telemetry warning messages

[Andrea]

Hi - 

I am having difficulty adding telemetry data and I'm hoping someone can help. I have formatted my data following the description in the secr telemetry vignette. The first few rows of my data are:

BM16_17	F2	20	671887	6339907
BM16_17	F2	20	671892.2	6339907
BM16_17	F2	20	671723.2	6340445
BM16_17	F2	21	670293.7	6342067
BM16_17	F2	21	669099.8	6342688
BM16_17	F2	21	669099.1	6342686

When I attempt to load the telemetry data:

CHt <- read.telemetry(file = "telemetry_UTM.txt")

I get the warning message "Telemetry data (telemetryxy) do not match capture histories." I cannot figure out how to resolve this.

I have matched the occasion number to the occasion number in my trapping file. For example, a date of November 30th corresponds to occasion 20 in the trapping file so I called a telemetry location from November 30th occasion 20 as well. Perhaps that is not correct? 

If I ignore the warning and go on to combine the trapping and telemetry data:

combinedCH <- addTelemetry (CHBM, CHt, type = c("dependent"), collapsetelemetry = TRUE, verify = TRUE)

I get another warning message: 

In addTelemetry(CHBM, CHt, type = c("dependent"), collapsetelemetry = TRUE,  : covariates in telemetryCH do not match detectionCH so covariates discarded

The covariates in my detection CH include 'sex' and 'trap type' but there are no covariates in the telemetryCH. Again, I'm not sure how to resolve this error. 

I've plotted the data and the plots seem ok.

If I continue to ignore the warning messages and attempt to run a model:

m0.combined <- secr.fit(combinedCH, model=list(D~1,g0~1,sigma~1), mask = maskBM, detectfn = 'HHN', trace = FALSE) 

I get the following warning message:

Warning messages:
1: In subset.capthist(capthist, occasions = det != "telemetry", traps = 1:(nrow(traps(capthist) -  :
  no detections on occasion(s) 1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 15, 20, 22, 25, 26, 28, 29, 30, 31, 33, 34, 35, 36, 38, 40, 41, 44, 45, 49, 50, 55, 56, 57, 62, 63, 66, 67, 68, 69, 71, 72, 73, 74, 75, 76, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 103, 105, 106, 107, 108, 109, 114, 115, 116, 117, 118, 119, 120, 127, 128, 129, 134, 135

2: In subset.capthist(ch, occasions = detector(traps(ch)) != "telemetry") :
  no detections on occasion(s) 1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 15, 20, 22, 25, 26, 28, 29, 30, 31, 33, 34, 35, 36, 38, 40, 41, 44, 45, 49, 50, 55, 56, 57, 62, 63, 66, 67, 68, 69, 71, 72, 73, 74, 75, 76, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 103, 105, 106, 107, 108, 109, 114, 115, 116, 117, 118, 119, 120, 127, 128, 129, 134, 135

Obviously, I've got something incorrect but I can't figure out what the issue is. The trap CH models run fine without the telemetry. 

Any ideas?

Thanks,

Andrea

---

[Murray Efford]

For a start, which version of 'secr' are you using?

--
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.
To view this discussion on the web, visit https://groups.google.com/d/msgid/secrgroup/c8bfeebf-7092-4c0e-ab5b-5d4af3b9a332%40googlegroups.com.

[Andrea]

Hi Murray - 

I'm using secr 4.1.0 in R 3.6.2

Andrea

To unsubscribe from this group and stop receiving emails from it, send an email to secr...@googlegroups.com.

[Andrea]

Hi Murray - 

I'm still struggling with these warning messages. I've asked a few different colleagues as well but no luck yet resolving this error. Any suggestions would be greatly appreciated. Thanks!

Andrea

[David Bernasconi]

Andrea,

Myself and others have also had issues with integrating Telemetry data with secr, particularly with multisession data.  Here are some other folks that have asked similar questions.

https://groups.google.com/forum/#!searchin/secrgroup/nathan%7Csort:date/secrgroup/jXUqznWyGWQ/uQKe7afXCAAJ

https://groups.google.com/forum/#!searchin/secrgroup/nathan%7Csort:date/secrgroup/G2Vcq3Jo7PE/YYIzf8KYAgAJ

The second link addresses your occasion question.  I have gotten passed issues with ocassions in telemetry data by setting them all to 1 since telemetry occasion is rarely used in secr models

Ultimately if you cannot figure out how to solve your problems secr, there is also the option of migrating over to oscr. That is what had to do. The data management is way more involved with oscr ( requires a solid understanding of data tables and matrices in r)  but the developers are super responsive and helpful.

Best,

David

[Andrea]

Thanks David. 

Yes, I've seen those two links. Unfortunately, I'm still struggling with this. 

Your suggestion of setting all occasions in the telemetry file to 1 resolves the first issues I was having. Loading the file with read.telemetry no longer producers a warning message. However, addTelemetry still produces the same warning message as described previously. As does any attempt to run a combined model. 

I tried a slightly different approach using the following code:

msCH <- MS.capthist(CHBM, CHtsingle_nom6)

fit.ms <- secr.fit(msCH, mask=maskBM70, detectfn = 'HHN', trace = TRUE)

predict(fit.ms)

This will run, but no parameters are estimated. The trace detail shows that no LL are being calculated - all say NA. I've tried providing starting values as well as using a different detection function. In addition, I get the warning message: Warning message: In verify.capthist(capthist, report = 1) :  Levels of factor covariate(s) differ between sessions

I'm not familiar with oSCR but will have a look. I'm hoping someone might have some insight here rather than having to learn a new program. 

I appreciate your suggestions. 

Thanks

Andrea

On Friday, January 10, 2020 at 10:52:34 AM UTC-7, David Bernasconi wrote:. 

[Andrea]

While I haven't solved all warnings, I wanted to provide an update in case it helps others. I've been able to resolve 2 of the 3 warning messages. If anyone has insight to the 3rd message, I would greatly appreciate any help.  

The first warning message "Telemetry data (telemetryxy) do not match capture histories" I have resolved by using setting all occasions in my telemetry file to 1. Hopefully this will not cause problems further down the line as the top model without the telemetry data includes bk on lambda0.

The second warning message "In addTelemetry(CHBM, CHt, type = c("dependent"), collapsetelemetry = TRUE,  : covariates in telemetryCH do not match detectionCH so covariates discarded"  I have resolved using the following code:

covariates(CHtsingle_nom6) <- data.frame(sex = covariates(CHBM)$sex[match(rownames(CHtsingle_nom6),rownames(CHBM))])
covariates(CHtsingle_nom6)$sex[is.na(covariates(CHtsingle_nom6))]<-'F'

However, when I try to run a basic model:

fit0 <- secr.fit(combinedCH_nom6, mask = maskBM, detectfn = 'HHN', trace = TRUE)

I still get the following warning message: 

1: In subset.capthist(capthist, occasions = det != "telemetry", traps = 1:(nrow(traps(capthist) -  :
no detections on occasion(s) 1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 15, 20, 22, 25, 26, 28, 29, 30, 31, 33, 34, 35, 36, 38, 40, 41, 44, 45, 49, 50, 55, 56, 57, 62, 63, 66, 67, 68, 69, 71, 72, 73, 74, 75, 76, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 103, 105, 106, 107, 108, 109, 114, 115, 116, 117, 118, 119, 120, 127, 128, 129, 134, 135

2: In subset.capthist(ch, occasions = detector(traps(ch)) != "telemetry") :
no detections on occasion(s) 1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 15, 20, 22, 25, 26, 28, 29, 30, 31, 33, 34, 35, 36, 38, 40, 41, 44, 45, 49, 50, 55, 56, 57, 62, 63, 66, 67, 68, 69, 71, 72, 73, 74, 75, 76, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 103, 105, 106, 107, 108, 109, 114, 115, 116, 117, 118, 119, 120, 127, 128, 129, 134, 135

 

In addition, no log likelihood is calculated. Trace reveals these as NA or Inf (depending on the things listed below).

Things I have tried:

-providing starting values for sigma

-fixing sigma to the average sigma of telemetered individuals

-adding a telemetry scale argument

-setting fast proximity to false

None of these resolve the error. There are not trap detections on the occasions listed, but I can't figure out why that should matter - surely it's common for there to be occasions without any detections. The models run fine without the addition of the telemetry data.

Thanks

Andrea