Quantification of transpirable soil water explains tree water use dynamics in a semi-arid pine forest (original) (raw)

Knowledge of the relationship between soil water dynamics and tree water use is critical to understanding forest response to environmental change in water-limited ecosystems. However, the dynamics in soil water availability for tree transpiration (T t ) cannot be easily deduced from conventional measurements of soil water content (SWC), notably because T t is influenced by soil water potential (Ψ s ) that, in turn, depends on soil characteristics. Using tree sap flow and water potential and deriving depth-dependent soil water retention curves, we quantified the 'transpirable soil water content' (tSWC) and its seasonal and inter-annual variations in a semiarid Pinus halepensis forest. The results indicated that tSWC varied in time and with soil depth. Over one growing season T t was 57% of rain and 72% of the infiltrated SWC. In early winter, T t was exclusively supported by soil moisture at the top 10 cm (tSWC = 11 mm), whereas in spring (tSWC > 18 mm) and throughout the dry season, source water for T t shifted to 20-40 cm, where the maximum fine root density occurs. Simulation with the soil-plant-atmosphere water and energy transport model MuSICA supported the idea that consistent tSWC at the 20-40 cm soil layer critically depended on limited water infiltration below 40 cm, because of high water retention below this depth. Quantifying tSWC is critical to the precise estimation of the onset and termination of the growing season (when tSWC > 0) in this semi-arid ecosystem. Figure 7. (a) Transpirable soil water content (tSWC, n = 2-3) in three depth layers and precipitation in Yatir forest during 2006-2010 (annual precipitation amounts were 224, 308, 200, 160 and 289 mm, respectively). Droughts during 2008 and 2009 resulted in long intervals (up to 5 months) of zero water availability, ultimately leading to large-scale tree mortality. (b) Zooming in on the linear decline of tSWC in layer III after 30 April. Regression slopes are -0Á012, -0Á011, -0Á012, -0Á011 and -0Á009% d À1 for 2006-2010, respectively (R 2 = 0Á92-0Á97), allowing early estimate of the date of tSWC exhaustion in this forest ecosystem.