topic Re: Least-cost paths between n=1471 populations in ArcGIS Spatial Analyst Questions

Least-cost paths between n=1471 populations

DataOfficer — Fri, 03 Mar 2017 15:27:29 GMT

I am trying to calculate and map least-cost paths connecting every single population of great crested newt that has been observed in a study area in North East Wales across a cost-weighted distance surface.

I have been trying to do this using Linkage Mapper 1.1.0, a set of Python scripts that automates the mapping of wildlife corridors (Linkage Mapper - CIRCUITSCAPE.ORG ). Unfortunately, I keep getting error messages that I am unable to make sense of, and the program's developer has informed me these are likely to be bugs that cannot currently be fixed.

With that in mind, I am trying to find an alternative approach. I am aware of the suite of Distance tools that are part of the Spatial Analyst extension in ArcGIS 10.3. What I don't know is how to use the Cost Path tool to automatically construct least-cost paths between a large number of data points (1471 species records in my case), rather than a single source and destination point. Ultimately, I want to end up with a map like the attached image. What is the best way to go about this? I have never used ModelBuilder but have a feeling this might be the solution? Is anyone aware of any other tools for wildlife corridor mapping that are freely available?

Thank you all!

Arne

Re: Least-cost paths between n=1471 populations

ChrisDonohue__GISP — Fri, 03 Mar 2017 15:38:15 GMT

Here's a similar recent GeoNet post that may may be of help:

https://community.esri.com/thread/188140-home-range-tool-or-program

Chris Donohue, GISP

Re: Least-cost paths between n=1471 populations

DataOfficer — Fri, 03 Mar 2017 15:40:46 GMT

Hi Chris,

Thanks for your reply. Did you mean to attach a link to your post? I can't see any 😕

Re: Least-cost paths between n=1471 populations

ChrisDonohue__GISP — Fri, 03 Mar 2017 15:49:53 GMT

It's linked - click on the text "home range tool or program". At least, it works for me. Does the link go anywhere for you? If it only goes to the GeoNet homepage and not to the specific post, that apparently is a JIVE issue (the software that runs GeoNet).

Chris Donohue, GISP

Re: Least-cost paths between n=1471 populations

ChrisDonohue__GISP — Fri, 03 Mar 2017 15:54:52 GMT

When the going gets tough, the tough go to raster:

Chris Donohue, GISP

Re: Least-cost paths between n=1471 populations

DataOfficer — Fri, 03 Mar 2017 15:58:16 GMT

Ah yes that's working now, thank you. I have actually already downloaded Conefor 2.6, one of the tools linked to in the post you mentioned. The reason I am running Linkage Mapper is a) to be able to produce a nice looking map of least-cost paths and b) to get a text file that quantifies cost-weighted distance between populations so I can use this as the distance file in Conefor (rather than simply using Euclidian distance).

Is there perhaps another, simpler way of getting a text file/CSV for cost-weighted distances between points across a resistance surface?

Re: Least-cost paths between n=1471 populations

ChrisDonohue__GISP — Fri, 03 Mar 2017 17:48:09 GMT

I've not done many Least Cost Path runs, so am not sure. However, there are some possibilities out there:

As you mentioned, one possibility is to use Modelbuilder to batch process all the individual runs. One would employ an Iterator in Modelbuilder to loop through all the populations by Least Cost Paths. You'd have to play with Modelbuilder a bit to figure it out - it's not hard to learn, but does have its quirks.

ArcGIS Help (10.2, 10.2.1, and 10.2.2)

If you are skilled at programming, another option is to use Python to do the batching. This is a more elegant and efficient way to do it, but requires more skills.

ArcGIS Help (10.2, 10.2.1, and 10.2.2)

Dan Patterson

Chris Donohue, GISP

Re: Least-cost paths between n=1471 populations

DanPatterson_Retired — Fri, 03 Mar 2017 17:55:56 GMT

Chris, I am still stuck on the need to run 1471 individual cases. One would need to know more about the 'from' locations and the 'to' locations and the intervening landscape to see if the problem could be generalized more.

Re: Least-cost paths between n=1471 populations

MervynLotter — Fri, 03 Mar 2017 17:59:54 GMT

You should also consider Circuitscape (http://www.circuitscape.org/) for

modelling multiple paths (this does work well but be careful of cell size

as it takes an inordinate amount of time to run and requires LOTS of RAM).

Then if you are using Pro 1.3 and above, or ArcMap 10.5 (possibly 10.4),

consider using the new "Cost Connectivity" tool that can connect up

multiple patches using corridors and a cost surface. Not sure of the

limits.

See

https://pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/cost-connectivity.htm

Re: Least-cost paths between n=1471 populations

DataOfficer — Mon, 03 Apr 2017 15:14:20 GMT

Thanks Mervyn, I ended up using LinkageMapper after all (which is part of the Circuitscape project as far as I'm aware). I wasn't aware of the new Cost Connectivity tool, this sounds incredibly useful for the type of work that I do. I will make sure to update our organisation's version of ArcMap (currently still on 10.3.1).