search

orcoastalmgmt / orSeaSketch

Repo for tracking issues with Oregon SeaSketch
1 stars 0 forks source link

Marine Mammal Haulouts Geometry Problems (DF) #29

Open mmosesOCMP opened 4 years ago

mmosesOCMP commented 4 years ago

Marine mammal haulout data – The haulout data are spatially depicted as lines that represents a stretch of coastline or a group of offshore rocks. If the proposed area polygon does not touch the line, the attribute will show zero marine mammal use, even though the polygon may actually contain the haulout sites. Not sure if there is an easy solution to this.

tchaddad commented 4 years ago

One approach might be to buffer the lines somewhat to create areas that are larger and easier to hit with a report. We would also want to change cartography.

Open to other suggestions on this one... Andy do you have any ideas? @DrewLanier

DrewLanier commented 4 years ago

I'm having trouble envisioning the scenario referred to above not returning any values. I'm wondering if the queries are using an overlay, or intersect geoprocess to evaluate the data. Certainly there should be values returned if a haulout falls entirely within an area drawn on the map, it should identify at least the presence/absence of a haulout. I'm wondering if we need a two step query, or if that's even possible.

DrewLanier commented 4 years ago

While all that I indicated above is true, I think the reports that are returned are hard to understand, and should be restructured.
For example, see my area drawn and the reports, along with my discussion post in the staff forum. https://www.seasketch.org/#projecthomepage/5c1001699112e049f68fc839/forum/5c100de39112e049f68fc8a9/topic/5ee157a2495fe47f9229d3ee

tchaddad commented 4 years ago

I think the first post from Moses had the right idea - this data is ODFW data and they have historically coded all their data to a different shoreline. So that is what is causing the underlying problem...

I don't disagree about the report, but that is probably a separate issue to this geometry problem.

DrewLanier commented 4 years ago

Understood. I conducted a quick and dirty assessment of the distance between the mammal haulout lines and the MHW shoreline, and it looks like a distance of 150m buffer would get most of the identified haulouts. There were a couple that were ~ 1000m away, but I believe it's okay to let those remain as they are, given they are likely offshore rocks or islands and not physically connected to a shoreline.

I'm not sure about changing the cartography or visual representation of the data in the SeaSketch data layers catalog. I think it might be enough to have them processed in the queries, without changing the representation in the tool.

DrewLanier commented 4 years ago

Suggestion: Instead of changing the source data by buffering each of the 7-8 data layers related to haulouts, could we have the geoprocessing query buffer the sketched area by the required distance? It would be adding a geoprocessing step for the query, but it would be a single change rather than having to manage the additional data layers.

tchaddad commented 4 years ago

I wonder if Dan can use the unclipped polygon for this query?

tchaddad commented 3 years ago

I'm leaning towards buffering these layers as that would make this a simple data layer substitution fix, rather than something that involves a code change that could have unintended effects elsewhere.

If all we do is substitute the buffers on the GP back end, then the front end can remain the same.