Richard Comments May 19, 2024

[rfernand387]

Comments on https://ronnyhdez.github.io/reclaimed_sites_ab/abandoned_wells.html as of May 19, 2024.

Acceptable sites

• [x] 1. Time: You can sample all sites that have been abandoned and no longer in production irrespective of date. I would also filter out wells with abandoned<max production date or reclamation date<abandoned date. There are issues:
• [x] 1.1 What is difference between last production date and abandoned date. Should we only consider wells with abandoned date?

No. We can data mine time between abandoned date and subsequent production date as well.

• [x] 1.2 Abandoned date is sometimes zero, does this means it is still in production?

SUBIR

• [x] 3. Area: We need at least 90m x 90m square footprint to get 1 landsat pixel. This is a gis filter but will reduces number of sits considered.

YES, This is OK.

• [x] 4. We would process based on footprint id and not well footprint id since we care abput each patch. We can synthsize results later based on well footprint id.

Yes.

• [x] 5. There are 89915 observations with a year reclamation date value of 0. How many are abandoned>max production date?

What is the Abandoned date. According to subir an abandoned site with a reclamation certificate was reclaimed. We can still analysze these sites.

ASK SUBIR where the record of dates of RECLAMAtIOn certificate issue is.

[ronnyhdez]

1.1 What is difference between last production date and abandoned date. Should we only consider wells with abandoned date?:

• there are 779 polygons where the production date is more recent than the abandoned date, 55022 where the abandoned date is more recent than the production date, and 7435 with both dates being the same.
• When we have reclamation status as reclaimed, the number of abandoned dates with zero goes pretty low:

I think we can leave out the polygons without abandoned dates out. That would be 543 observations.

[ronnyhdez]

1.2 Abandoned date is sometimes zero, does this means it is still in production?

There is one variable feature_ty which definition is: "The type or category of human footprint feature". In this dataset every observation has the category: "WELL-ABAND", so all the polygons should be abandoned.

When we have 0 in the abandoned date, we have:

• 4687 observation with max last production date in 0.
• The maximum date value in the range is 2022 (This is the maximum date value found for every date variable)

And then, the distribution of the last production date observations that are not 0:

[ronnyhdez]

3. Area: We need at least 90m x 90m square footprint to get 1 landsat pixel. This is a gis filter but will reduces number of sits considered.

• In ref #23

[ronnyhdez]

5. There are 89915 observations with a year reclamation date value of 0. How many are abandoned>max production date?

779 polygons where the production date is more recent than the abandoned date (when not filtering by 0 in reclamation date:

1 less    779
2 more  55022
3 same   7435

With reclamation date not equal to zero:

1 less    155
2 more  16900
3 same   2883

Filtering just by observation with reclamation date == 0:

1 less    624
2 more  38122
3 same   4552

In all of the 3 cases, I'm excluding the observations where there is a 0 in the abandoned date or the last production date

[ronnyhdez]

Changes in ref #22