| dc.description.abstract | The stability of the Amazon rainforest against precipitation changes has been commonly studied by means of conceptual models, from which the existence of two equilibria (savanna and tree-covered) emerges. These conceptual models are based on large spatial and temporal averages of climatic variables. However, no final, comprehensive proof of bi-stability is available for the Amazon, but results are supported by a growing evidence for it. On the other hand, biome models coupled to atmospheric models better reproduce the temporal and spatial dynamics of the Amazon basin, at the cost of higher computational cost. In literature, the only available studies which make use of a coupled vegetation-atmospheric model to show multiple equilibria have a spatial resolution of ∼2° (∼220km): too coarse to properly represent the local variations of the precipitation field, especially in a system (such as the Amazon) where precipitation is one of the main drivers of change. Here we used the results of a fully-coupled, fixed-vegetation climate model run under a yearly 1% pCO2 increase with a spatial resolution of 0.25° (∼28km) and a temporal resolution of 3h to project the climatic variables (relevant for the Amazon) for end-of-century. We then fed an equilibrium biome model with the output of the climate model, ran sensitivity analyses of the biome model to different parameters, and framed the end-of-century projections among the simulated equilibrium states of the rainforest. Although substantial changes (-24% trees) in the biomes of the rainforest are projected for end-of-century, the lack of vegetation feedbacks in this setup prevented us from investigating any non-linear behaviour. Thus, we successfully implemented a simple evaporation advection scheme, which shows a coupled response of precipitation and biome variations, and suggests a stronger response between the rainforest to climate change if advection is included. This method paves the way for similar studies that could resolve non-linearities, exploiting state-of-the-art climate models and sophisticated biome models to explore spatially-resolved biome distributions with wider ranges of climate forcings. | |
