Detecting the Amazon’s Seasonal Signal

  page 2page 4

Trying to describe the big picture of seasonal dynamics across the entire Amazon Basin puts scientists in a catch-22. The forest is so big that satellites are the only way to make observations of the entire forest. But measurements collected from hundreds of kilometers above can sometimes be hard to tie to specific biological processes on the ground. On the other hand, a handful of ground stations scattered throughout 7.5 million square kilometers of forest can’t tell the forest’s entire story, either. To make a convincing case for an Amazon-wide, dry-season green-up, Huete knew that he would need both perspectives: space-based and ground-based.

  Map of flux tower locations in the Amazon

For the satellite-view, Huete and his research team compiled 5 years of satellite vegetation data from the Moderate Resolution Imaging Spectroradiometer sensor on NASA’s Terra satellite. The maps are based on the relative amounts of red and near-infrared light that the sensor detects over a location on Earth. Chlorophyll in vegetation absorbs red light, while “scaffolding” (like cell walls) in the plants’ leaves reflect near-infrared. An area that reflects very little red light but a lot of near-infrared light back to space is likely covered in vegetation; scientists call this signal “greenness.” Greenness is an optical (light-based) way to measure forest productivity.


Flux tower sites (red dots) provide on-the-ground evidence of forest processes, but are too widely scattered to describe seasonal changes across the whole Amazon. Satellite data cover the whole area, but can be hard to link to specific forest biology, like carbon uptake. Huete’s team used both types of data to describe basin-wide seasonal changes in Amazon vegetation. (Map by Robert Simmon.)

  Photograph of the structure of a leaf

Like most tropical locations, the Amazon is very cloudy, especially in the rainy season. In the dry season, it can be pretty smoky from slash and burn deforestation and agricultural fires. The high humidity (water vapor) in the atmosphere can also interfere with measurements from satellites. To make sure their vegetation maps were free of clouds and other data contamination, Huete and his colleagues selected only the best-quality data from 2000-2005 and averaged them into a single “typical” Amazon year.

Once they had their example year, the team pored over the maps, looking at seasonal changes in greenness at three different scales: the whole Amazon, regional slices called transects, and small areas surrounding research towers like the one where Saleska worked. They subtracted wet season and dry season greenness values to identify seasonal patterns. At tower locations in undisturbed forests and sites that had been converted to pasture, Huete and Saleska made meticulous week-by-week comparisons of satellite greenness and ground-based measurements of carbon dioxide uptake.


Leaves interact with different wavelengths of light in different ways. Chlorophyll and other light-harvesting pigments absorb red light very strongly, while a layer of spongy “scaffolding” tissue in leaves reflects near-infrared light. Land covered with vegetation will absorb red light and reflect near-infrared light. Remote-sensing scientists call this light signature greenness. (Photograph copyright M. J. Davidson.)

  Map of difference in vegetation index between dry and wet seasons in the Amazon

At each scale, they saw the same pattern: undisturbed rainforests became “greener” and increased their photosynthesis throughout the dry season. In the regional transects, the scientists discovered that the longer the area’s dry season was, the greater the greening effect was. Even though it might seem like months with little or no rain ought to slow down the forests’ ability to photosynthesize, Huete says the reverse appears true. “The dry season, with less clouds and higher sunlight, is actually the ‘good’ season.”


After combining five years of satellite vegetation data, Huete and his team subtracted wet season measurements from dry season measurements. This map shows the results: vegetation index values were higher (greener) during the dry season than the wet season in places where the Amazon forests were undisturbed. At pasture sites, such as in the southeastern Amazon, the vegetation was less productive (browner) during the dry season. Apparently, lack of sunlight in the wet season limits forest growth more than lack of water during the dry season. (Map by Robert Simmon, based on data from the University of Arizona Terrestrial Biophysics and Remote Sensing Lab.)

Print this entire article