Combine low level moisture with a cloud mask #3022
Conversation
Combine ESSL colorized low level moisture with a cloud mask. Work in progress.
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #3022 +/- ##
=======================================
Coverage 96.11% 96.11%
=======================================
Files 383 383
Lines 55673 55673
=======================================
Hits 53511 53511
Misses 2162 2162
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
Pull Request Test Coverage Report for Build 12950493562Details
💛 - Coveralls |
If I don't give its own enhancement to the masked colorised llm, either the masked or the unmasked version doesn't find its enhancement.
gerritholl
requested review from
djhoese,
mraspaud,
adybbroe and
pnuu
as code owners
|
Great idea! a couple of questions:
|
fixing one merge conflict
The cloud mask claims a sharp edge there between cloud-free and cloudy:
The cloud mask stops.
Maybe a resampling artefact. Redoing the resampling with gradient search, the result looks different, but still has artefacts:
Oh, gradient search is of course not great for a cloud mask. |
|
Resampling with nearest neighbour again, but with a larger radius of influence, the result looks better:
|
|
With the modifier
Version filled with IR 105 instead (using
|
|
Nice imagery! do you have the date and time for this slot? |
2024-12-17 12:10 |
|
Case from today (2025-01-80 15:30), with the night part blended out using the
|
|
Nice, happy to see the L2 data cloud mask in action with users :) Regarding the sharp edge over the Baltic sea, that's caused by a known bug in one of the spectral tests. A fix for this will come with the next version of the processing facility in a few weeks. As you probably know. the cloud mask is also "only" pre-operational and we are still tuning it. It seems like we are also picking up some false alarms in the Baltic sea and we also have some coastline artefacts. Wrt. the higher viewing angles, also keep in mind that the L2 products are not generated for the full disk, but up to about 75 degrees viewing zenith angle. |
For the create reference script, add the ability to use multiple readers.
Improve behave tests. Allow to have multiple readers. When clipping, use the configuration option, because reader_kwargs with multiple readers is less practical / does not work very well.
|
Starting to clone and test the repository pytroll/satpy |
|
The testing process was executed successfully. See the test results for this pull request here! |
|
Starting to clone and test the repository pytroll/satpy |
|
The testing process was executed successfully. See the test results for this pull request here! |
|
Starting to clone and test the repository pytroll/satpy |
|
The testing process was executed successfully. See the test results for this pull request here! |
Add calibration to the fake reader channels matching 0.86 and 0.91, so that the workaround added in 6a7a78a does not change the total number of expected datasets in test_load.py
Use nearest neighbour when combining with cloud mask
|
Starting to clone and test the repository pytroll/satpy |
|
The testing process was executed successfully. See the test results for this pull request here! |
The area and the resampler were swapped for one case in the feature definition. Fix this.
|
Starting to clone and test the repository pytroll/satpy |
|
The testing process was executed successfully. See the test results for this pull request here! |
|
Starting to clone and test the repository pytroll/satpy |
|
The testing process was executed successfully. See the test results for this pull request here! |
Document the composites combining low level moisture with a cloud mask, and add credit to ESSL for the colours.
Combine ESSL colorized low level moisture with a cloud mask.
Example for eurol: