jeffreyhanson
commented
2 years ago Thanks! Yeah, this FAQ was meant to be about processing data for thousands of species globally (e.g., all amphibians, birds, or mammals). Yeah. the recommendation was specifically for processing all birds globally. If you have >16GB RAM, I don't think chunking is needed for processing amphibians, mammals, or reptiles. I'll update the FAQ to make this clearer (see below). @ctelhado, how does this sound?
Can I produce Area of Habitat data for thousands of species globally?
Yes, the package can generate Area of Habitat data for all terrestrial amphibians, mammals, birds, and reptiles. To accomplish this, you will need a system with at least 16 Gb RAM and 65 Gb disk space. The processing could a couple of days (e.g., if processing all amphibian species) or weeks (e.g., if processing all bird species) to complete. An example script for processing global data is available on the online code repository (see here). Additionally, since a lot of memory is required to process data for all bird species globally, it is recommended to split the full dataset containing all bird species into multiple chunks (e.g., six chunks) and process each of these chunks separately.
ctelhado
https://github.com/prioritizr/aoh/blob/ae92b467770035d6c03170505bf54b64fcbbdce5/vignettes/aoh.Rmd#L215
I think this question is a bit confusing. The tutorial shows how to create AOH for the complete/global range of a single species or few (4) species without issues related to computer power. I think here you want to explain the difficulties around scaling that up to hundreds/thousands species, no? So, do you mean "Can I produce Area of Habitat data for multiple species globally?" (or anything similar to several)
Swap birds for multiple species / all species of a class / thousands of species? Or is the recommendation just for birds? "Additionally, to improve performance when processing data for multiple species globally (e.g., all bird species), it is recommended to split the full dataset into multiple chunks (e.g., six chunks) and process each of these separately."