Tropical forest transpiration is a key driver of the global hydrological cycle and is often estimated through sap flow measurements. In large-scale water budget models, transpiration is estimated based on average sap flow among cohorts of functionally and dimensionally similar trees; however, sap flow variability in homogeneous cohorts has not yet been assessed in highly diverse tropical forests. We aimed to quantify sap flow variation among a cohort of 10 similarly sized conspecific trees in an Australian tropical rainforest. Over three field campaigns (dry, wet, dry seasons), we measured sap flow simultaneously on north and south aspects of the trunks using heat-ratio sap flow sensors and estimated crown exposure. Between the first and second campaigns, the forest was struck by a tropical cyclone, which led to a two-fold increase in our focal trees’ average crown exposure. Irrespective of aspect, within-tree variation ranged between 25% and 41% (median 34%) across the seasons, being higher in the dry seasons and lower in the wet season. Sap flow variation across the cohort was up to ten-fold under the same environmental conditions and was largely explained by crown exposure. Across the dry seasons, the north aspect maintained constant sap flow, while the south increased significantly by 44% under higher crown exposures following the cyclone. Our findings enable us to make two recommendations to decrease uncertainty in large-scale water budget models in tropical forests. First, sap flow measurements should be taken in at least two points in a tree stem to account for the 34% difference in readings. Second, crown exposure should be incorporated into sample stratification, since it not only may be a more reliable scalar than diameter and height, but may greatly reduce within-cohort variation.
