Minimize Cloud Identification

Cloud identification processing is a series of tests aiming at identifying different types of cloud (or snow) present on output images acquired by SLSTR. These tests are mainly inherited from those applied on AATSR data, with new tests added mainly on the new visible channels.

The identification of cloud-affected pixels is accomplished by applying a series of tests in turn to the brightness temperature data in the 12, 11 and 3.7 micron channels, and to the reflectance data in the 1.6 micron channel, and in the visible channels. Present tests are derived from heritage of AATSR, hence new channels at 2.25 and 1.375 microns may permit the implementation of new tests. The pixel is flagged as cloudy if any one of the tests indicates the presence of cloud. The table below summarises the cloud clearing tests to be applied, including one test flagging snow-covered surfaces.

 

 

Tests name Views Day/night time Land/sea
Gross cloud test nadir and along-track views separately both both
Thin cirrus test nadir and along-track views separately both both
Medium/high level cloud test nadir and along-track views separately night only both
Fog/low stratus test nadir and along-track views separately night only both
11 micron spatial coherence test nadir and along-track views separately both both
1.6 micron histogram test nadir and along-track views separately day only sea only
11/12 micron nadir/along-track test both views both sea only
11/3.7 micron nadir/along-track test both views night only sea only
Visible channel cloud test nadir and along-track views separately day only land or both
Infra-red histogram test nadir and along-track views separately both sea only
2.25 micron histogram test nadir and along track views separately day only sea only
1.375 micron threshold test nadir and along-track views separately day only both
Snow-covered surface test nadir and along-track views separately day only both

Cloud identification test list

 

Some of the tests depend on results from the tests performed previously and hence the order in which they are applied is important.

The infra-red histogram test is applied after the other tests, and only uses those pixels that have not been flagged as cloudy by any of the preceding tests.

The 1.6 micron test operates only on pixels not previously flagged as cloudy by the gross cloud test or the thin cirrus and 11 micron spatial coherence tests, and must therefore follow these tests.

Each test makes use of a look-up table of parameters with which the brightness temperature or reflectance data is compared. Where tests are applied to along track and nadir view images separately, the parameters may be defined separately for the two cases. More generally, the comparison parameters may depend on the air mass in the line of sight, and this is implemented by allowing the tabular parameters to depend on the across track position.

Some tests are only applied under certain conditions (during day or night time only, over land or sea pixels only), as described in the table above.

Note that, in the frame of coming evolutions, two tasks of improvement of the cloud masking are planned:

  1. The thresholds used in the current cloud masking will be tested and tuned
  2. The Bayesian method will be implemented 
Logical Flow of Cloud Identification