COS 50-9
Spatial fire spread and the evolution of flammability among grasses

Tuesday, August 11, 2015: 4:20 PM
329, Baltimore Convention Center
A. Carla Staver, Ecology and Evolutionary Biology, Yale University, New Haven, CT
Emmanuel Schertzer, Laboratoire de Probabilités et Modèles Aléatoires, CNRS, UPMC Univ. Paris 06, Paris, France
Background/Question/Methods

The evolution of flammability has attracted substantial attention from both empirical and theoretical perspectives. The paleo-literature debates whether flammability alone can explain diversification and ecological dominance of some plant lineages (e.g. tropical C4 grasses). From the theoretical perspective, flammability has been considered as both a cooperative and a competitive one. On the one hand, communities are only flammable if many members are flammable; on the other, flammability is a beneficial trait only inasmuch as it excludes non-flammable competitors.

Here, we consider a theoretical model for the evolution of flammability in the context of spatial fire spread. We consider fire as an infection process, such that only when a substantial fraction of a community is flammable can fire spread. Flammability is partially environmentally controlled (arid environments dry fuels) and partially an evolvable trait (some fuels dry more readily than others). Flammability traits are based on perennial tropical grasses; moist grasses decompose but do not burn readily, while drier grasses burn but do not decompose readily. Grasses that neither burn nor decompose self-shade and thus die more readily. After a mortality event, a patch is colonized by the genetic offspring of surviving neighbors, with some probability of a small trait mutation.

Results/Conclusions

We find that flammable and non-flammable landscapes represent alternative stable states in some environments, where fires favor flammability traits but the absence of fires favors decomposition. However, the evolution of a flammable trait, from a non-flammable ancestor, can only happen when the environment is so dry that fires are inevitable (i.e.when flammability is the only stable landscape configuration). In a temporally variable environment, flammable traits are thereafter maintained even when the environment becomes wetter, via a positive feedback with fire. In a spatially variable environment, flammable traits that evolved in a locally dry environment can spread, via nearest neighbor effects, into wetter areas of the landscape.

Thus, we find that fire alone probably did not drive the evolution and rise in ecological dominance of C4 grasses. However, flammability traits may have evolved in either a globally or locally fire-prone (e.g. a dry or high oxygen) environment, and fire might then plausibly have promoted the widespread persistence or invasion of flammable plants.