GEOG 372: Introduction to Remote Sensing

Fall 2007

November 14^th and 15^th

Laboratory 11

Vegetation Indices for Disturbance and Burn Scar Mapping

No late labs can be accepted. Please type the answers just below the questions and hand the printouts at the beginning of the next lab: November 28th (0102) and Nov 29th (0101). Also do not forget to mention your MAJOR on your labs

The MODIS image we will be working with today is a subset of a much larger image. The area we are focusing on is in central Alaska, where fires in the boreal forests often cover a large extent and can be quite severe. For the purposes of this laboratory, we will be examining burn scars in this image and methods for extracting information about the damage caused by fire. We will be exploring how to use multi-channel (or band) data for image analysis and will use standard image manipulation techniques to see if we can derive improved burn area estimations.

Open ENVI 4.2 and go to File - Open Image File. Navigate to U:\\g372\372-fall07\lab11data. Select the file ak_modis_sub4 without the .hdr extension and click Open. Click on the RGB radio button and load band 7 as Red, band 2 as Green, and band 1 as Blue. Click Load RGB.

Look at the colors in the image and notice the band combination you used to load your RGB. Click on and drag the red boxes in the displays to move the location of the zoom windows.

1. What color is healthy vegetation in this image? Why?

2. Based on the purpose of this lab, what do you think the dark brownish/red features are?

3. Notice the bright bluish/white features in the lower left corner of the scroll image. What land cover feature is this?

4. What is the spatial resolution of each band you are viewing?

Go to Tools – Pixel Locator.

5. What is the dark feature located at 5716, 3403?

Some information about MODIS imagery:

Band	Bandwidth (as measured in micrometers)	Spatial Resolution
1 (Red)	.620-.670	250 m
2 (Near IR)	.841-.876	250 m
3	.459-.479	500 m
4	.545-.565	500 m
5	1.230-1.250	500 m
6	1.628-1.652	500 m
7 (Mid Wave IR)	2.105-2.155	500 m

The MODIS data set you are using was created by implementing a compositing technique to eliminate clouds and smoke. All MODIS passes between 28 August and 14 September 2004 were downloaded. There were 2 passes of MODIS data each day that were used (one in the morning and a second in the afternoon). An algorithm was used to select the pixel over the time period which contained the least cloud cover and smoke contamination.

A visual examination of an image can often provide us with valuable information concerning disturbance in a landscape. However, it does not provide a standardized product for evaluation or quantitative comparison. Now we will use automated techniques to pull out our features of interest. We can spectrally isolate the burn scars by using Vegetation Indices (VIs). You will experiment with three such VIs; a basic NIR/Red band ratio, the Normalized Difference Vegetation Index (NDVI), and the Normalized Burn Ratio (NBR). The NDVI and NBR are based on the same equation, however the NBR substitutes the MIR band (MODIS band 7) for the red band.

Step 1: create a simple vegetation index (VI) ratio.

VI = NIR/Red = Band 2/Band 1

To do this you will need to use the Band Math Tool which can be accessed from the Basic Tools menu. Enter the equation "float(b2)/ float(b1)" without the quotations under 'Enter an expression'. It is important to create a float output to avoid truncating the integer values. Click OK. With 'B1 undefined' highlighted, select band 1 from the list. Next, highlight 'B2 undefined' and select band 2 from the list. Make sure to check the radio button to output the results to memory and click OK.

When the VI band math image loads to Available Band List, select the VI band math band, click on display 1, new display, and load the image.

In the VI Band Math “band” display toolbar, click on Tools - Link - Link Displays - OK.

6. What 2 landscape features have the lowest index values (represented by the darkest shades of grey)?

7. What landscape feature has the highest index values (represented by the lightest shades of grey)?

Step 2: create an NDVI (normalized difference vegetation index) image of the area.

NDVI = (NIR-Red)/(NIR+Red) = (B2 - B1) / (B2 + B1)

Follow the same steps above, but this time substitute the equation "(float(b2) - float(b1))/(float(b2)+float(b1))" without the quotations. When the NDVI band math image loads to Available Band List, select the NDVI band math band, click on display 1, new display, and load the image.

In the NDVI Band Math “band” display toolbar, click on Tools - Link - Link Displays. Use the arrow button to change display 3 to yes and click OK.

8. Visually, how are the results of the NDVI different from the VI?

9. What Landsat band numbers would be used in this equation?

Step 3: next we will create an NBR (normalized burn ratio) image of the area

NBR = (NIR-MIR)/(NIR+MIR) = (B2 - B7) / (B2 + B7)

Follow the same steps above, but this time substitute the equation "(float(b2) - float(b7))/(float(b2)+float(b7))" without the quotations. Make sure you select the correct bands from the list this time since they have changed. When the NBR band math image loads to Available Band List, select the NBR band math band, click on display 1, new display, and load the image.

In the NBR Band Math “band” display toolbar, click on Tools - Link - Link Displays. Use the arrow button to change display 4 to yes and click OK.

10. Visually, how are the results of the NBR different from the NDVI and VI?

11. What Landsat band numbers would be used in this equation?

12. Which vegetation index works best to contrast burn scars and the surrounding landscape? Why?

13. What 2 portions of the electro-magnetic spectrum have the greatest difference for healthy vegetation? How do you know?

14. What 2 portions of the electro-magnetic spectrum have the greatest difference for burned vegetation? How do you know?

15. Some of the fires were still smoldering in this image. Why is there no smoke present?

Be sure to submit a screenshot showing all 4 linked displays with your lab. Arrange all your windows on the desktop so that they do not block each other. Press “Ctrl” and “Print Screen” on the keyboard at the same time. Go to Start – Programs – Accessories – Paint and open the Paint program. In Paint, go to Edit - Paste. You will see the screen shot appear in the Paint window. On the Paint tool bar go to File – Save As, navigate to your directory and give it a name, click “save”. Close the Paint program.

Copy your screenshots to a CD or USB jump drive, email them to yourself, or print them using your print account.

Log off your computer and turn off the monitor when you are finished.

Typed answers with screenshots are due November 28th (0102) and Nov 29th (0101), 2007