projection and area

[perrine lequitte]

Hi, I have data issued from movebank, and the coordinates are in WSG84. I wonder if it is nessecary to project them in (in lambert 95) to have a good area estimate? Moreover I don't understand what are the values return by "summary(UD20)" (in the vignette)?

Thank you very much,

Perrine

[Christen Fleming]

Hi Perrine,

WSG84 is the assumed by default datum in as.telemetry(). Imported data will be assigned an automated projection (2-point equidistant) that is generally safe for most calculations, but you can specify a different projection: https://ctmm-initiative.github.io/ctmm/reference/projection.html

You want a projection that's locally flat and in meters to prevent distortion.

With summary(UD), the CI slot gives you the confidence interval of the home-range area estimate. The DOF slot gives effective sample sizes that are related to that calculation. I will update the documentation on that.

Best,

Chris

[perrine lequitte]

Hi Chris,

Thank you very much for your answer,

Best,

Perrine

[perrine lequitte]

Hi,

I have other questions. I have GPS data for 14 crows, that i iported with as.telemetry. But then I tryed to use ctmm.fit(data), it didn't work, so I made as you suggested : ctmm.fit(data[[1]]). So I have to do the calculations for each crow, one by one. Is it possible to do them for all the crows at the same time by giving them an identifier?

When I plot the kernels for a crow, I obtain two groups of points, each having its own contour. Does that mean that the crow has two seaprate homeranges?

is it possible to plot the kernels on a map (using ggplot for example)?

Thank you very much,

Best,

Perrine

[Christen Fleming]

Hi Perrine,

Unless you know what model is going to be selected, I would use ctmm.select() rather than ctmm.fit(). There are a lot of options to run multiple individuals at once. You can do this in a for() loop, parallelized foreach() loop, or in parallelized ctmmweb commands. Semantically, ctmm.fit() and ctmm.select() don't take multiple individuals, because eventually that operation will eventually be used for population model fitting.

If you get a bimodal distribution, then that could be switching between two different nests. It could also be a nest area and a very localized foraging area that isn't adjacent to the nest.

You can export ctmm objects in various formats useful for plotting in other packages: https://ctmm-initiative.github.io/ctmm/reference/export.html

I haven't built in direct map support into ctmm yet, but it is in ctmmweb.

Best,

Chris

[perrine lequitte]

Thank you very much, I will try this!

Best,

Perrine