3.1 Becoming familiar with AATSR GBTR data

Many satellite images come in hierarchically structured formats that may include considerable amounts of metadata and ancillary data sets in addition to the satellite measurements. Metadata tags or attributes are used to describe the contents of a data file in such a way that the data is self-describing i.e., no other additional information should be required to unpack and use the data held within the file.

N1 metadata

In the case of an ENVISAT N1 data file, extensive metadata information is provided that describes various aspects of the data file and the data held within it For example, the filename and the time of data acquisition is stored in ASCII text format in the Main Product Header (MPH) section of the metadata. Here we will investigate some of the metadata supplied with AATSR TOA images.

Activity/Question 1

Open the data file ATS_TOA_1COLRA20020727_211146_-_000000812008_00071_02129_1339.N1 A new window will open which contains the N1 hierarchical file structure. The satellite image data with associated metadata, pixel flag coding and tie point grids ( giving the geographical location of the image pixels) are listed as a series of nested folders in the left frame of the file structure window. The Metadata folder is open and its constituent folders are shown in the right pane. To examine the metadata folders you can

Some metadata are text documents (displayed as Bilko Formula documents); some are tables (displayed as Bilko Table documents). Use the information in the MPH folder and the sidebar information on the Envisat orbit and filenames, to help you answer the questions below.

a)	At what time exactly was the first and last scan lines of the image data set acquired by the AATSR sensor?
b)	Is this scene obtained during an ascending or descending orbit?

AATSR bands

In the left frame select the bands folder (highlight it by clicking once). This will display all the image datasets as folders in the right frame, giving the numerical format (data type) and image size (width x height) for each image.

• Folders named btemp_{view}_{band} or reflec_{view}_{band} are 16-bit integer data sets containing the forward view and nadir view image data. For example, btemp_nadir_1100 is the 11 µm nadir brightness temperature (Tb) image data set.

• 2 sets of 16-bit unsigned bitfields contain data flags for each pixel provided separately for the nadir and forward view images.

• The confidence flags (confid_flags_nadir or confid_flasg_fward) contain information that can be used to assess the quality of each pixel value in the data set.

• The cloud flags, cloud_flags_nadir and cloud_flags_fward, contain information about the cloud clearing tests that have been applied to each pixel value. We will use the flags later in this lesson.

As you can see from the description in the right frame, the btemp_nadir_1100 image is a 16 bit signed integer image 512 pixels wide (X) and 544 scan-lines long (Y). It has 5 metadata attributes. Right-click on the folder icon and select Open Properties from the pop-up menu to display this metadata information.

Question 2. Right-click on the folder icon and select Open Properties from the pop-up menu to display the metadata information for the btemp_nadir_1100 band. You will need this to answer some of the questions below. To calculate true brightness temperatures from the image a scaling factor (b) and an offset (a) have to be applied to the 16-bit values (DN), using a formula that follows the standard convention Geophysical Value = a + b * datavalue.
a)	What are the values of the scaling factor and offset in this image?
b)	The measurements are accurate to about 0.1 K at best. How would you create a simplified formula to calculate temperatures in degrees C from the current data values? (Use @1 to represent the input image with datavalues, and @2 to represent an output image with the temperature values. 0 °C = 273.15 K). Note: @ is how Bilko and other image processing software refers an image plane, when carrying out calculations. @1 would be image plane 1.
c)	As seawater freezes at about -1.9 °C, we can safely assume that values cooler or equal to -2 °C are not true SST values - that is they may be treated as invalid data (null values). What is the image integer value corresponding to this temperature?
d)	Climatology tells us that in July, SST values below 13 °C are unlikely in the Mediterranean. What is the integer value corresponding to this minimum SST value?

Displaying the band data sets

To open and display the 11µm brightness temperature data set you simply

1. double click on the btemp_nadir_1100 folder in the right frame;
2. accept the default selection in the Extract Box, and
3. set the Redisplay dialog as follows:

   select a null value range that includes all values of less than or equal to -2 °C (<= -2 °C), using your answer to question 2c to guide you. (NB! you must set the null value before you click the null-value check box), in the stretch settings area set the minimum to 13 °C (see your answer to question 2d), and click OK to accept the Redisplay settings (figure).

Open the other AATSR frame (ATS_TOA_1COLRA20020727_211258_000000812008_00071_02129_1339.N1), and display the data using the same null value range (<= -2 °C) and minimum stretch value (13 °C) as for the previous frame.

Note: You will probably have noticed in the Redisplay dialog that the second image has a higher maximum value. To display both images with the same stretch you also need to set the Max value used for the display stretch. Open the Redisplay dialog again by right-clicking on each image, and change the Max value to a suitable number, for example 30000.

Question 3.
a)	Bearing in mind the orbit type and the time of acquisition which of these scenes shows the North African coast, and which the coast of Europe?
b)	Is there an overlap between the two standard scenes, and if so, can you give a rough idea of how many lines are involved? (The AATSR scan cycle is repeated 6.6 times per second, and the satellite moves forward by 1km (1 pixel) per scan cycle).

Because it was obtained on an ascending orbit, these images are actually 'upside down' - i.e. North is towards the bottom of the image. It is much easier to work with the data when the orientation is set so that the northern part of the image is at the top of the image display. To do this,

1. open the Bilko View menu and
2. select Reflect > in Both .

This will reorient the image so that north is up and west is to the left.

Question / Activity 4

Use the map in figure 6 as a guide and familiarise yourself with the image location and its content. You will find it easier to do this if you enhance the contrast within the display using the Bilko manual stretch:

1. Open the Stretch menu and select Manual.
2. In the Stretch window, double-click on the diagonal line to add an anchor point.
3. Drag the anchor point down to the bottom of the grame at a suitable position along the x-axis, for example 050, 000 (you can read the original (050) and new (000) display values on the Bilko status bar).
4. Double-click to add a second anchor point, and drag this up to a suitable position at the top of the frame, for example 210,255.