In Mediterranean areas high precipitation variability and crop dependence on soil water availability make the interaction between climate and agricultural management a key issue for mitigating N2O emissions. In this study we used the STICS model to capture the effect of a water deficit gradient and precipitation variability on N2O emissions and mitigation strategies (i.e. N fertilizer type, grain legumes introduction in crop rotations and crop residues management) in a rainfed Mediterranean transect (HWD-Senés, MWD-Selvanera and LWD-Auzeville, i.e. high, medium and low water deficit, respectively). The model was first tested against a database of daily N2O fluxes measured during twelve growing seasons of winter crops at the LWD site. Several scenarios were then run on each site, always over 9 successive growing seasons to take into account precipitation variability. STICS showed a good ability to simulate the driving variables of N2O fluxes at the daily time scale. The mean observed and simulated cumulative emissions during the growing season were 0.71 and 0.82 kg N2O-N ha-1 , respectively. The simulated N2O emissions (mean of all scenarios) decreased with increasing water deficit being 2.51, 0.65 and 0.26 kg N2O-N ha-1 yr-1 for LWD-Auzeville, MWD-Selvanera and HWD-Senés, respectively, which is consistent with published results. The lower N2O emissions in the driest sites were not only related to lower fertilization rates but also to other factors associated with the Mediterranean characteristics, particularly, the drier water regime. Simulated N2O emissions were highly sensitive to the interannual variability of the climatic conditions. According to the simulations, urea fertilizer would lead to slightly higher N2O emissions (+6 and +8%) than ammonium-and calcium nitrate, respectively. The incorporation of winter pea in the traditional cereal-based Mediterranean rotations would reduce by ca. 22% the N2O emissions in HWD-Senés without changing wheat yields. Differently, in MWD-Selvanera and LWD-Auzeville, N2O emissions would remain unchanged since the emissions associated to the decomposition of low C:N ratio pea residues would 3 counteract the lower application of N fertilizer. The systematic removal of crop residues at LWD-Auzeville would decrease the N2O emissions by 20%. However, this practice seems not recommendable if tillage is practiced due to the concomitant decrease of soil organic matter, fact that would worsen the C footprint of the system and increase the susceptibility to soil erosion. Our work highlights the interest of combining experimental and modelling approaches to account for climatic variability and evaluate long-term effects of N2O mitigation practices under Mediterranean conditions.