Imprimir Republish


Amazonia threatened by drought

Experiment in the world’s largest tropical forest reveals tree collapse and parched soil

The view from the top of a 40-meter tower shows that the tallest trees, prominent above the forest canopy, have died

Rafael Oliveira/UnicampThe view from the top of a 40-meter tower shows that the tallest trees, prominent above the forest canopy, have diedRafael Oliveira/Unicamp

When you drink juice through a straw, you have to be careful to keep the tube immersed. Otherwise, air bubbles will form and break the structure of the liquid column that brings the juice from glass to mouth. By increasing the scale of this scenario to the height of a 10-story building, you can get an idea of how water flows inside one of the gigantic trees found in the Amazon. Transpiration through leaves creates suction that moves water from the roots to the immense tops of the trees, which can exceed 40 meters in height, and launches moisture into the atmosphere that accounts for 35% to 50% of the region’s rainfall—a process that has a major impact on global hydrology. When this system fails, the water cycle is not the only thing affected. The trees, which have appeared to function normally until that point, suddenly die off. An experiment headed by British ecologist Patrick Meir of the University of Edinburgh, Scotland, and the Australian National University induced a 15-year drought in a section of the Amazon, thereby revealing the role of this mechanism, according to a paper published in November 2015 in the journal Nature.

Building the experiment required 500 cubic meters of wood, five metric tons of plastic, two metric tons of nails and 23,000 man-hours (10 men working continuously for a year), according to Antonio Carlos Lola da Costa, a meteorologist at the Federal University of Pará (UFPA). Their efforts produced 6,000 plastic panels measuring 3 meters by 0.5 meters each, interspersed with 18 100-meter-long gutters to prevent 50% of the rainfall from reaching the ground in a one-hectare parcel of the Caxiuanã National Forest in northern Pará State, where the Emílio Goeldi Museum of Pará maintains a scientific station. “Patrick contacted me in 1999 with this crazy idea,” Lola recalls. The meteorologist didn’t know where to begin, but he studied the photos that Meir sent him from a similar experiment, Seca Floresta (forest drought), that was being conducted in the Tapajós National Forest in western Pará, and went into the field. “In a year it was done.”

It was no trivial logistical feat. Getting to Caxiuanã entails leaving from Belém and spending 12 hours aboard a boat piled up with colorful netting, landing in Breves. In that city of about 100,000, Lola obtained the materials for his project, such as tubes of galvanized iron for building two 40-meter towers. From there, it is 10 hours by boat to Caxiuanã, where the materials had to be carried through the dense forest.

The experiment, known as Esecaflor, an abbreviation for Efeitos da Seca da Floresta (effects of forest drought), is the most extensive and longest-lasting ever undertaken to evaluate the effect of drought on a tropical forest. The only comparable experiment is Seca Floresta, which covered a similar area and ended after five years (see Pesquisa FAPESP Issue nº 156). For the past decade and a half, Antonio Carlos Lola has had primary responsibility for monitoring the reaction of the forest and keeping the experiment running despite constantly falling branches and trees—an effort that requires R$10,000 to R$15,000 each month. The cost is trending upward, now that more trees are succumbing to the drought and destroying parts of the structure. “I spend around six months each year in the middle of the forest, with a few interruptions,” says Lola, who has coordinated a number of master’s and doctoral student projects in connection with the experiment.


Long-term observation
The two Amazonian experiments tell similar stories overall, as indicated in a review article published by Meir and colleagues in the journal BioScience in September 2015. In the first few years, the forest appears to ignore the lack of rain and continues to function as normal. A few years into the drought, however, branches begin to fall and trees start to die, in particular the largest and the smallest specimens. Experiments in other countries have analyzed smaller areas or lasted for a shorter duration. The largest of these, in Indonesia, ran for two years.

Experimental fire in Mato Grosso State: under normal moisture conditions, fires are low-energy and minimally destructive

Paulo Brando/IPAMExperimental fire in Mato Grosso State: under normal moisture conditions, fires are low-energy and minimally destructivePaulo Brando/IPAM

The Caxiuanã study is yielding unprecedented findings because of its long duration: the collapse of the largest trees occurred only after 13 years of experimental drought, and this may represent a point of inflection, the point at which the forest undergoes a transformation. Since 2001 the researchers have been taking physiological measurements of the trees, comparing the rain-restricted area against a similar parcel that received no intervention. In the past two years they have begun to note drastic mortality among the tallest trees, which are rare by nature. When they fall, they cause destruction and transform a thriving forest into one with a degraded appearance. “Of the 12 tallest trees with a trunk diameter greater than 60 centimeters, only three remain,” says Lucy Rowland, a British postdoctoral researcher in Meir’s group at the University of Edinburgh, who has headed the project since 2011. The surprise came when they identified the hydraulic system as the internal cause of the tree mortality. Reduced water supply in the soil increases the tension in the water column inside a tree’s vascular system, known as the xylem. The integrity of this column, which depends on natural adhesion between water molecules, is compromised by air bubbles—a process that specialists refer to as cavitation. As a result of this collapse, which occurs suddenly, water cannot move from the roots to the leaves, and rapid tree death ensues. Meir points out that this hydraulic failure acts as a trigger that initiates the process of tree death, though it is not necessarily the final cause, which is as yet unknown.

Another theory advocated to explain tree death in drought situations is what researchers call “carbon hunger.” When leaves close their  stomata (the pores that enable transpiration and gas exchange) to prevent the leaves from drying out, they also reduce carbon absorption. The two processes most likely occur simultaneously, but in the case of Caxiuanã, the researchers dismissed carbon deficiency as the principal factor when they found that the trees contained a normal supply of that element and did not stop growing until they died.

“We measured the vulnerability of the plants’ hydraulic system to cavitation and saw that it is related to tree diameter,” says biologist Rafael Oliveira of the University of Campinas (Unicamp), who has worked on the project for two years. That observation is consistent with the preponderance of large-sized trees affected: 15 trees more than 40 centimeters in diameter fell in the experimental area, compared to only one or two in the control zone, where rain was not excluded. The impact is great, because these giant trees form a substantial portion of the forest‘s biomass and moisture-emitting canopy. At the same time, medium-sized trees grow taller because of the light that reaches them as the forest thins out and leaves gaps in the canopy.

Plastic panels prevent half the rainfall from reaching the ground...

Rafael Oliveira/UnicampPlastic panels prevent half the rainfall from reaching the ground…Rafael Oliveira/Unicamp

Oliveira has studied the relationships between the soil, plants and the atmosphere and, in a review published in 2014 in the journal Theoretical and Experimental Plant Physiology, he showed that changes in the precipitation regime can cause lethal water stress due to cavitation, although drought can be offset by a period of intense rainfall, so that the total annual rainfall remains unchanged. In his view, we need to get a better understanding of the physiological and anatomical functioning of the trees under these conditions in order to predict how they will react to expected climate change. These features also likely explain why the reaction varies from one species to another. The Caxiuanã study, for example, cites the genus Pouteria as very vulnerable to drought, and Licania as the most resistant of the trees examined. Plants use a variety of mechanisms, such as absorbing water through their aerial parts—the leaves, and even the branches and trunk. “We need to see which trees in the Amazon do this,” he says, looking ahead.

Another effect of tree mortality is the accumulation of more leaves and branches on the forest floor. “People who work with fire refer to this layer as fuel,” jokes ecologist Paulo Brando, a researcher at the Amazon Environmental Research Institute (IPAM) and the Woods Hole Research Center in the United States. Brando participated in the Seca Floresta experiment, whose extensive database is still being analyzed nearly 10 years after the project was completed. More recently he conducted a study of forest fires in an experiment in the Upper Xingu, the driest region in the Amazon. According to the findings presented in a 2014 article in PNAS, the trees readily resisted the first fire that occurred in 2004, partly because the moisture in the forest prevented the fire from achieving devastating proportions. The pivotal finding came in 2007, when the controlled burn coincided with a significant drought and, in the authors’ interpretation, represented a point of inflection in the forest. “What we saw was a very intense fire that killed everything, particularly the small trees,” he says, concluding that the interaction between drought and fire intensifies the forces that drive degradation.

Less water in the soil, less moisture in the air and more fuel on the ground act in combination and greatly increase the probability of fire. And not to be ignored is human activity on the agricultural frontiers of the forest, where fire, a common management practice, adds to the effects of deforestation, creating islands of forest with vulnerable edges. “A forest frontier with a soybean crop, for example, is five degrees Celsius warmer, and drier, than the forest interior,” Brando notes.

...causing trees to collapse...

Rafael Oliveira/Unicamp…causing trees to collapse…Rafael Oliveira/Unicamp

He coauthored a paper with geographer Ane Alencar, also at IPAM, that used satellite images to analyze records of fires in the Amazon between 1983 and 2007. The findings, published in September 2015 in Ecological Applications, show that there has already been an increase in the number of forest fires in response to a drier climate. In a comparison of three types of forest in eastern Amazonia, the group found that dense forest is sensitive to climate change, while open and transitional formations are more subject to human activity through deforestation.

Since there is no crystal ball to picture what lies ahead, several groups are seeking to develop climate and ecological models. Brando took part in a study headed by Woods Hole’s Philip Duffy that compared how well climate models have been able to take into account the 2005 and 2010 droughts in the Amazon, which were so drastic that they were thought to be once-in-a-century events. The findings, published in October 2015 on the PNAS website, project a significant increase in droughts, and an increase in the area affected by these droughts in the Amazon Region. The problem, Brando says, is that many of the models employ averages, but the issue in question is climate extremes. In November of 2015—a year characterized by an El Niño phenomenon that was stronger than average—the Esecaflor team found a forest that had received practically no rainfall for over two months. The researchers expect to monitor the consequences of that period over the next few years.

gutters carry the water away in a 1-hectare section of the Caxiuanã National Forest

Rafael Oliveira/Unicamp…gutters carry the water away in a 1-hectare section of the Caxiuanã National ForestRafael Oliveira/Unicamp

“The 2013 report of the Intergovernmental Panel on Climate Change (IPCC) pointed to our inability to predict drought-related mortality in the forests as one of the scientific uncertainties related to vegetation and climate,” Meir says. “Our findings indicate which physiological mechanism needs to be well represented in the models in order to predict tree mortality,” he explains. In an attempt to reduce uncertainties and anticipate the future, Lucy Rowland—who specializes in using field data to feed into models—has been working in partnership with the group headed by Stephen Sitch of the University of Exeter, England, to improve the representation of how tropical forests respond to drought in the vegetation model known as Jules.

The Amazon Region clearly demonstrates the importance of policies aimed at reducing climate change, a subject that has dominated the news recently because of the Paris Climate Conference (COP21), which took place in December 2015. The experiments show localized effects, but natural droughts such as those of the past decade can affect an extensive forest area. Meir stresses the need to break the cycle: as they decompose, large dead trees release a certain amount of carbon into the atmosphere, and this tends to aggravate the greenhouse effect. “It is possible to develop rules on energy and land use that are economically beneficial, without harming the environment in the long term,” he points out.

Soil-plant-atmosphere interactions in a changing tropical landscape (nº 2011/52072-0); Grant Mechanism Research Partnership for Technological Innovation (PITE) and FAPESP-Microsoft Research Agreement; Principal Investigator Rafael Silva Oliveira (IB-Unicamp); Investment R$1,082,525.94.

Scientific articles
ALENCAR, A. A. et al. Landscape fragmentation, severe drought, and the new Amazon forest fire regime. Ecological Applications. V. 25, No.  6, p. 1493-505. September 2015.
BRANDO, P. M. et al. Abrupt increases in Amazonian tree mortality due to drought-fire interactions. PNAS. V. 111, No. 17, p. 6347-52. April 29, 2014.
DUFFY, P. B. et al. Projections of future meteorological drought and wet periods in the Amazon. PNAS. On-line. October 12, 2015.
MEIR, P. et al. Threshold responses to soil moisture deficit by trees and soil in tropical rain forests: insights from field experiments. BioScience. V. 65, No. 9, p. 882-92. September 2015.
OLIVEIRA, R. S. et al. Changing precipitation regimes and the water and carbon economies of trees. Theoretical and Experimental Plant Physiology. V. 26, No. 1, p. 65-82. March 2014.
ROWLAND, L. et al. Death from drought in tropical forests is triggered by hydraulics not carbon starvation. Nature. On-line. November 23, 2015.