Area estimation from multiple land cover product

[adebowaledaniel]

The src/compare_covermaps module already provided all necessary functions and classes to support area estimation from multiple land cover maps. However, there is no provision to select maps available for a year of interest, which is important for crop area estimation. So, I modified the module and updated the intercomparison notebook template to include the area estimates in the PR.

• Add year variable to the Covermap class. This is needed to select land cover maps available for the reference year and add year attributes to each map in the module.
• The major updates in the notebook are filtering maps with the year attribute, CEO reference sample agreement, and area estimation compute function.
  • Select maps ±2 years of the reference year.
• The number of pixels for the DynamicWorld dataset returns a None array due to EE computation timed out. [Temporarily fix] - I changed the None array to nan to avoid errors in the notebook.
• Run the area estimate for Kenya and Senegal
  • Include FAOSTAT - cropland area

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[ivanzvonkov]

I recommend moving this inside the maps/Kenya_2019 folder. I think its only used there and will be easier to find that way.

[adebowaledaniel]

Noted!

[ivanzvonkov]

Why is this necessary? Is it intended to be run from Google Colab?

Reply via ReviewNB

[adebowaledaniel]

Yes, Google Colab for now.
I'll add modifications for local later.

[hannah-rae]

@adebowaledaniel It seems like the goal of this PR is similar to my other PR #374. Have you reviewed that PR already? Can you comment on how the goals of this are different from that PR, or do you see this as superseding the other PR?

[adebowaledaniel]

@hannah-rae, The primary goal of this PR is to get the area estimates of a reference year from all the land cover products available for that year, including the ones ± 2 years. So, my focus was on the area estimation. I didn't notice #374 until I finished this draft. I suggest we complete and merge #374 first. Or integrate both PRs since this has an extra layer for area estimates.