Yes, this is correct.

Yes, that's correct. They were non-permanent islands, or most likely floating veg, that aren't there anymore.

Thank you for catching that! It's supposed to be start_date. I'll get those changed.

Absolutely! I just found out I should first reach out to the community before importing data like this. I'll be doing that shortly. These are Marsh terraces and some more information about them can be found at https://www.ducks.org/conservation/where-ducks-unlimited-works/gulf-coastal-prairie/marsh-terraces-as-a-conservation-practice

In short they are built in an attempt to capture sediment and increase vegetation growth to increase marsh accretion. We want to keep marsh there and reduce subsidence.

These are constructed by excavating the marsh soil and piling it up in the lines that you may see in the imagery. Once piled up crews will plant marsh specific vegetation on them.

We were adding the dataset with a year_of_construction tag to provide data to potential users for research purposes. Do these help marsh growth? How long does it take? What form factor seems to work best? Does juxtaposition within open water matter?

It's extremely difficult to classify open water or land in this area because wetlands change so rapidly. A change in the tide between two images could drastically change a users perspective on what is under water and what is not. This is why we decided on linear features instead of enclosed.

Thank you for looking and reviewing. I appreciate it.