Increased aridity may have severe effects on productivity of dry forests. However, it remains unclear to what degree the positive effects of elevated CO2 (both increased carboxylation rates and enhanced water-use efficiency) may offset the negative effects of drought and climate warming. In forest ecosystems, it is particularly challenging to evaluate CO2 effects on productivity because the impacts of climate variability, competition, and management, combine to have longlasting effects on stand-level productivity. Here we address this problem using a unique long-term database containing repeated inventories of wood biomass for every decade from 1912 to 2002 in a pine forest (Pinus pinaster Ait.) in central Spain (≈7,500 ha.). The approach is based upon a combination of statistical analyses of long-term historical management data and mechanistic modeling which allows us to evaluate the effects of potential CO2 fertilization, climate, and stand structure on woody net primary production (W-NPP). We found a significant negative effect of drought on W-NPP during the first half of the twentieth century that diminishes at the turn of the century. Simulations with a process-based ecosystem model, ORCHIDEE, suggest that wood production under conditions that included CO2 fertilization produced a more highly correlated long-term W-NPP than simulations keeping CO2 values in preindustrial levels. Interestingly, however, the CO2 effect was only apparent when accounting for confounding factors such as competition and management legacies. Identifying CO2 fertilization on forest growth is a critical issue, and requires partitioning CO2 effects from confounding factors that have jointly shaped stand dynamics and carbon balance during the twentieth century.