The significant worldwide increase in observed river runoff has been tentatively attributed to the stomatal "antitranspirant" response of plants to rising atmospheric CO(2) [Gedney N, Cox PM, Betts RA, Boucher O, Huntingford C, Stott PA (2006) Nature 439: 835-838]. However, CO(2) also is a plant fertilizer. When allowing for the increase in foliage area that results from increasing atmospheric CO(2) levels in a global vegetation model, we find a decrease in global runoff from 1901 to 1999. This finding highlights the importance of vegetation structure feedback on the water balance of the land surface. Therefore, the elevated atmospheric CO(2) concentration does not explain the estimated increase in global runoff over the last century. In contrast, we find that changes in mean climate, as well as its variability, do contribute to the global runoff increase. Using historic land-use data, we show that land-use change plays an additional important role in controlling regional runoff values, particularly in the tropics. Land-use change has been strongest in tropical regions, and its contribution is substantially larger than that of climate change. On average, land-use change has increased global runoff by 0.08 mm/year(2) and accounts for approximately 50% of the reconstructed global runoff trend over the last century. Therefore, we emphasize the importance of land-cover change in forecasting future freshwater availability and climate.