We have calculated stand-scale transpiration (Ep) by sequential scaling up and down in a Stipa tenacissima grassland in the SE of Spain. To this end, we differentiated three tussock size groups obtained from the extrapolation of transpiration measured from leaves, applying corrective functions for overestimation of individual transpiration. The transpiration found that way was compared with the transpiration of the S. tenacissima grasslands (Ep) estimated by a multi-source evapotranspiration model that divides the surface into vegetated and non-vegetated evaporative components (Clumped Model). We conducted the research in two seasons with high and low water availability in the study area. The daily Ep estimated by the Clumped Model was positively correlated to the daily stand transpiration scaled down from individual transpiration (average difference - 6.5%). The strong coherence between Ep found using a stand-scale model and the stand transpiration calculated by sequential scaling (leaf-individual-population or stand), represents a breakthrough for future population or stand-scale study of these semiarid S. tenacissima grasslands. The difficulty and complexity involved in some cases in implementing and setting up evapotranspiration models in the field could be partly overcome by means of sequential scaling. © 2007 Elsevier B.V. All rights reserved.