"Conservation of high-biodiversity tropical forests is sometimes justified on the basis of assumed hydrological benefits – in particular, reduction of flooding hazards for downstream floodplain populations. However, the ‘far-field’ link between deforestation and distant flooding has been difficult to demonstrate empirically. This simulation study assesses the relationship between forest cover and hydrology for all river basins intersecting the world’s tropical forest biomes. The study develops a consistent set of pan-tropical land cover maps, gridded at one-halfdegree latitude and longitude. It integrates these data with existing global biogeophysical data. We apply the Water Balance Model – a coarse-scale process-based hydrological model -- to assess the impact of land cover changes on runoff. We quantified these impacts of forest conversion on biodiversity and hydrology for two scenarios: historical forest conversion and the potential future conversion of the most threatened remaining tropical forests. A worst-case scenario of complete conversion of the most threatened of the remaining forested areas would mean the loss of another 3 million km2 of tropical forests. Increased annual yield from the conversion of threatened tropical forests would be less than 5% of contemporary yield in aggregate. However, about 100 million people – 80 million of them in floodplains -- would experience increases of greater than 25% in annual water flows. This might be associated with commensurate increases in peak flows, though further analysis would be necessary to gauge the impact on flooding. We highlight basins in Southeast Asia, southern China, and Latin America that warrant further study.