COS 49-1 - On the origin and maintenance of savanna biome in Madagascar

Wednesday, August 14, 2019: 8:00 AM
L006, Kentucky International Convention Center
Nikunj Goel1, Erik Van Vleck2, Julie C. Aleman3 and Carla Staver3, (1)Department of Integrative Biology, University of Texas at Austin, Austin, TX, (2)Department of Mathematics, Kansas University, Lawrence, (3)Ecology and Evolutionary Biology, Yale University, New Haven, CT
Background/Question/Methods

Savannas are widespread in Madagascar, covering more than 70% of the island’s area. Despite their ubiquitousness, the origin and maintenance of the savanna biome in Madagascar has been a long-standing debate. On one hand, many naturalists argue that before human arrival 90% of the total land area (TLA) of Madagascar was covered with rainforest. However, following human occupation—around 10,600 yrs. BP—90% of the forests were converted into degraded savannas due to Tavy farming. However, paleo reconstruction, phylogeographic analysis, and species diversity maps indicate that pyrogenic savannas pre-date human arrival. Yet, the debate surrounding the origin of savanna biome—anthropogenic vs. natural—remains unresolved, for it rains enough in currently occupied savanna regions to climatically support forest. An examination of biome distributions in Madagascar reveals that the savanna-forest boundary is coincident with the eastern edge of the Central Plateau. We hypothesize that this plateau could have acted as a dispersal barrier, preventing forest expansion into mesic savannas of Central Madagascar.

Results/Conclusions

Here, we use a reaction-diffusion model and large-scale simulations to show that—the Central Plateau could have prevented forest expansion into mesic savannas of Central Madagascar by limiting dispersal. Thus, the savannas in Madagascar are maintained naturally via dispersal limitations and are not anthropogenically derived. These findings have implications for understanding the phylogeographic and biogeographic distribution of Madagascar’s endemic biota and have the power to inform conservation policies on the island nation. More broadly, this case study of Madagascar illustrates the ability of dispersal limitation to alter biome distributions, and thus incorporating dispersal constraints into global vegetation models might be crucial to predicting biome changes in the face of global change.