Uncovering Spatial Feedbacks At Alpine Treeline Using Spatial Metrics In Evolutionary Simulations

The recognition of feedbacks between plants and their physical environment is important in biogeography. At alpine treeline, positive feedbacks for trees versus tundra include warming the canopy via lower albedo, increasing nutrients and water by increasing local atmospheric deposition, and reducing transpiration and abrasion by slowing wind; negative feedbacks include shading and cooler soil. To study the relations between the spatial pattern of trees, the feedbacks that ensue, and the potential change in the spatial pattern that results, we apply genetic algorithms to a cellular automaton model to simulate the advance of trees into tundra. In one model, spatial metrics on trees in the neighborhood of a tundra cell determine its probability of becoming a tree cell and are used in the fitness function. In a second model, the genetic algorithm addresses a polynomial series for the feedbacks and the frequency distribution of a spatial metric is the fitness function. The simulations presented here support the interpretation of self-organization in the alpine treeline ecotone. Moreover, in self-organizing systems, evolutionary computation has the potential to help understand the broad form of the functions describing system behavior.