Search for a command to run...
Abstract Aboveground ecosystem structure moderates and even confers essential ecosystem functions. This includes an ecosystem’s carbon dynamics, which are strongly influenced by its structure: for example, tropical savannas like those in central Kenya store substantial amounts of carbon in soil. Savannas’ belowground allocation of carbon makes them important for global carbon sequestration, but difficult to monitor. However, the labile soil carbon pool is responsive to changes in ecosystem structure and is thus a good indicator of overall soil organic carbon dynamics. Kenya’s savanna structure is controlled by belowground ecosystem engineers (termites), ambient weather conditions, and the aboveground engineering influences of large-bodied, mammalian consumers. As a result, climate change and biodiversity loss are likely to change savannas’ aboveground structure. To predict likely outcomes of these threats on savanna soil carbon, it is critical to explore the relationships between labile soil carbon and ecosystem structure, local climate, and mammalian consumer community composition. In a large-scale, long-term herbivore exclosure experiment in central Kenya, we sampled labile carbon from surface soils at three distinct savanna structural elements: termite mounds, beneath tree canopies, and the grassland matrix. In one sampling year, we measured total extractable organic carbon (TEOC), total extractable nitrogen (TEN), and extractable microbial biomass for each sample. Across three sampling years with varying weather conditions, we measured rate of labile soil carbon mineralization. We quantified areal coverage of each structural element across herbivore community treatments to estimate pool sizes and mineralization dynamics at the plot scale. Concentrations and stocks of soil TEOC, TEN, and microbial biomass were driven by the structural element from which they were sampled (soils collected under tree canopies generally had the highest of each). Large-bodied herbivore community composition interacted variably with concentrations, stocks, and carbon mineralization, resulting in apparently compensatory effects of herbivore treatment and structural element with no net effects of large herbivore community composition on plot-scale labile carbon dynamics. We confirmed engineering of structural heterogeneity by consumers and identified distinct labile carbon dynamics in each structural element. However, carbon and nitrogen were also influenced by consumer community composition, indicating potentially compensatory interacting effects of herbivore treatment and structural element. These results suggest that one pathway by which consumers influence savanna carbon is by altering its structural heterogeneity and thus the heterogeneity of its plot-scale labile carbon.