L2C computation

[Lucile Gaultier]

Computing L2C from L2B is now possible using

Code:

skimul2c [yourparameterfile]

One method has been implemented, there will be more options in the future.
So far, two parameters can be indicated in the parameter file: resol and posting
By default:
 

Code:

resol = 40 # in km
posting = 5 # in km

[Lucile Gaultier]

An new option has been added to the parameter file.
It is possible to remove across track current at the center of the swath as there is a high uncertainty in this area:

Code:

# Mask data at less than ac_threshold km from the nadir
ac_threshold = 20 # in km

By default, ac_threshold = 20

[Lucile Gaultier]

A new option has been added to the parameter file.
You can add to the name of the l2c file a specific attribute to refer your test case

Code:

# config_l2c:
config_l2c = 'my_test'

[josborne]

Values near the nadir are masked because the radials are mostly across track. Near the outer edges of the swath, radials are mostly along track. Presumably, observations near edge suffer uncertainty similar to the near-nadir values. . Is guidance available for what across-track distance to mask (e.g., mask values for x_ac > 140)?

Thanks,

John O

[Lucile Gaultier]

Indeed, there will be also indefinite estimation of the 2d velocity near the edge of the swath, though the hole is not as big as around the nadir. Using neighbours, a computation can be made at the edge but with an error that can be important if the area is very dynamic. You can compute the error of 2d reconstruction by comparing the true and the reconstructed velocity without error. I usually remove 10km at the edge.

[josborne]

Thanks!