Seeing Through Clouds

   
 

Unlike traditional optical sensors, radar is considered active as opposed to passive remote sensing. Instead of passively recording how much energy is being reflected by or emitted from the Earth as the spacecraft travels overhead, radar works by sending out a pulse of radio waves toward a target and then recording the strength and return time of the signal as it bounces back. That information tells the scientists both how far away the target is and what the surface looks like, since different surfaces will absorb and reflect the pulse in different ways.

  Page 3Page 5
 

 
Radar Maps of Flooded Areas of the Amazon
 

 

Although LBA is a Brazil-led study, it is an international affair. The National Space Development Agency (NASDA) of Japan mapped the Amazon floodplain as part of their Global Rainforest Mapping Project, using radar data collected by the Japanese Earth Resources Satellite (JERS-1). As the satellite mapped tropical rainforests around the globe, different groups around the world became responsible for processing the data and making them available to the scientific community in an easy-to-use format.

Bruce Chapman is a senior engineer at NASA’s Jet Propulsion Laboratory (JPL) in California, which is the organization selected by NASDA to handle the data coming in from South America. Chapman was a principal investigator on the project. “With an optical sensor,” he says, “it can take years to create a cloud-free image of the Amazon. Even the supposedly ‘cloud-free’ image still has some clouds because there are places in the Amazon where the clouds just never go away. Radar wavelengths penetrate the clouds and provide a detailed image of the forests below. The radio waves can even penetrate the forest canopy and reveal the layers of structure within the forest right down to the ground.”
 

 

Radar maps of the Amazon Basin reveal the seasonally flooded forest. In the pair of images above, black represents permanent waterways, dark grey represents forest, and light grey represents flooded areas. (Images based on data provided by the Global Rainforest Mapping Project)

 

It’s this ability to see the underlying structure that enabled them to map the extent of the flooding. The water underlying the forest canopy provides a kind of amplification of the returned radar signal. Explains Chapman, “The water underneath the canopy provides something we call a ‘double bounce reflection.’ This double bounce occurs when the radar waves bounce off two perpendicular structures: the very reflective surface of the water and the tree trunks. This double bounce makes the return signal very bright. When we see that really bright signal in the Amazon, there is a good chance there are partially submerged trees.”

Making the maps
The mapping of the Amazon took place in two phases: one data collection for the dry season and a second one for the wet. The first strip of radar data was obtained on September 27, 1995, over the east coast of South America. The satellite mapping progressed about 75 kilometers westward each day for the next 62 days, with the last strip collected over the west coast in mid-November. Beginning May 4, 1996, the satellite mapped the Amazon in flood. The picture was complete by July 3. Chapman and his team at JPL made the final maps available to the scientific community in March 2001.

Even with the radar data, though, there were limitations. The radar could only see rivers and streams at least 100 meters wide, but hundreds, possibly thousands of small streams branch across the Amazon. “To get those streams,” explains Richey, “we had to drill down even further, using Geographic Information Systems (GIS) data sets that had been collected over the years.” For the smallest streams they had computer models predict the volume and area based on topographic and geologic features.

next Putting Together Maps and Measurements
back Using the Right Tools

 

Photograph of Water and Trees
Flooded areas appear bright to radar because the radar waves are reflected directly back at the sensor. The first bounce, off the water surface, is away from the sensor, but the second bounce, off the tree trunks and canopy, redirects the beam back towards the source. (Photograph courtesy Jeffrey Richey)