Senitel-2 Harmonized Set Collection Update?

[jlzamary]

I’ve been working on a project collecting imagery using the Sentinel-2 satellite for about a year now. One issue has been retrieving imagery from the 2017–2019 period. When calling dates prior to 2020, my functions would not run, either in my IDE or in the JavaScript online editor.

Recently, while working through one of my Jupyter notebooks, I decided to try collecting imagery from an earlier date range again, and it worked for the year 2019, but not 2018 or 2017. I was able to access 2019 data in my IDE, but not 2018 or 2017. However, online in the Javascript editor, I have been able to view images from the 2017-2018 range.

I’m not sure if there have been any updates related to this. Checking the Sentinel-2 catalog on Earth Engine, it appears that SR imagery should be available from March 2017 onward. I remember this being the case in the past as well, but I still had trouble accessing it.

I’m a bit stumped and wondering if the collection was updated without me realizing and why I still might be having trouble with data from 2017 and 2018.

[jdbcode]

wondering if the collection was updated without me realizing

Within the last year there have been no changes other than the rolling addition of new scans.

why I still might be having trouble with data from 2017 and 2018

If you can access the images in the Code Editor, you can access them from Python. The requests are all hitting the same backend and catalog. Something about your Python script might be incorrect. Please share minimal reproducible scripts that demonstrate the problem.

If the issue is unrelated to geemap please post it directly to the Earth Engine date team, using these instructions: https://developers.google.com/earth-engine/help#dataset_problem

[jlzamary]

Please share minimal reproducible scripts that demonstrate the problem.

Here are the two scripts I have. In the online code editor, an image for my ROI appears. I know my viz parameters for both are fine, as other dates load in alright. Thank you for your help!

Earth Engine JavaScript

var trueColorImage = Sentinel2
  .filterBounds(ROI)
  .filterDate('2017-06-01', '2017-06-25')
  .map(maskS2clouds)
  .median() 
  .clip(ROI);

Python in Jupyter Notebook

trueColor = (
  ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")
  .filterDate('2017-06-01', '2017-06-25') 
  .filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE', 30))
  .map(mask_s2_clouds)
  .median()
  .clip(ROI)
)

[jdbcode]

You don't provide an ROI - this is not reproducible.

In the Python version you are filtering for low cloud percent, but not in the JS script. It's likely that there are no images meeting your criteria in the Python script.

[jlzamary]

In the Python version you are filtering for low cloud percent, but not in the JS script. It's likely that there are no images meeting your criteria in the Python script.

I ran the code without filtering for low clouds and still no image.

You don't provide an ROI

My ROI is a .geojson file (not sure if that is easy to share here). The location is Krusenstern Lagoon in Alaska (should be able to search and draw a polygon around it).

[jdbcode]

Please see: https://developers.google.com/earth-engine/help#help_request_best_practices

People can't help if they can't reproduce the issue.

[jlzamary]

Sounds good, appreciate your help! I'll gather all the needed information.

Would you recommend posting on here, making a post on the GEE community forum, or the Google Earth Engine Developers Group?

[jdbcode]

Thanks for considering reproducibility. Since the topic is not directly related to geemap, the Google Earth Engine Developers Group is a good space to post your question.