Knowledge on the water (sources and amount) used by vegetation is crucial for many ecohydrological studies and applications ranging from single-tree to catchment- and ecosystem-scales. Despite urgently needed for sustainable management of water resources under a changing climate, holistic assessments of the sources for plant transpiration for ecosystems are rare and methodological constraints persist, in particular in bridging the scales (i.e., single tree to plot, plot to catchment, catchment to ecosystem). Stable isotopes of hydrogen and oxygen are excellent tools to estimate sources and depths of water uptake but recent research has questioned their reliability in achieving robust source quantification even in small-scale studies. This is due to the combination of natural (mostly related to intrinsic heterogeneity of ecohydrological systems) and artificial effects that generate potentially large uncertainties in the estimates of water sources for plant transpiration. This uncertainty is rarely quantified and can have severe consequences when upscaling the results from small to larger areas. In this opinion paper we aim at i) raising awareness in the ecohydrological community of the potential sources of uncertainty in isotope-based estimation of water sources; ii) stressing the need to take this uncertainty into account in small-scale estimates of water sources; and iii) stimulate discussion and efforts to find new ways to upscale findings from small-scale investigations incorporating this uncertainty to provide more reliable conclusions for larger areas such as forests, catchments, and ecosystems.