Drought responses in trees are context-dependent: the role of experimental design

Drought is a major driver of vegetation change yet estimates of plant responses to drought are derived from experiments that vary widely in design. This raises an important question: to what extent do reported drought responses reflect biological processes rather than experimental context? We conducted a meta-analysis of drought experiments in Eucalyptus and Quercus, comprising 712 effect sizes from 160 studies and 106 species across photosynthesis, stomatal conductance, intrinsic water-use efficiency (iWUE), leaf water potential, soil water content and root-to-shoot ratio. We evaluated variation associated with growth environment, drought-imposition method, study duration and sample size, and tested whether experimental-design effects persisted after accounting for realised drought intensity, quantified from drought-induced changes in leaf water potential. Drought responses varied substantially among experimental contexts but realised drought intensity overlapped broadly among pot and field experiments and between binary and progressive dry-down treatments, demonstrating that these experimental categories did not correspond to fixed levels of water stress. Accounting for realised drought intensity substantially improved model fit and reduced the estimated difference between pot and field experiments, although experimental-design terms remained jointly significant, indicating that realised drought intensity did not fully account for variation associated with experimental context. Trait-specific responses also differed. Greater realised drought intensity was associated with increasingly negative photosynthetic responses, whereas the corresponding relationship for stomatal conductance was weaker, and no clear relationship was detected for iWUE. Sensitivity analyses identified the photosynthetic response to realised drought intensity as the most consistently supported trait-specific relationship. Our findings show that tree drought responses are context-dependent and that categorical experimental designs should be interpreted alongside the water stress experienced by plants. Incorporating realised drought intensity together with methodological context should improve cross-study comparisons and the interpretation of experimental drought responses.