Experiments and simulations of Turing patterns in vegetation*

Many pattern formations exist in ecosystems. From labyrinth patterns of bushy vegetation and spotted patterns of isolated tree patches in Niger to regular maze patterns of shrubs in West Siberia, and striped pattern of tree lines in tape forests in the USA, among various others. Several of these systems can, and have been described by reaction diffusion equations, however because in general ecosystems have large spatiotemporal scales there is little direct experimentation to quantify and validate the models which in addition have limited prediction powers.  We present controlled experiments of Chia seeds growing in half-meter square domains growing in four different substrates with different diffusion coefficients, amounts of daily irrigation and evaporation that display evolving transitions that include Turing patters ranging from spots, labyrinth, and mazes to continuum. We then model the experiments with a reaction diffusion system with parameters fitted to the experiments and used it to predict the patterns obtained in two different diffusive substrates as a function of daily irrigation and evaporation. We believe this is the first time a vegetation model has been validated directly with experiments and used for predictions.