Add tm5_constant_a and tm5_constant_b for tropomi_l2

[zxdawn]

Because of different shape, tm5_constant_a and tm5_constant_b are ignored by the tropomi_l2.py reader.
I added them in tropomi_l2.yaml now.

[djhoese]

CC @tommyjasmin for opinions.

[coveralls]

Coverage decreased (-0.003%) to 89.532% when pulling e83e49c on zxdawn:tropomi into d12bb7a on pytroll:master.

[codecov]

Codecov Report

Merging #1128 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1128   +/-   ##
=======================================
  Coverage   89.53%   89.53%           
=======================================
  Files         200      200           
  Lines       29356    29356           
=======================================
  Hits        26284    26284           
  Misses       3072     3072

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d12bb7a...e83e49c. Read the comment docs.

[mraspaud]

LGTM

[ghost]

Congratulations 🎉. DeepCode analyzed your code in 0.007 seconds and we found no issues. Enjoy a moment of no bugs ☀️.

👉 View analysis in DeepCode’s Dashboard

---

[tommyjasmin]

Hi all - @zxdawn Xin - can you explain the use of these constants and which products they are expected to be present in? In the current YAML, we reference them directly under the PRODUCT group:

'PRODUCT/tm5_constant_a'

This appears to be correct for NO2 data, which I presume you are working with.

But with a very brief look, I do not see these constants in the CO sample data I have. I do see these constants in my SO sample data. However, for those files they are present in a support data subgroup:

'PRODUCT/SUPPORT_DATA/INPUT_DATA/tm5_constant_a'

So we can't assume these constants are always present in the top level PRODUCT group.

[zxdawn]

@tommyjasmin Hi Tommy, Yes, I tested it for NO2 files.

The constants are used to calculate the TM5 model pressure levels which are used in the NO2 retrieval, like this:

P = A + B*PSFC

I will fix this and test all kinds of tropomi products soon. Thanks.

[zxdawn]

@tommyjasmin I've tested the newest YAML and it works well for NO2, SO2, O3 and CO data. Because the var_name in YAML is tested whether exists in the file:

satpy/satpy/readers/tropomi_l2.py

Line 98 in 2ce07c0

if (matches and var_name in self) or (assembled and bounds_exist):

So, is it OK to just keep these two variables in the YAML? @mraspaud @djhoese