Link to the UNESCO-IOC website
Link to the Bilko website
Earth from space
Annual sea surface temperature

3.4 Using stretches to interpret data

Identifying water masses     Visual comparison of different images    

TUTORIAL

3. Stretches

to bottom of page next page

Images used:

Description of mer_20040201 _algal1.dat

Download mer_20040201 _algal1.dat
(83 K)

Description of ATS_TOA_1C_ 20040201~.N1

Download ATS_TOA_1C_ 20040201~.N1
(7.7 MB)

Useful information:

Numerical data formats

Null values and missing data

In this section you will use manual stretches to identify different water masses in the two images from the Benguela region which you have been working with in previous sections.

You should have to images open from your previous work:

1.

The btemp_nadir_1100 band of the image ATS_TOA_1COLRA20040201~.N1, loaded with a linear Redisplay stretch range of 28400 - 29350, a range that covers the water temperatures only, while assigning clouds to 0 and land to 255 in the display.

2.

The image 8-bit image mer_rr_20040201_082314_algal1.dat, currently minimized to an icon.

The amount of detail in the AATSR brightness temperature image after a Redisplay stretch optimised for the ocean is considerable. However, there may be even more structure to discover, by experimenting with a manual stretch. Examination of the image indicates that there are two main areas of water, one cooler (darker) area near land and to the North, and another warmer (brighter) area off-shore and to the South. A histogram taken of the whole image, with the Redisplay stretch applied confirms that there are two peaks in the frequency distribution.

Question 1

What is the display value of the boundary between these two peaks?

to bottom of page    

Identifying water masses

A bimodal frequency distribution usually suggests the presence of two water masses separated by a fairly distinct boundary. As we have seen earlier the boundary value is found in the gap between the two peaks. A stretch with a knee at this boundary value should therefore help to discriminate between them.

  1. Open a manual stretch document ( [ALT+S] > [M] ) and insert a knee at the x-position indicated by your answer to Question 1.
  2. Drag the knee upward to a y-position of 255 and observe how this stretch reveals greater detail in the cooler area.
  3. Now open another stretch document and insert a knee at the boundary x-position
  4. Drag this downward to a y-position of 0 and observe how this stretch reveals details in warmer area.

The streches each reveal greater detail in the temperature patterns, but in mutually exclusive areas.

Splitting the image window

If you want to compare these stretches to the original data loaded with the Redisplay stretch, you can do so by splitting the image window.

  1. From the menu bar select Windows > split ( [ALT+W] followed by [p] for split.)
  2. A cross appear in the centre of the image. Drag this down to the bottom centre of the image window to give a vertical split.
  3. A second image appears in the right frame. Right-click this frame and zoom to preserve shape.

Note how the stretched image now appears to the left, and a second image with no manual stretch appears to the left of the bar dividing the window. You can change the appearance of the right window to the other manual stretch:

  1. Activate the stretch document you want to apply
  2. Click on the knee to apply the stretch. split.

To remove the split simply drag the vertical bar to the side of the image window.

Comparing images

You can compare the thermal patterns in the AATSR image and the patterns of pigment concentration in the chlorophyll image, by using manual stretches in a similar way in both images.

  1. Maximise the image mer_rr_20040201_082314_algal1.dat
  2. Clear any stretches that may be left from previous sections.
  3. Take a histogram of the whole image, and use the cursor to find the boundary value between the peaks in this image.
Question 2
a)

How many distinct peaks are there in this image?

b)

What is the boundary value(s) between the main peaks?

c)

Where would you put the knees in order to create a manual stretch optimised for each of these peaks and the boundary zone between them?

Create manual stretches for the MERIS pigment image based on your answers to question 2c, apply each in turn and watch how they reveal details in mututally exlusive areas of the image.
 

Answers:
(Resizable
pop-ups)

Answer 1

Answer 2

Answer 3

Back up to:
Q1   Q2  

Activity / Question 3

Place the MERIS and AATSR image on top of each other and adjust their size until the two windows have identical dimensions. Move them to a position size by size, and right-click on each in turn to select Zoom > Preserve shape. This makes the two images easier to compare. Use the manual stretch documents for each image to help you answer the questions below.
 

a)

Which peak in the frequency distribution of algal pigment concentrations corresponds to the low-temperature region in the AATSR image?
 

b)

What is the relationship between algal pigment concentration and measured brightness temperature seen in these images?
 

c)

How would you explain the feature seen at an approximate geographic position of 16° 7' E, 32° 47' S ? Use the Go To dialog ( [CTRL+G]) to find this position in both image (see figure (5K) for how to enter the position). Adjust the stretch documents for both images until you find a way of displaying this feature to show the relationship between temperature and chlorophyll as clearly as possible.

to top of page next page

 
Next: Saving the display

  Link to ESA's Envisat website   Link to NOC's website