Substrate type and termite discovery shape decomposition at the dry end of an African savanna rainfall gradient.

Decomposition is the process by which dead plant biomass is recycled and made available again for uptake by other plants. It is largely mediated by microbes and soil invertebrates. Global decomposition studies have demonstrated that decomposition is primarily temperature-driven with rainfall playing a secondary role, although to date, all global decomposition studies have used a single substrate to measure decomposition meaning differences with substrate type may be missed. Here we explored the decomposition of wood, dung and grass along a semi-arid rainfall gradient (380-650 mm/year) at three savanna sites in north-eastern South Africa. As the three sites experience a similar temperature range, we effectively controlled for temperature allowing us to explore the effect of rainfall in isolation. We used decomposition bags - with and without termite access - and measured mass loss after set time intervals. We predicted that the absolute decomposition of all three substrates would increase along the rainfall gradient. This assumption held for dung and grass, but we found the reverse for wood. Our study highlights the importance of considering multiple substrates as decomposition patterns and dominant agents can vary. Unexpectedly low wood decomposition at our wettest site suggests that the amount of dead wood available, and therefore probability of discovery by termites, may be an important factor explaining findings from decomposition studies. Resource availability and termite discovery is a neglected subject within decomposition studies and warrants further investigation.