r-spatialecology/landscapemetrics

Hectare/Metrics

Closed this issue · 6 comments

This came up in the testing issue:

Hectare vs map units.
Fragstats uses hectare and assumes everything you put into fragstats is in meters and has a resolution > 0.0001. We are obviously not limited by that and could use map units instead. Nevertheless, from the reproducibility and science perspective, one would lose the capability to compare results against fragstats ... but would have much more intuitive results.

Option 1: We leave it as it right now (assuming resolution/grain in meters) and if people use the package and find it weird we change it into map units.

Option 2: They are better reasons for switching to map units than we can come up with. If so, I am happy to change everything to work with map insteads instead of hectare.

Many of the FRAGSTATS metrics do not make any sense for non-cartesian coordinates. E.g. patch area in squared degrees !?... Therefore, in my opinion, an input map needs to have projected coordinates (meters, feets, etc).

It would just be another unit for the spatial scale, wouldn't it?

Anyways, you are for leaving at is and make the warning more prominent in the readme?
I am just not a big fan of implementing warnings everywhere and check for the projection.

No, It would not be. Area of e.g. 1 sq km is always (relatively) stable. However 1 sq degree (sic) is different near the equator and near to the poles... (https://www.stuffintheair.com/images/World_Map.png)

That's why one always needs an expert in the group - makes perfect sense, I am total map dummy ...

Teamwork FTW.

Put a hint in the README that one needs cartesian coordinates and another one is in the getstarted vignette.