Data on temporal variability in Mg isotope ratios of atmospheric deposition and runoff are critical for decreasing the uncertainty associated with construction of isotope mass balances in headwater catchments, and statistical evaluation of isotope differences among Mg pools and fluxes. Such evaluations, in turn, are needed to distinguish between biotic and abiotic contributions to Mg 2+ in catchment runoff. We report the first annual time-series of δ 26 Mg values simultaneously determined for rainfall, canopy throughfall, soil water and runoff. The studied 55-ha catchment, situated in western Czech Republic, is underlain by Mg-rich amphibolite and covered by mature spruce stands. Between 1970 and 1996, the site received extremely high amounts of acid deposition and fly ash form nearby coal-burning power plants. The δ 26 Mg values of open-area precipitation (median of À0.79‰) at our study site were statistically indistinguishable from the δ 26 Mg values of throughfall (À0.73‰), but significantly different from the δ 26 Mg values of soil water (À0.55‰) and runoff (À0.55‰). The range of δ 26 Mg values during the observation period decreased in the order: open-area precipitation (0.57‰) > throughfall (0.27‰) > runoff (0.21‰) > soil water (0.16‰). The decreasing variability in δ 26 Mg values of Mg 2+ from precipitation to soil water and runoff reflected an increasing homogenization of atmospheric Mg in the catchment and its mixing with geogenic Mg. In addition to atmospheric Mg, runoff also contained Mg mobilized from the three major solid Mg pools, bedrock (δ 26 Mg of À0.32‰), soil (À0.28‰), and vegetation (À0.31‰). The drought of summer 2019 did not affect the nearly constant δ 26 Mg value of runoff. Collectively, our data show that withincatchment processes buffer the Mg isotope variability of the atmospheric input.