the largest trunkload on offer. Consequently, they are grazers when suitable herbaceous forage is available, and browsers when it is scarce. Green grass provides adult bulls with disproportionately large trunkloads and, therefore, adult bulls are predicted to have a strong preference for green grass. Availability of suitable green grass during the dry season may therefore buffer woodlands from heavy impact by adult bulls. Consequently, where possible, protected areas with elephants should aim to include key grass resources.