Groundwater recharge by precipitation is often assumed by practitioners as well as scientists to be a slow process of filtration through layers of uniform texture analogous to Darcy's law. In most basins, however, rainwater also finds its way through macropores and preferential pathways to the shallow unconfined aquifers within hours of falling. Recharge phases may extend over several days, increasing groundwater levels, stored volume and hence baseflows into adjoining rivers. In this study, groundwater recharge and storage are computed from baseflow as separated with a nonlinear reservoir algorithm from time series of daily flows at gauging stations in northern Germany. Results are compared to groundwater level fluctuation in the catchments and to daily seepage rates measured in a lysimeter station. Peak times of the fast transfer of rain water through the vadose zone are generally the same. However, while recharge from baseflow ends when baseflow assumes its typical recession, the attenuation of lysimeter seepage may last much longer. The volume of lysimeter seepage is generally higher than the recharge in catchments due not only to different vegetation but also to rim effects impeding direct runoff. Furthermore, the lysimeter walls allow vertical fluxes only. Without further evidence or improved devices, lysimeter seepage should therefore not be indicated as groundwater recharge for the site or catchment. Findings also indicate that the shape of derived recharge unit responses is practically time invariant but with a strong seasonal variation in the recharge-rainfall ratio of precipitation events.