long periods of inactivity and/or no location data

[Sean Sutor]

Hello Chris,

I'm working on estimating cumulative and annual home ranges from GPS telemetry for Sonoran desert tortoises. I've got two questions, one regarding methods in your 2020 home range paper with this species, the other with my choice of estimator:

 1. In your 2020 paper reptile homeranges revisited, I see that you and your co-authors excluded all locations between the first and last dates of hibernation. As you may know, the Sonoran desert tortoise hibernates through the winter, then may become active for a few months in spring before returning to inactivity again for the early parts of summer. Could you clarify why locations for hibernation were removed (stationary animal, so it seems logical), and why locations for spring inactivity that can be comparably long were not removed? Or, in the case of my animals that did become active in the spring, should I remove locations I know are stationary when estimating cumulative home ranges?

 2. The data were collected at 30 minutes 24/7 during their most active period, then either at 30 or 60 min intervals from an hour before sunrise to an hour after sunset during the species less-active period. Furthermore, there are occasional gaps that correspond to shelters that preclude satellite connection. Am I correct that wAKDEc would be the most appropriate estimator for the data I am working with?

 I've included a plot depicting GPS fixes over time for four animals, the gaps being when the animal enters hibernation/estivation and loses satellite connection.

 Thanks in advance! I tried to find and bug you about this last week at TWS, but failed to do so :) I appreciated the workshop (2023) and enjoyed your talks this year.

 

Sean

(apologies to those who already read this - I needed to delete and re-upload to replace the figure I included)

[Christen Fleming]

Hi Sean,

1. Generally, you want to segment distinct behaviors and estimate separate ranges for them. I am not sure why the spring inactivity was not segmented, unless perhaps the difference wasn't so large. You would have to ask Roy.

2. Definitely if you switched the sampling rate from 30min to 60min.

Best,

Chris

[Sean Sutor]

Hi Chris,

Thanks for getting back to me. I've calculated wAKDEc for my study animals and plan to use the home range estimates in a mixed effects meta-analysis in the metafor package, and using the log transformed home range estimates in my model. I was wondering if you could shed some light on what kind of log transformation the ctmm's Log() function is using. 

Using your example (https://ctmm-initiative.github.io/ctmm/reference/log.html), Cilla's log transformed home range estimate (est = 376.2317 km^2) using the ctmm's Log() function is 19.77354. However, using base R's log() function returns 5.930205. Is there a particular advantage to using ctmm's Log() function versus another log transformation?

Thanks in advance, and my apologies if the answer to this is hiding from me in plain sight.

Sean

[Christen Fleming]

Hi Sean,

There is a bias correction included in ctmm's Log function, but the predominant difference there is that ctmm is transforming the area in square meters rather than square kilometers.

Best,

Chris

[Sean Sutor]

Excellent - thank you for explaining that!

Sean