Irregular Sampling Handling

[Noah Osterhoudt]

Hi Chris,

I have been testing out doing home-range estimations using ctmm with my data collected on movements of Black Vultures. The tags I use have a dynamic sampling interval that goes up and down with battery life, and these are setting sI can change at any point. In the process of running the tags from summer to winter and the loss of sunlight hours,  I have had to alter my sampling intervals consistently until finding the right combinations. This results in some points being taken from every minute, to 105 minutes, to 120 minutes. Because of this, I have been running akde models with weights = TRUE, but was wondering if I should be pre-processing my data to resample it down to a coarser timescale, or if the wADKE is robust enough to handle this variation? I've included a photo of the sampling interval plot here as well.

Additionally, I've included a sample of a bird whom I believe classifies as a range resident. In the variogram with no alterations using the dt function, you can see the variogram is quite noisy, and when reduced to 10 hours, you can see you lose much of this, but it doesn't exactly smooth out. I wanted to double check that this passes the eye test of home-range residency?

[Jesse Alston]

Hi Noah,

The weights can handle this--no coarsening is needed.

These variograms look fine.

Jesse

--
You received this message because you are subscribed to the Google Groups "ctmm R user group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ctmm-user+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ctmm-user/44b63cca-0711-4f4f-b947-165eabe220a1n%40googlegroups.com.

--

Jesse Alston

+1 252 532 5236

jmalston.com

[Noah Osterhoudt]

Jesse,

Great, thank you! I also had one follow up question if you don't mind!

I was wondering how to interpret or to handle these tiny satellite ranges that are formed with the pHREML wAKDE model? Looking at the data from one of these small ranges, the bird only spent 2 days at this place (1 day in July and 1 in September), but I assume the intense use of this one small area caused this formation? 

I included a screenshot of the plot formed in R and and a screen shot of only the est home range plotted in arcgis pro. 

Please let me know what you think, thank you!

Best,

Noah

[Noah Osterhoudt]

I also figured this may be helpful to see the intensity of points in the areas.

Best,

[Jesse Alston]

Yes, the 95% coverage area includes the smallest area that adds up to 95% cumulative probability, which can create lacunae and disjunct portions. I imagine these were large-ish carcasses that the birds found and hung out near for a stretch of time. If that is true, these are transient resources and they may not be predictive, but they could also be roosts or some other predictable resource that the bird would occasionally visit in the future, meaning that it would be predictive. It is hard to say without ground truthing. In any case, I would not cut these out of the home range if you are calculating home range areas--even if these birds are not visiting these specific areas in the future, they will likely visit similar areas in a similar way moving forward.

Jesse

To view this discussion visit https://groups.google.com/d/msgid/ctmm-user/7aa9111e-9f8b-46c5-8271-73835eed8a83n%40googlegroups.com.