Lethal Psi: Characterising critical embolism thresholds for Amazon tree survival

The Southern Amazon faces the greatest climatic threat of all Amazon regions. This region is drier and warmer than 'core' areas of the Amazon and has been subject to the most pronounced drying and warming trends. It is also the region of the Amazon where increases in tree mortality have been most marked and where atmospheric measurements suggest forests are no longer acting as a carbon sink but as a net source of carbon to the atmosphere. Given that Southern Amazon is at the front line of the Amazon's battle against climate change, it is essential that we better understand how resistant its forest species are to climate stress.

In Lethal Psi, we will construct a new 1-hectare drought experiment to better understand the physiological survival limits of southern Amazon trees. It has become increasingly clear that the process of hydraulic failure plays an important role in drought-induced tree mortality. Water is transported from the soils to the canopy under tension. As drought ensues and the soil dries, the tension in the xylem vessels that transport water intensifies and this can lead to the formation of air bubbles (embolism) in xylem vessels, disrupting water transport to the canopy and ultimately resulting in tree death. While this process is understood in general terms, one critical current knowledge gap is that we don't know the thresholds in embolism formation that result in the death of tropical trees. This lack of understanding of the physiological thresholds that result in death constitutes a key uncertainty for accurately modelling tree mortality under climate change.

Determining the hydraulic thresholds of tree death is not an easy task and requires monitoring tree hydraulic status up to the point of death. In Lethal Psi, we track key indicators of hydraulic function (e.g. leaf water potentials and sap flux) from the beginning of our imposed drought all the way to the death of the tree to quantify how loss of xylem conductance translates into mortality risk. While other drought experiments have been set up in Amazonia, these did not monitor embolism status before and during the mortality process and were thus unable to provide insights into physiological thresholds of survival. Up to now, drought experiments have only been set up northeastern Amazonia, where annual rainfall is almost twice that of our study site and where changes in climate have been much less pronounced than in southern Amazonia. Given their ecotonal nature and the rapid climate change experienced in southern Amazonia, we expect that trees in this region are much closer to their climatic limits and will experience much more accentuated mortality under imposed drought than observed in northeastern experiments.

Ultimately, we plan to use the newly acquired field data to develop improved mortality functions that we will apply more broadly across southern Amazonia to better predict drought mortality risk of this critically important region. This will be done by updating a unique trait-based model specifically developed to simulate Amazon forests and their responses to environmental change.