We still lack substantial understanding on how landscape characteristics shape the storage and release of water at the catchment scale. Here we use 13 years of fortnightly precipitation and streamflow δ<sup>18</sup>O measurements together with hydrometeorological data from 12 nested catchments (0.5 to 247.5 km<sup>2</sup>) in the Alzette River basin (Luxembourg) to study bedrock geology and landcover controls on streamflow generation.</p> <p>Streamflow responses to precipitation were highly variable. Runoff coefficients were typically higher in catchments dominated by less permeable bedrock (i.e. marls and claystones, <em>R<sub>c</sub></em> = 0.43 to 0.52) than in catchments with a high fraction of permeable bedrock (i.e. sandstones and conglomerates, <em>R<sub>c</sub></em> = 0.19 to 0.40). The fraction of new water (<em>F<sub>new</sub></em>, water younger than ~16 days in this study) determined via ensemble hydrograph separation was strongly related to differences in bedrock geology. <em>F<sub>new</sub></em> was highest in impermeable bedrock catchments (i.e. with a dominance of marls and claystone, <em>F<sub>new</sub></em> = 4.5 to 11.9 %), increasing with higher specific daily streamflow (<em>F<sub>new</sub></em> up to 45 % in one catchment). In catchments with an important fraction of permeable sandstone and conglomerates, high <em>F<sub>new</sub></em> variability with specific streamflow (<em>F<sub>new</sub></em> rising to 25 % in one catchment) was also found, despite a damped and delayed hydrograph response to precipitation and low <em>F<sub>new</sub></em> (means of 1.3 to 2.7 %). In the weathered bedrock catchments (i.e. dominated by schists and quartzites), rapid infiltration led to large fractions of water that was older than 12 weeks (~80 %) and very small fractions of water younger than two weeks (~3.5 %). <em>F<sub>new</sub></em> variability with streamflow was near zero, contrasting with the rapid response of the hydrograph to precipitation events. At high specific streamflow, <em>F</em><sub><em>new</em> </sub>was also correlated with bedrock geology and certain land use types. The extensive data set of streamflow δ<sup>18</sup>O enabled us to link water storage and release to bedrock geology. Such information is key for a better anticipation of water storage and release functions under changing climate conditions, i.e. long dry spells and high-intensity precipitation events.