EXERCISE 2: Green, Greener, Greenest
Down here on Earth, it’s easy for us to spot vegetation and figure out where
it is thick and healthy and where it is sparse or struggling. The grass in our
lawns, the woods at the park, or the corn field at the farm down the road are
easy for us to identify. For a satellite orbiting 700 kilometers above the Earth,
it’s not so easy. In the next paragraphs is an explanation of the most common
way satellites identify vegetation.
Vegetation reflects both visible light (the kind we see with our eyes) and other radiation
from invisible parts of the electromagnetic spectrum in unique
ways. These unique reflection patterns are vegetation’s “spectral
signature.” Obviously, one of the spectral signatures of leafy vegetation
is that it’s green. Vegetation is green because it absorbs almost all other
wavelengths of visible light except green, which it reflects back to our eyes
or to a satellite.
The most commonly used technique for identifying vegetation in satellite data
is to compare the amount of visible light reflected by a particular place on
Earth to the amount of infrared light. Areas reflecting low levels
of visible light and high values of infrared light are likely covered
by leafy vegetation. Scientists use computer programs to subtract the amount of visible
light observed by satellites from the amount of near-infrared light, and
then divide that by the total amount of reflected light in those wavelengths.
Deciduous trees comprise as much as 40 percent of the forest in some parts
of Panama. Many of these trees lose their leaves during the dry season, from
January through February. The two images above show the same tropical forest tree, looking down
from above. This tropical tree is leafless in the dry season (like forests in North America that
lose their leaves in the winter), so only the branches are left. Below this leafless tree is a
shorter tree that is full of leaves. In the red band, the fully-leaved tree below absorbs red
light and the wood from the leafless tree above reflects red light. In the infrared, the
fully-leaved tree below reflects a lot of light, more than the bare branches above. (Photos
courtesy Stephanie Bohlman, U. of Washington)
This calculation produces an estimate of vegetation called the “Normalized
Difference Vegetation Index,” or NDVI for short. NDVI values have no units; they can
range from - 0.1 (low vegetation) to 0.9 (maximum detectable vegetation). NDVI images
are usually colored on a scale of brown for low vegetation to dark green for
dense vegetation. NDVI can reveal where natural and agricultural vegetation is
under drought or temperature stress and where it is flourishing. Over many years,
NDVI observations can show how vegetation and ecosystems may be appearing or
disappearing in response to human influence or climate change.
In this next exercise, you can use the Image Compite Editor to produce your
own NDVI measures. Using the pop-down menus, please select Landsat channel 4 as
the first channel and channel 3 as the second channel. You can use the ICE
tool’s math mode to perform the following operation: NDVI equals band
4 minus channel 3 divided by channel 4 plus channel 3 or written as a formula:
NDVI = (near infrared - red light) divided by (near infrared + red light).
Landsat: March 2000
Landsat: March 1991
Questions to consider:
- Do you see any differences in the features of the tree in the red and near-infrared
photos above? Explain any differences you see and what makes them look different.
- How is the way a satellite sees Earth’s vegetation similar to the way our
eyes see it? How is it different?
- What kind of “spectral signature” would you expect a rainforest
to reflect back to the satellite? Would it have a high NDVI value or a low NDVI
value? Explain why you think so.
- Would the NDVI value of the Panama Rainforest be higher or lower than vegetation
in a state park near where you live?
Helpful hints for using the ICE tool:
- Select the bands you wish to use from the pop-down menus above
and then click “build.”
- ICE displays the bands you chose in the first two thumbnails, and
the sum of those two bands added together in the third thumbnail.
- Use the pop-down menus under the thumbnails to set up the desired math
operation—the available options are add, subtract, multiply or divide.
- Use the numbers in the fields under each thumbnail as multipliers.
The default values are 1.0, which means “1.0 multiplied by the pixel values shown in the
thumbnail.”
- Once the multipliers and math operations are set as desired, click
“compute.” The result of your operation will be displayed in the large
window.
- Using ICE to calculate NDVI is a two-step process. Once bands 4 and 3 are shown
in the first two thumbnails, and their sum is shown in the third, go back and change the math
function from “add” to “subtract” between first two thumbnails, and then change
“result” to “divide” between the second two thumbnails. Clicking the
“Compute” button performs the calculation and displays the resulting NDVI product in the large
window.
- To learn more about what you can do with ICE,
read the User’s Guide.
Exercise 1 | Exercise 3