secr

397 views
Skip to first unread message

Samundra Subba

unread,
Aug 7, 2018, 8:20:05 AM8/7/18
to secr

Hi all,

 

I’m trying to combine snow leopard’s telemetry data with camera trapping data by following secr-telemetry.pdf.

 

My input data are as follows:

 

Telemetry data input (example) (file name= te.csv)

Session

ID

Occasion

X

Y

1

7

1

589016

3052965

1

7

2

588916

3052875

1

13

1

588918

3052883

1

13

2

588916

3052874

1

4

1

589010

3052973

1

4

2

593561

3053625

1

4

3

593569

3049035

1

5

1

595696

3049438

1

5

2

598960

3052654

1

5

3

569353

3066356

 

Camera trapping data input (examples)

 

Trap file (file name= trap.csv)

 

Detector

X

Y

1

598960

3052654

0

1

1

1

1

1

2

569353

3066356

1

1

1

1

0

0

3

566963

3060337

0

1

1

1

0

0

 

Capture file (file name= capture.csv)

 

Session

ID

Occasion

Detector

1

1

3

2

1

1

2

2

1

2

4

3

1

2

5

3

1

3

2

1

1

3

5

1

1

4

4

3

1

4

6

1

 

I used following codes

 

trap<-read.csv("D:/Documents/ trap.csv")

 

capture<-read.csv("D:/Documents/capture.csv")

 

trCH<- read.capthist("capture.csv", "trap.csv", detector="proximity", fmt= "trapID", noccasion=60)

 

teCH<-read.telemetry("D:/Documents/te.csv")

 

My two queries are as follows:

 

  1. My halting problem is when I used the function addTelemetry I get the following error in ‘red’

 

combinedCH<-addTelemetry(trCH, teCH, type = "concurrent", collapsetelemetry= TRUE)

 

Error in `covariates<-`(`*tmp*`, value = rbind(covariates(detectionCH),

length of covariate does not match
 
I haven’t kept any covariates so I’m quite confused here.

 

  1. How good it is to use fixed value for sigma of initialsigma value while running secr.fit model: secr.fit(trCH, fixed = list(sigma=7500))? The initial sigma value of teCH object was 7500.

 

Thank you so much in advance, will look forward for your valuable inputs,

Apologies if I couldn’t explain you nicely,

Best wishes,

Samundra

 

Samundra Subba

unread,
Aug 14, 2018, 2:13:14 AM8/14/18
to secr

Hi everyone,

 

With no intention of bothering anyone, was just wondering if my concern could be enlightened.

 

I would be very grateful if my queries could be answered (email below).

 

With best intentions,

--------------------------------------------------------------------------------------------------------------------------------------------------

Samundra Ambuhang Subba, Research Officer, WWF Nepal

P.O Box 7660 | Baluwatar, Kathmandu

Tel: +977 1 4434820- Ext. 024| Fax: +977 1 4438458

     

2.       How good it is to use fixed value for sigma of initialsigma value while running secr.fit model: secr.fit(trCH, fixed = list(sigma=7500))? The initial sigma value of teCH object was 7500.

 

Thank you so much in advance, will look forward for your valuable inputs,

Apologies if I couldn’t explain you nicely,

Best wishes,

Samundra

 

--
You received this message because you are subscribed to the Google Groups "secr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to secrgroup+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

ngwilh...@gmail.com

unread,
Aug 16, 2018, 2:47:23 PM8/16/18
to secr
Samundra, 

I posted a similar inquiry about a month ago with no replies. The only workaround that I was able to find for the "addTelemetry" error was to use the ms.capthist function. I believe the only thing to be concerned about is if you had multiple sessions (I did not). But I was able to combine both datasets with "ms.capthist" and the parameter estimates post-telemetry were much more precise - as expected! 

Please keep us updated on any other solutions you might discover! 

Regards, 
Nathan Wilhite

Samundra Subba

unread,
Aug 20, 2018, 3:06:57 AM8/20/18
to ngwilh...@gmail.com, secr

Hi Nathan,


Thank you so much for your inputs. I have multisession telemetry data so I'm worried if ms.capthist can work. The snow leopards were collared in 4 years interval and the data were kind of in successive sequence (e.g ID-1 had data from 2014-2015 and ID-2 had data from 2016-2017 and the camera trap survey was conducted for 60 days in the summer of 2017).


I'm trying to use oSCR package to combine these data with the help of Chris and Daniel, let see what will come out of it.


Best wishes,

Samundra


From: secr...@googlegroups.com <secr...@googlegroups.com> on behalf of ngwilh...@gmail.com <ngwilh...@gmail.com>
Sent: Friday, August 17, 2018 12:32:22 AM
To: secr
Subject: Re: secr
 
--

Eric H

unread,
Sep 27, 2019, 9:34:51 AM9/27/19
to secr
Good day

I agree with Nathan that it could help to format the data for multiple sessions (which can be done without MS.capthist, see the data input vignette). However, while assuming constant detection parameters across sessions will generally improve precision, it's not valid if those parameters were actually different. I suggest using model selection to see if differences in detection parameters across sessions are supported. If not, you can use the more precise estimates from the model that pools data across sessions (constant parameters across sessions).

E.g. 

mod1: model = list(g0~1, sigma~1)
mod2: model = list(g0~session, sigma~1)
mod2: model = list(g0~1, sigma~session)
mod2: model = list(g0~session, sigma~session).

If you have other covariates, they could also be included.

All the best,
Eric
Reply all
Reply to author
Forward
0 new messages