From comfort zone to mortality: Sequence of physiological stress thresholds in Robinia pseudoacacia seedlings during progressive drought.

Abstract

Parameterizing the process of trees from the comfort zone to mortality during progressive drought is important for, but is not well represented in, vegetation models, given the lack of appropriate indices to gauge the response of trees to droughts. The objective of this study was to determine reliable and readily available tree drought stressindices and the thresholds at which droughts activate important physiological responses. We analyzed the changes in the transpiration (T), stomatal conductance, xylem conductance, and leaf health status due to a decrease in soil water availability (SWA), predawn xylem water potential (ψpd), and midday xylem water potential (ψmd) in Robinia pseudoacacia seedlings during progressive drought. The results showed that ψmd was a better indicator of drought stress than SWA and ψpd, because ψmd was more closely related to the physiological response (defoliation and xylem embolization) during severe drought and could be measured more conveniently. We derived the following five stress levels from the observed responses to decreasing ψmd: comfort zone (ψmd > -0.9 MPa), wherein transpiration and stomatal conductance are not limited by SWA; moderate drought stress (-0.9 to -1.75 MPa), wherein transpiration and stomatal conductance are limited by drought; high drought stress (-1.75 to -2.59 MPa), wherein transpiration decreases significantly (T< 10%) and stomata closes completely; severe drought stress (-2.59 to -4.02 MPa), wherein transpiration ceases (T< 0.1%) and leaf shedding orwilting is > 50%; and extreme drought stress (< -4.02 MPa), leading to tree mortality due to xylem hydraulic failure. To our knowledge, our scheme is the first to outline the quantitative thresholds for the downregulation of physiological processes in R. pseudoacacia during drought, therefore, can be used to synthesize valuable information for process-based vegetation models.