Dynamic Global Vegetation Models (DGVMs)
Dynamic global vegetation models aim to use biophysical principles to predict how vegetation is influenced by climate, soils and disturbances. We developed a DGVM in 2009 (which we called aDGVM, Scheiter and Higgins 2009, Global Change Biology) that allowed a better representation of how fire and plants interact. We have used this model to explore potential future vegetation states in Africa and Australia (e.g. Scheiter et al. 2015, New Phytologist). A few years ago we switched to developing a trait-based DGVM (which we called aDGVM2). aDGVM2 allows individual plant traits to evolve within simulation runs (see Scheiter, Langan and Higgins 2013, New Phytologist for an overview). Essentially this forces the model designer to focus more on trade-offs between traits than the trait values themselves. Although this dramatically simplifies the model parameterisation process, it does challenge our knowledge of fundamental trade-offs in ecology: this in turn set priorities for empirical work. Working with DGVMs is a sobering experience since it confronts one with the fact that our capacity to anticipate the plant communities that will assemble as climates change is rudimentary (see Higgins 2017, Ecosystems). Our work with the aDGVM2 is an attempt to model the assembly process by using a trait based approach that primarily focuses on trade-offs between traits and competitive interactions between species.