Clarified impact of intensity rating

[claire-simpson]

Added a note to clarify that the intensity rating that a user assigns to a stressor will only impact the footprint of the stressor and will not affect the buffer area around that stressor

Fixes #150

[claire-simpson]

@phargogh There are multiple places that this note could be added so I'd appreciate feedback on whether this is the best location!

[phargogh]

Yep, I totally agree that there are multiple places where this information could be displayed (a downside of having a too-complicated chapter), and might I suggest we move this note to the spatially explicit criteria heading?

I also had a suggestion about the nature of the note, since I think (but could be wrong about this!) that a scalar rating does apply throughout the whole buffer zone, while a spatially explicit rating covers only the area it covers when rasterized. If this is true, moving the comment to the spatially explicit criteria area and also clarifying the coverage the rating layer must have might help clear up the original confusion.

Curious what you think, and I may well be wrong about the above!

[phargogh]

I agree that a .. note:: is a great way to present this information!

On the content of the suggestion, though, I'm pretty sure a scalar rating does impact the area impacted, including through the buffered area. The confusion for the user was that they provided spatially explicit criteria scores, but only provided them for the geometry of the stressor, not for the whole area that the stressor affected.

So I think I have 2 suggestions here:

1. Might I suggest we move the .. note:: to the section below titled "Using Spatially Explicit Criteria"?
2. I might update the text of the message to clarify what the user needs to provide, maybe something like the below?

When providing spatially explicit criteria layers, each layer must have pixel values or features that, when rasterized to the output resolution, cover the entire area affected by the stressor.

[claire-simpson]

Great points! I don't think I fully realized that the rating field was applied differently to the buffer zone depending on whether the rating was constant or spatially explicit.

1. Might I suggest we move the .. note:: to the section below titled "Using Spatially Explicit Criteria"?

I agree that adding the note in the spatially explicit criteria section is a better option.

When providing spatially explicit criteria layers, each layer must have pixel values or features that, when rasterized, cover the entire area affected by the stressor.

I am slightly concerned that this revised comment isn't explicit enough about how the buffer is utilized. What about combining the notes to have it be:

When providing spatially explicit criteria layers, each layer must have pixel values or features that, when rasterized, cover the entire area affected by the stressor. The spatially explicit intensity rating will not extend through the buffer zone that may be defined for that stressor.

[phargogh]

the rating field was applied differently to the buffer zone depending on whether the rating was constant or spatially explicit

The spatially explicit intensity rating will not extend through the buffer zone that may be defined for that stressor.

My understanding is that the rating calculation uses pixel values that represent criteria scores, where the pixel values are either from:

1. The buffered, decayed stressor. If a buffer is provided, then the criteria criteria score applies through that buffer, modified by a decay equation if selected.
2. A spatially explicit criteria score. If a user provides a spatially explicit criteria score, then we just use the raster version of the criteria layer and buffering isn't a part of the criteria score at all. The implication of this is that if there's no score for part of the area that is buffered, then that will have unexpected results as this user experienced.

So for this part of the note:

The spatially explicit intensity rating will not extend through the buffer zone that may be defined for that stressor.

I'm pretty sure the spatially explicit rating will cover whatever areas the user has it cover ... which might mean that it extends throughout the buffer zone! I don't think we can say that in all cases the intensity rating will not extend through the buffer zone, because it depends on what the user provides.

[claire-simpson]

I'm pretty sure the spatially explicit rating will cover whatever areas the user has it cover ... which might mean that it extends throughout the buffer zone! I don't think we can say that in all cases the intensity rating will not extend through the buffer zone, because it depends on what the user provides.

I guess my desire is to clarify that a user who adds a criterion with spatially explicit rating should not expect that rating to automatically extend through the buffer area. I.e., the user would need to manually buffer that area and add this buffered file into the Criteria Scores table. To me its not necessarily intuitive that a rating would be automatically extended throughout the buffer area if the user provides a constant rating in the Criteria Scores table, but the rating would not extend through the buffer area if the user provides a spatially explicit criterion for rating. Am I interpreting this correctly?

[phargogh]

Yes, I believe you are interpreting this correctly! The assumptions that you describe are also really helpful, so thank you for clarifying where you have been coming from.

What if we made 2 distinct edits, then?

1. We could update the body of the model (under Step 1 of Cumulative Risk to Habitats or Species from Multiple Stressors) to better describe how scalar ratings behave (as a constant value) vs how spatially explicit ratings behave (which we really don't yet do very well, if at all)
2. We could update this section under spatially explicit criteria to reflect how spatially explicit criteria should be formatted, that there should be data for each pixel throughout the buffer zone, and that if there are missing data for a pixel within the buffer zone, that the criterion will be ignored on that pixel (it's effectively a criterion of 0)

Maybe that would better clarify things?

[claire-simpson]

I tried finding a place in the Cumulative Risk section but it was a bit clunky as the spatially explicit aspect and buffer option arent really explained here. No matter where I added this note it felt a bit awkward, so decided to contain it to the sections about spatially explicit criteria, but let me know what you think!

[phargogh]

I think this looks good! Thanks for finding where to add this info. I agree, it's a little awkward, and it doesn't help that this is the longest chapter in the UG, despite the relative simplicity of the model. Anyways, I'll approve and merge. Thanks!

[claire-simpson]

@phargogh Is it possible to reopen this PR? I had a realization about my changes, and I think I actually was a bit misleading in that the buffer areas can still be affected by a stressor, however the effect of the stressor will not be influenced by the spatially explicit rating unless the rating is explicitly specified throughout the buffer zone.. To that end, I am inclined to delete the last sentence I added in each of the two sections. Thoughts?

[phargogh]

I think maybe since you had created the PR, maybe you'd be able to reopen it? If not, you could always make the change on the same branch and open a new PR on the same branch.

Personally, I think the two sentences are probably fine as-is, but please feel free to edit/update them for clarity!