How large of a landscape can I analyze with Circuitscape?
Short answer: 1-6 million cells per GB of RAM on your system if you are running 64-bit Windows. The size of grids that can be analyzed depends on how much RAM Circuitscape can address. We have solved grids with 100 million cells on 64-bit systems. However, 32-bit Windows and Mac operating systems limit the amount of RAM that Python can address, meaning that only landscapes in the neighborhood of 1-6 million cells can be solved on these systems, even when they have lots of RAM.
Users can generally coarsen their grids and get results that closely approximate those run at fine-scale resolution (see McRae et al. 2008). To solve the largest grids, we recommend that users find a 64-bit Windows or Linux system with lots of RAM. Installing on Windows is easy, just see the download page. Linux can be more difficult unless you are accustomed to working with Python and dependencies. See the user guide.
We are working to push these limits, through the use of more efficient algorithms and parallel computing. Please see the user guide for hints on getting the most out of the memory you have.
What OS is best for Circuitscape?
If you want to run large problems (> 4-6 million cells), the only option is 64-bit Windows or Linux - for Linux, we recommend the latest Ubuntu release. For smaller problems, Circuitscape 32-bit Mac and Windows releases will work fine.
What kind of computer is best for Circuitscape?
Basically, spend all your money on RAM. Today, for about $10,000, you can get a computer with > 128G of RAM.
Do I need ArcGIS to run Circuitscape?
Nope- our standalone version is entirely open source and operates without an ArcGIS license. Input and output files use an ASCII grid format that can be read and written by ArcGIS and other common GIS platforms. We also have an ArcGIS toolbox that calls Circuitscape directly from ArcGIS, so that grids don't have to be exported to ASCII. You can also export to ASCII using the Export to Circuitscape ArcGIS tool.
Is Circuitscape 'better' than least-cost corridors?
Both approaches provide valuable tools to understand connectivity, and we think Circuitscape and least-cost modeling complement each other. Linkage Mapper now hybridizes least-cost modeling with Circuitscape (see the Pinchpoint Mapper module).