Climate change scenarios for the Iberian Peninsula predict increasing temperatures and increasingly variable precipitation regimes, which will challenge the sustainability and biodiversity of Mediterranean ecosystems such as the semi-natural evergreen oak woodlands. To assess the effects of precipitation variability on productivity, species composition and vegetation gas exchange of the understorey vegetation in a typical managed cork oak woodland, a large-scale rainfall manipulation experiment was established. We studied the impacts of a change in the timing of precipitation events on this ecosystem, without altering total annual precipitation inputs. The two water manipulation treatments were: 'weekly watering treatment', where natural conditions were simulated with a normal dry period of 7 days, and '3-weekly watering treatment', with the normal dry period increased threefold to 21 days. Our experimental precipitation patterns resulted in significant differences in temporal soil moisture dynamics between the two treatments. Average soil water content (SWC) at 3 cm depth during the growing season was 16.1 ± 0.17% and 15.8 ± 0.18% in the weekly and 3-weekly watering treatments, respectively, with a mere 5% increase in the variability of SWC when extending the dry period from one to three weeks. Water infiltration into deeper soil layers (>50 cm) was significantly higher in the 3-weekly watering treatment as compared to the weekly watering treatment. This might be beneficial to Quercus suber, the tree component in this ecosystem, as its extensive tree root system enables water acquisition from deeper soil layers. However, manipulation of the within-season precipitation variability, with a shift to fewer, but larger rain events, without change in total precipitation amount, had no significant effect on aboveground net primary productivity (ANPP), belowground net primary productivity (BNPP) and species composition, with average values of peak biomass of 385 g m −2 and 222 g m −2 for ANPP and BNPP, respectively. The experimental precipitation patterns did not result in significant differences in the vegetation gas exchange between the two watering treatments. The CO 2 and H 2 O exchange parameters correlated well with air temperature. In addition, evapotranspiration showed a good correlation with SWC. Incorporating the data of SWC in the conceptual 'bucket model' showed that, independently of the watering regime, soil water availability during the life-cycle of these annual plants did not reach severe water stress conditions, which can explain the lack of a significant treatment effect in our study. In addition, our results showed that the annual plant community in these Mediterranean ecosystems is well adapted to short-term drought, through their phenological patterns and physiological adaptations.