The AR4 is a good starting point, followed by a more detailed look at the papers it cites. They mention a savannah would be more resilient to stresses like droughts, so the forest would "probably" be replaced by such an ecosystem. This is the relevant chapter:

http://www.ipcc.ch/publications_and_data/ar4/wg2/en/ch13s13-4.html#13-4-1

Concerning precipitation, models still disagree as to what could be the end result. Some suggest more rain, others suggest less. The intense droughts of the last decade may suggest the latter would be the case. A somewhat outdated paper (2006), but still interesting, is this:

Marengo, Jose Antonio. On the Hydrological Cycle of the Amazon Basin: A historical review and current State-of-the-art. Revista Brasileira de Meteorologia. 2006

http://www.rbmet.org.br/port/revista/revista_artigo.php?id_artigo=202

It suggests an El Niño-like condition made permanent, with stronger equatorial winds bringing rain from the Amazon basin to be poured down South, at the La Plata basin.

Some brazilian researchers point out that the amazon basin hydrology is mainly ruled by water transfers derived from the oceans. Does someone knows evidences for that, or otherwise?

Thanks a lot, guys.

From what I have read, global warming caused by CO2 would have little impact of equatorial regions. The greatest increases in temperature relative to normal (in general) should be in the polar regions, but as stated above, there is little agreement among the general circulation models concerning temperature and precipitation, especially at the regional level. Climate does change, it has been changing for millions of years, but the most likely factor altering equatorial climate is changes in ocean circulation patterns.

There are certain parts from the Amazon Rainforest territory where the savannas occur naturally. This does necessarily mean that it should be the next ecosystem in case of devastation. the BDFFP project (INPA/Smithsonian joint program) strongly research on forest fragmentation and ecology dynamics on central Amazon, give a look on their site and search for their works, they have like 30 years of data collected (4 months of that by me :D ). I did not hear about ocean waters controlling the dynamic of the basin but as far as I studied the combination of equatorial climate and topography of the region within the lack of strong winds keeps the flow of rainy season between north and south parts. That would be such a break-thru study to find such relations...

Thanks for all the information, I really needed it. I study Geography and I have a subject that called Geography of Humid Tropics where we study all the process and dynamic that occur in this region, specially the relationship between human being and his geographical space, that in this case would be the amazonian rainforest.

Alfredo, I am a geographer too and have worked on tropical rainforests in Brazil and El Salvador. One of my students worked on human land use patterns of deforestation, we have also published on use of satellite imagery for mapping dark earth soils (Terra Preta) under the rainforest in the Amazon region of Southern Brazil. We have used satellite imagery to study deforestation for cropping and grazing, and for mapping forest biodiversity in El Salvador.

I have an article on the subject: http://ssrn.com/abstract=1995799.

The Future of the Amazon Forest: Trends and Debates. http://ssrn.com/abstract=1995799.

Abstract

The future of the Amazon rainforest is a matter of much concern worldwide. On the one hand, climate change has been said to threaten the forest via possibly increased drought, prompting fires and fragmentation, accelerated by man-made deforestation. On the other hand, the Amazon forest itself is a major carbon store area, and carbon dioxide release from deforestation or forest degradation may strongly contribute to overall future carbon concentrations in the atmosphere, thus increasing global warming. Some authors have suggested the possibility of a catastrophic savannisation of the forest in a relatively short time due to the combination of several processes. This paper, sticking to the climate projections of the latest (2007) report from the Intergovernmental Panel on Climate Change (IPCC) as well as available data and scientific results, examines the evidence about deforestation trends and about expected changes in climate over the Amazon, and concludes that deforestation rates are much lower than previously thought and rapidly decreasing. Results on gross deforestation in 2010-11, released in late December 2011, confirm and continue the declining trend of previous years. Gross deforestation reflects conversion of forest into non-forest. Net deforestation of primary forest (or decrease in primary forest area) is even lower and declining faster, once re-growth and the clearing of secondary forest are taken into account. Total forest area change would also be lower once these factors, plus plantations, are considered. Even in the absence of such declining trends in deforestation, catastrophic forecasts of rapid Amazon 'die-back' or savannisation, resting on the hypothesis of decreasing moisture in the Amazon basin, are also shown to lack scientific basis and to be not supported by IPCC climate projections, especially when such events are predicted to occur in a few decades or during this century.

When anthropogenic earth art are being frequently found under the understory of the Amazonian forests, one has to wonder if we currently have all the facts right about past land use of this area. See the link to a new article about some of the more recent discoveries.

http://www.chinadaily.com.cn/sunday/2012-01/29/content_14498163.htm

Very relevant article:

Richard T., Corlett. “Impacts of Warming on Tropical Lowland Rainforests.” Trends in Ecology & Evolution 26, no. 11 (November 2011): 606–613.