How do large herbivores, both wild and domestic, influence carbon cycling at the landscape scale in a Kenyan savanna? We explored carbon cycling within a large-scale, 21-year large herbivore exclosure experiment: the Kenya Long-term Exclosure Experiment, KLEE. A vibrant community of large herbivores (including functionally unique megaherbivores like elephants) persist there, and a millenia-long history of pastoralism in the region means domestic cattle herds coexist with wildlife. Within each plot, we also asked how large herbivores indirectly influence savanna carbon cycling via reassembly of ecologically important landscape features: Acacia drepanolobium trees and termite mounds, in addition to the background matrix of the savanna. We used classic soil science methods in addition to novel remote-sensing tools to measure soil organic carbon, nitrogen, soil microbial biomass and activity, and heterotrophic soil respiration across the experiment.
We determined that each feature within each herbivore treatment had a different size net pool of bioavailable carbon, nitrogen, and microbial biomass. We also found that landscape feature was the primary driver of the rate of soil microbial activity, with soils beneath tree canopies always respiring at faster rates than soils from the other two features. Because we saw significant differences in how carbon cycles and is stored in each landscape feature, it stands to reason that changes to the relative coverage of each feature on the plot scale would ultimately result in different total pool sizes in each herbivore treatment. However, these differences did not ultimately sum to significant differences in pool sizes at the treatment scale. While changes to large-bodied herbivore community composition, density, and identity may not be creating the distinct changes in pool sizes we hypothesized, they do clearly drive consistent patterns of landscape reassembly. These results supplement existing research in KLEE that found significant differences in pool sizes of total carbon across the herbivore treatments, and indicate that despite total carbon storage changes in response to community manipulation, the pool of carbon most readily available for uptake and cycling remains consistent though spatially redistributed.