Prospecting water tension with the new bottom-up tool STRATEAU-Application to water usage in France Water use in France in 2006 has been quantitatively evaluated from final use data using the decision support tool STRATEAU. Indeed, declarative water-withdrawal tax data, often used for diagnosis, are inprecise and agregated. Instead, we use national data bases assessing human actitivy (such as demography, tertiary employement, industrial and power productions, agricultural inventory and yield, etc.), often more precise and systematic to estimate water need, demand and theoretical consumption. A similar approach for water resources could not be achieved due to the lack of data at that scale. The elementary descriptive mesh is the disctrict, nudged into two agregating spatial scales (administrative or hydrographic) up to national scale so as to attribute data to their representative scale-here the 24 hydrographic regions. Lower scale water balance is hindered by inter-basin transfert inconsistencies, uncharted usage or misplaced individual industries. Similarily, time-steps are one month and one year. Anthropic activities is also available in nudged definitions, following NAF nomenclature, from sectors to very refined categories. Data is therefore used at its best available representative level. Results confirm the dominance of power plant cooling in water withdrawal and of agriculture in water consumption. Estimates honor tax-related data with a mean error about 5%, hiding higher discrepancies according to sectors of activity and areas. Despite the lack of definite validation data, this good fit conforts the reconstitution method as well as withdrawal data while it indicates ways of progress. The originality of the results resides in their monthly description, their sub-territorialisation and their refinement over the sub-sectorial activities. For example with respect to drinking water, domestic usage largely dominates industrial or tertiary ones. With respect to agriculture, effectively irrigated areas have been estimated, as well as withdrawals and evapprative losses. Whence grassland, when irrigated, ranks first as a withdrawer per unit area, whereas grain corn remains the first absolute water consumer.