The response of sap flow to pulses of rain in a temperate Australian woodland

Abstract

In water-limited systems, pulses of rainfall can trigger a cascade of plant physiological responses. However, the timing and size of the physiological response can vary depending on plant and environmental characteristics, such as rooting depth, plant size, rainfall amount, or antecedent soil moisture. We investigated the influence of pulses of rainfall on the response of sap flow of two dominant evergreen tree species, Eucalyptus crebra (a broadleaf) and Callitris glaucophylla (a needle leaved tree), in a remnant open woodland in eastern Australia. Sap flow data were collected using heat-pulse sensors installed in six trees of each species over a 2 year period which encompassed the tail-end of a widespread drought. Our objectives were to estimate the magnitude that a rainfall pulse had to exceed to increase tree water use (i.e., define the threshold response), and to determine how tree and environmental factors influenced the increase in tree water use following a rainfall pulse. We used data filtering techniques to isolate rainfall pulses, and analysed the resulting data with multivariate statistical analysis. We found that rainfall pulses less than 20 mm did not significantly increase tree water use (P > 0.05). Using partial regression analysis to hold all other variables constant, we determined that the size of the rain event (P < 0.05, R2 = 0.59), antecedent soil moisture (P < 0.05, R2 = 0.29), and tree size (DBH, cm, P < 0.05, R2 = 0.15), all significantly affected the response to rainfall. Our results suggest that the conceptual Threshold-Delay model describing physiological responses to rainfall pulses could be modified to include these factors. We further conclude that modelling of stand water use over an annual cycle could be improved by incorporating the T-D behaviour of tree transpiration.