Image data for this section is contained in the /DATA/MODS1TO5/MODULE04/IMAGES/BELIZE directory and should be copied over to your computer before beginning. The files you will need are:
An explanation of these images and their source follows. You may also find a directory within the /BELIZE directory, entitled /MOREBELL. This directory contains the different L-band polarization images which can be generated using Sigma0.
SEASATim.gif (261K GIF) - SEASAT radar image of the North-West corner of Belize, in Central America. The area is known as the Rio Bravo. The image was taken in 1978 and shows an area of about 100kmx100km. The white lines outline on the image are the borders with Guatemala, to the West (right side of image) and Yucatan, Mexico, to the North (top part of image). This area was settled by the Mayan civilization over a thousand years ago, and has Mayan ruins dotted all over it.
SEASATzm.gif (373K GIF) - A zoom-in of part of SEASATim.gif (261K GIF) showing more detail.
Mosaicim.gif (257K GIF) - radar image mosaic using AIRSAR data. P-band HH was colored red, L-band HV green and C-band VV blue to make this Red/Green/Blue overlay of AIRSAR data. The image was taken in 1990 and shows an area of about 50kmx50km. Comparing Mosaicim.gif with SEASATim.gif (261K GIF) you should be able to see two things immediately:
Lbandful.gif (1,260K GIF) - L-band HH radar image of part of the image seen in the mosaic. This image is upside down when compared with the Mosaicim.gif image but was collected at the same time (April 1990) by the NASA/JPL AIRSAR imaging radar system. The radar was flying parallel to the top of the image and looking down and out to the left (i.e. down the page).
Belizkey.gif (8K GIF) - key to the radar image shown in Lbandful.gif. This schematic diagram, drawn to the same scale as the radar image, shows the size of the area covered (12.3kmx12.6km) and identifies some features seen in the radar image. Most of the image frame contains broadleaf upland forest, or tropical rain forest. At the top left is a small triangular shaped patch of a type of stunted forest, known as Bajo, which grows on poorly drained clay soils. The Bajo is separated from the upland forest by an escarpment, which rises about 100 feet. The most interesting feature in the image is a ranch known as Gallon Jug in the center right of the image, which was settled in about 1985, and contains: farmland (for raising corn and cattle); bare soil fields (just recently plowed); an area of regrowth, where the soil was too poor to support agriculture and the rancher is allowing the forest to grow back again; a recently bulldozed clear-cut, with trees still lying scattered on the ground; and a small coffee plantation. The coffee plantation was an area in the upland forest where the understory and middle canopy vegetation had been cleared, leaving only large trees. Although some coffee bushes had been planted, the forest floor was generally open. Also shown in the key is a local village where the ranch hands live, and two small lakes: Laguna Verde (green lagoon) and Laguna Seca (dry lagoon). Following the path of a small stream is a patch of flooded forest and marshland. Several dirt roads or tracks are indicated on the key as well.
VegmpByt.gif (16K GIF) - vegetation map of the Gallon jug area generated automatically from the AIRSAR data, using all three frequencies and all the different polarizations.
Comparing the vegetation map with the key shown in Belizkey.gif you should be able to see the following:
Note, for example, that 63.3% of the image was classified as forest, and 6.1% classified as Forest (D).
GroundV1.gif (257K GIF) - digitized photograph of upland forest, taken from the ground, looking down from a small rise. Photograph shows rain forest canopy stretching for miles in the background, with trees and other foliage in the foreground.
GroundV2.gif (11K GIF) - digitized sketch map of a walk (or transect) through a stretch of tropical rain forest within the Gallon Jug image. The image. which was drawn by a botanist, shows trees and shrubs seen along the walk.
LBandRed.STK - LBand AIRSAR image data file containing all polarizations. To display this image, run the Sigma0 application, select the LBandRed.STK file and then select the Open Cmprsd Stokes option under the File menu. Select the HV polarization for display.
Now point to a patch of upland forest with the mouse. The HV backscatter, seen in the bottom left hand corner of the image window, should be in the region of -14.0 dB, give or take a few dB. Pointing to a patch of farmland, the backscatter value will change to around -23.0 dB. Thus we can tell the difference between forest and farmland by looking at the L-Band HV backscatter.
Next, draw a rectangle on the image within the upland forest by holding down the mouse while dragging, then release. Choose the Mean and Std Dev option under the Statistics menu and a set of numbers will appear in the Statistics window. The mean values corresponding to the selected rectangle under the Mean column should match closely the following numbers:
| HH | -9.0 dB |
| HV | -14.0 dB |
| VV | -9.0 dB |
| HHVV* phase | 0.0 deg |
| Corr coef | 0.29 |
These values correspond closely to the model for scattering from a vegetation layer, since HH and VV are equal, HV is 5.0 dB less than both, the phase difference between HH and VV is 0 degrees, and the HH-VV correlation coefficient is close to one-third.
Close the HV image, and then display the HH image by using the Open Cmprsd Stokes option again and selecting the LBandRed.STK file. Notice that the coffee plantation, the clear-cut and the flooded area are now significantly brighter than the surrounding upland forest. Pointing the mouse at any of these standout areas should reveal an HH backscatter of around -7 dB, 2 dB higher than the upland forest. Now select a rectangle around the coffee plantation and execute the Zoom In option under the View menu five times. Now use the Display Sigma Zeros option under the View menu to write the backscatter values in dB over each pixel. You should see that the backscatter values in the coffee plantation are indeed a couple of dB higher than the nearby upland forest.
| Teacher's Guide - Table of Contents |
Converted to the IBM-PC by Al Wong, sirced03@southport.jpl.nasa.gov
Jet Propulsion Laboratory
4800 Oak Grove Drive
Pasadena, CA 91109