warning message in revisitation function

24 views
Skip to first unread message

yu lei

unread,
Mar 8, 2022, 1:25:21 AMMar 8
to ctmm R user group
   Hi Chris,
   I am interested in the new funtion revisitation to explore the revisit range. But the output UD results always occur warning message like 
       "In npr(data, UD, variable = "revisitation", normalize = TRUE, debias = debias,  :
  Revisitation estimation requires a continuous-velocity model."
   then i found the results same as the akde function output, i'm confused about that,  is the model i choosed by AICc not a continuous-velocity model?

In details ,the best  models i have fitted are OU anisotropic and OUF anisotropic, and i put the ctmm.guess plots below: 
1st range.png2nd range.png1st breeding.png
 

Christen Fleming

unread,
Mar 8, 2022, 10:11:34 PMMar 8
to ctmm R user group
Hi Yu Lei,

The OU models do not support continuous velocities, and would result in a warning with this estimation target. This material is not published yet, but strictly speaking you can't estimate revisitation with fractal processes like OU motion. You can estimate what the revistation distribution limits too, but that further assumes stationarity at the moment, and there's no more information in those cases than in the common utilization distribution.

Also, the first variogram there doesn't look range resident, and you might consider segmenting (or running a sliding window) until the variogram flattens.

Best,
Chris

yu lei

unread,
Mar 9, 2022, 6:39:46 AMMar 9
to ctmm R user group
Hi Chris
Thank you for your comments,  i adjusted the first variogram by sliding window, but i'm not sure whether the variogram being flatten and looks like a range resident ? 
All the best,
Yu
slide window.jpg

Christen Fleming

unread,
Mar 9, 2022, 4:49:05 PMMar 9
to ctmm R user group
Hi Yu Lei,

It looks like you zoomed out to the full range of the variogram? By a sliding window, I'm referring to shorter segments of time where the individual might be range resident. It could also be the case that the individual was not range resident during any portion of the observation period. You might also consider looking at a scatter plot of the x-y data, colored by time.

Best,
Chris
Reply all
Reply to author
Forward
0 new messages