The effect of ENSO on transpiration in a dry evergreen forest in Cambodia: insights from a seven-year sap flow dataset

• Seven-year sap flow dataset (2011–2017) revealed responses of transpiration to ENSO. • Larger wet season transpiration was ascribed to a preponderance of younger leaves. • Transpiration decreased by 10–40 % during the drought, indicating leaf shedding. • Rapid recovery of transpiration after drought implied less effect on tree mortality. The effects of extreme drought on transpiration caused by El Niño in 2015–2016 have been investigated in tropical regions. Very few of these studies, however, have been conducted in tropical dry forests of Cambodia. We measured sap flux densities of Calophyllum thorelii Pierre and Monoon sp. in a dry evergreen forest in Cambodia for seven years from 2011 to 2017, and analyzed the long-term variations in scaled-transpiration with respect to the extreme drought. Prior to the drought, larger-than-expected wet season scaled-transpiration values were calculated from vapor pressure deficit and solar radiation. In the wet seasons, these tree crowns were dominated by younger leaves due to leaf shedding and flushing in the preceding dry period, leading to higher crown-level stomatal conductance. During the extreme drought, all test trees exhibited lower scaled-transpiration due to decreases in stomatal conductance spawned by the depletion of subsurface water, with the decline in transpiration exceeding its variation before the drought, thereby indicating the occurrence of soil drought-induced leaf fall. Despite the effects of extreme drought, transpiration recovered rapidly with the onset of rainfall in wet season and exceeded the expected value. These trends resulted from the prompt leaf recovery and the coincident preponderance of younger leaves in crowns, showing the limited effects on tree mortality by the drought. Because the test trees grew near a stream where the groundwater table was originally shallow, additional measurements are recommended for areas with deeper groundwater levels.