Strategies and dynamics of plant water potential regulation under drought: Implications for the Amazon rainforest stability from an ecosystem modelling perspective

Capturing the complex interplay of plant hydraulic mechanisms and its role for vegetation responses under drought remains challenging for model development and parameterization. Here, we present a new dynamic modelling approach that simulates leaf water potential dynamics by taking differences in species-specific strategies of water-potential-regulation mechanisms into account and we show first results using it as part of the dynamic vegetation model LPJ-GUESS applied to the Amazon Rainforest. Our approach successfully simulates spatial patterns of changes in aboveground biomass and tree mortality induced by recent drought-events in the Amazon Basin. A special focus is put on differences in hydraulic strategies where we show, that either isohydric or anisohydric species may dominate the Amazon Basin, however both may co-occur in areas with particular high rainfall and low soil moisture stress. Testing our model with multiple input datasets reveals high sensitivity of parameters related to competition between different plant functional types regarding their hydraulic strategies. Furthermore, we want to highlight that particularly psi_50, the water potential at which 50% of the xylem-conductivity is lost due to cavitation, greatly affects species competition and composition in our simulations. Our results indicate that particularly those areas, which are used to high annual precipitation and hence where plants are less adapted to resist drought stress, may suffer severely under ongoing climate change and intensification of droughts.