[R-sig-eco] bursts in adehabitat
Mathieu Basille
basille at ase-research.org
Thu Sep 8 20:55:43 CEST 2011
Dear Jacqueline,
Sorry if I'm a little late on this conversation. To complete Josh's
answer, a burst is intended to be a series of relocation from one
individual collected at regular interval. Imagine a monitoring project
based on GPS collars, with one relocation every 1h year-round. Every
time an individual is fit with a new collar, a lag might be included
between the two collars, so that each collar would define a new burst.
Now, if the schedule is set to 1 relocation every hour at day, and 1
relocation every 4h at night, every night and every day of each
individual should be considered as separate bursts (that can make a lot
of bursts).
It doesn't matter if you have some missing values within a burst (see
e.g. 'setNA' to add NAs) or if the interval is not precisely the same
(see e.g. 'sett0' to make the trajectory regular). What is important
though, is that the burst defines the individual set of relocations used
for many functions, as Josh already underlined. In your case, I would
probably use each movie as a burst, complete the trajectory and make it
regular if it is of any use for you.
You might have a look at:
vignette("adehabitatLT")
which is incredibly useful to me when I have doubts about the use of the
package.
Best,
Mathieu.
Le 2011-08-26 10:45, J. Augusiak a écrit :
> Many thanks Josh.
>
>
>
> This makes it much clearer now. I do rely on time intervals as I do the
> tracking based on an “1-image-per-second” basis as calibration for further
> modelling where I also plan to work with fixed time steps.
>
>
>
> Now I know how I will handle ltraj. :-)
>
>
>
> Thanks again,
>
>
>
> Jacqueline
>
>
>
> From: Josh Nowak [mailto:Joshnwk at yahoo.com]
> Sent: vrijdag 26 augustus 2011 16:18
> To: J. Augusiak
> Subject: Re: [R-sig-eco] bursts in adehabitat
>
>
>
> Jacqueline,
>
> In my experience it really only matters if the statistics you are doing
> require that the fixes be separated by regular time intervals. This is why
> bursts are useful to me, a GPS collar will often take fixes at regular
> intervals, but sometimes it misses a fix and so the fix interval is made
> irregular. If my goal is some sort of movement analysis that requires
> regular intervals then I will use bursts so as not to throw out data or rely
> on interpolating the missed fixes. This approach is a nice fix, but the
> regular interval requirement can likely be overcome too.
>
>
>
> Josh Nowak
> Département des sciences du bois et de la forêt
> Université Laval
> Québec, QC G1K 7P4
> (418) 656-2131 Ext.6110
>
>
>
>
>
> _____
>
> From: J. Augusiak <jaugusiak at googlemail.com>
> To: r-sig-ecology at r-project.org
> Sent: Fri, August 26, 2011 9:36:50 AM
> Subject: Re: [R-sig-eco] bursts in adehabitat
>
> Hi Roman,
>
>
>
> I use cameras which are fixed centered above an aquarium which is 1m2 of
> size. I record the movement for, say, one hour. Then I take the video file,
> dissect it into its single frames which I then import as image sequence in
> ImageJ, an open source image processing software. A custom made plugin for
> this software then analyzes frame by frame the position of my animal in the
> aquarium, so basically I get a set of plain x and y coordinates.
>
>
>
> Due to some setup restrictions on some images in the sequence the animal
> can't be detected by the software and it interrupts the current track. E.g.
> On images 1-300 the animal can be seen, I get a coherent track. On images
> 301 to 367 the animal cannot be seen but is visible again on image 368 to
> 400. This would give me track 1 (1-300) and track 2 (368-400).
>
>
>
> Adehabitat's definition of a burst is as follows:
>
> For a given individual, trajectories are often sampled as "bursts" of
> relocations. For example, when
>
> an animal is monitored using radio-tracking, the data may consist of several
> circuits of activity (two
>
> successive relocations on one circuit are often highly autocorrelated, but
> the data from two circuits
>
> may be sampled at long intervals in time). These bursts are indicated by the
> attribute burst. Note
>
> that the bursts should be unique: do not use the same burst id for bursts
> collected on different animals.
>
>
>
> I already had lengthy discussions with my colleagues about it.
> Theoretically, one could say that I have one "circuit of activity" with one
> movie (=one burst). On the other hand, I get those track fragments as
> described above. Those fragments themselves could be interpreted as bursts
> also.
>
>
>
> So what do I need to do when setting up my ltraj objects? Define burst as
> "1", because it's one movie, or say burst = 1, 2, 3,... depending on the
> number of track fragments for one movie?
>
>
>
> Many thanks for your reply and time!
>
>
>
> Jacqueline
>
>
>
> From: romunov at gmail.com [mailto:romunov at gmail.com] On Behalf Of Roman
> Lu¹trik
> Sent: vrijdag 26 augustus 2011 14:49
> To: J. Augusiak
> Subject: Re: [R-sig-eco] bursts in adehabitat
>
>
>
> Hi Jacqueline,
>
>
>
> what software/hardware are you using for tracking your animals?
>
>
>
> Cheers,
>
> Roman
>
>
>
>
>
> On Fri, Aug 26, 2011 at 2:09 PM, J. Augusiak <jaugusiak at googlemail.com>
> wrote:
>
> Dear list,
>
>
>
> I have a question concerning the definition of a "burst" in the adehabitat
> package. Somehow I 'm not able to wrap my mind around it and to adopt it
> appropriately for my data set.
>
>
>
> I want to analyze the movement behavior of aquatic invertebrates and how it
> may be impacted by different environmental influences. In order to study
> this I have set up an arena where I let the animals walk around and video
> tape their locomotion over time.
>
>
>
> However, my image quality is bad at times, or an animal hides under an
> object that I placed in the arena, which causes the tracking software to
> loose a track but it may pick it up again after some time when the
> individual is visible again. Those separate tracks get different numbers by
> default.
>
>
>
> My problem of understanding is now, are the different track ID's that I get
> from the tracking software "bursts" or not. It's only one animal per movie I
> look at and it's only one movie of, say, one hour that I record and for
> which I get track no.1, 2, 3. etc.
>
>
>
> According to my current understanding I am dealing with only one burst in a
> situation as described above and I would get several bursts, if I would
> record one animal for 10 minutes, have a break of 10 minutes, record again
> for 10 minutes, etc.
>
>
>
> I'm sorry if this is a trivial question. But the longer I think about it or
> the more often I read about it the more I get confused and unsure.
>
>
>
> Many thanks for your time!
>
>
>
> Jacqueline
>
>
>
>
>
>
>
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
--
~$ whoami
Mathieu Basille, Post-Doc
~$ locate
Laboratoire d'Écologie Comportementale et de Conservation de la Faune
+ Centre d'Étude de la Forêt
Département de Biologie
Université Laval, Québec
~$ info
http://ase-research.org/basille
~$ fortune
``If you can't win by reason, go for volume.''
Calvin, by Bill Watterson.
More information about the R-sig-ecology
mailing list