Different sources and "targets"

[jessjaco]

How possible would it be to implement the ability to have different source and "target" layers?

[vlandau]

You'd have to hack the code a bit, but it would be possible. First, you would need to add a new argument (raster input) for the targets, and feed that into the get_targets and get_ground function. One problem is that the total current injected for each moving window solve is equal to the source strength of the target pixel (see here), so you'd also have to hack get_source as well, so that it instead just uses the raw values of the source strength layer. I think you would also have to set any source pixels that fall within the target block to 0. There's probably a few things I'm missing, but I think those are the major things.

[vlandau]

Actually on second look, you wouldn't have to adjust get_source at all, as long as you're okay with the total current being equal to the strength of target pixels (see the strength argument to get_source).

[cbogdanc]

Hello,
Is anybody working on this issue? It doesn't look like this is possible yet in the latest version of Omniscape.

This would be very valuable for ecological studies where directed movement/connectivity is important, such as from one habitat type to a different habitat type. In a way similar to how pairwise mode works in Circuitscape, with a list of included pairs of source to ground nodes.
Thanks!

[vlandau]

@cbogdanc this project is currently in maintenance mode and as such there is not really any focus on development of new features, unfortunately. That being said, pull requests are always welcome!

[cbogdanc]

I'm not a Julia programmer, so I wouldn't know how to implement this as a new feature.

Is it possible to simulate separating the source pixels from the target pixels using the conditional connectivity options? If so, how would you do that?

Could you please provide some descriptive examples of using the conditional connectivity parameters, to better understand how they work (and if they would be suitable for the source - target separation)? The documentation for conditional parameters is not very clear for me.
Thanks!

[jessjaco]

I can't remember why I started this issue, but have you considered just using Circuitscape?

[emench]

@vlandau I am also very interested in this. Instead of supplying separate source and target strength rasters, could it instead be simpler/ more feasible to supply a single source strength raster with positive and negative values and have pixels only be considered a target if they have a negative value and only be considered a source when they have a positive value within each window?

This technique could be pivotal for my PhD work, so I have the impetus to learn Julia in order to apply it, but would appreciate any feedback on whether this might be feasible first!