Tree growth responses to extreme drought events are not well predicted by climate

The shifts in temperature, vapor pressure deficit, and soil moisture associated with anthropogenic climate change are causing new extremes for trees. Understanding their impact on tree growth, however, remains challenging given limited large-scale soil moisture data, the complex interaction between multiple climate drivers, and size-dependent tree growth. We developed a new modeling framework that integrates individual-level growth trends with species-specific climate sensitivities. This hierarchical Bayesian model can accommodate different sampling regimes and is specifically designed to capture extreme growth responses across trees, species and ecosystems. Using new soil moisture data from WLDAS, we apply the model to 1.6 million observations of tree-ring width across Western North America. We identify the significant drought period of the 2000-2007 as causing exceptional reductions in tree growth. These reductions are well beyond those predicted from direct responses to temperature, vapor pressure deficit and soil moisture or their interactions, suggesting gaps in our fundamental understanding of tree growth responses to climate. Ultimately, these results demonstrate that tree growth is a critical indicator of drought, and that many current models may underestimate growth declines associated with extreme drought events.