Decoupling the effects of deforestation and climate variability in the Tapajos river basin in the Brazilian Amazon

Tropical river basins are experiencing major hydrological alterations as a result of climate variability and deforestation. These drivers of flow changes are often difficult to isolate in large basins based on either observations or experiments; however, combining these methods with numerical models can help identify the contribution of climate and deforestation to hydrological alterations. This paper presents a study carried out in the Tapajos River (Brazil), a 477,000km(2) basin in South-eastern Amazonia, in which we analysed the role of annual land cover change on daily river flows. Analysis of observed spatial and temporal trends in rainfall, forest cover, and river flow metrics for 1976 to 2008 indicates a significant shortening of the wet season and reduction in river flows through most of the basin despite no significant trend in annual precipitation. Coincident with seasonal trends over the past 4 decades, over 35% of the original forest (140,000 out of 400,000km(2)) was cleared. In order to determine the effects of land clearing and rainfall variability to trends in river flows, we conducted hindcast simulations with ED2+R, a terrestrial biosphere model incorporating fine scale ecosystem heterogeneity arising from annual land-use change and linked to a flow routing scheme. The simulations indicated basin-wide increases in dry season flows caused by land cover transitions beginning in the early 1990s when forest cover dropped to 80% of its original extent. Simulations of historical potential vegetation in the absence of land cover transitions indicate that reduction in rainfall during the dry season (mean of -9mm per month) would have had an opposite and larger magnitude effect than deforestation (maximum of +4mm/month), leading to the overall net negative trend in river flows. In light of the expected increase in future climate variability and water infrastructure development in the Amazon and other tropical basins, this study presents an approach for analysing how multiple drivers of change are altering regional hydrology and water resources management.