The annual cycle of evapotranspiration (ET) is an important component of the Amazon hydrological balance, which is of critical importance to the global water cycle. Understanding the changing water balance in this region is particularly important to estimate future global and regional hydroclimate change in response to projected deforestation of the rainforest in this region.

Several methods have been used to estimate the annual ET cycle in the Amazon basin. These different methods, which result in a spread of annual means, ranges, and phases of the ET cycle, are evaluated. In an attempt to reconcile the differences between them, both the data and the assumptions upon which the methods are based are scrutinized. The differences seem to originate from the geographic site where radiation and ET are simulated and/or observed and, more significantly, from the way that vegetation controls ET in the different models being used.

While field campaigns conducted during the Large-Scale Biosphere Atmosphere (LBA) Experiment in the Amazon have provided many new insights into the Amazon hydroclimate, additional observations of ET and precipitation in that region are needed to understand better the processes involved.