Strange Results with col/ ext probabilities

[cheryl brehme]

Hi,

I was working on some power analysis when I came up with this issue.  I am using SymDynOcc to simulate occupancy data for 2 years- setting colonization to 0-0.2 and survival phi to 0.49-0.50.  The detection data comes out looking perfect.  Then I run the data in a simple colext model (all parameters constant) and am getting lots of models with low ext effect size and high negative coeff estimates for gamma.  Can anyone shed light on this?

Thank you!

Cheryl

Script: simulate data

L50d<-simDynocc(nsites = sites[j], nyears = years, nsurveys = reps, year.of.impact = NA,

                mean.psi1 = 0.42, beta.Xpsi1 = 0,

                range.phi = c(0.49, 0.51), sd.lphi.site = 0, impact.phi = 0, beta.Xphi = 0,

                range.gamma = c(0.00, 0.02), sd.lgamma.site = 0, impact.gamma = 0,                                  beta.Xgamma = 0,

                sd.lphi.lgamma.site = 0,

                range.p = c(0.90, 0.98), beta.Xp = 0,

                range.beta1.survey = c(0, 0), range.beta2.survey = c(0, 0),

                trend.sd.site = c(0, 0), trend.sd.survey = c(0, 0),

                trend.sd.site.survey = c(0, 0), show.plots = FALSE)

extractL50d <-as.data.frame(L50d$y)

UMFL50d <- unmarkedMultFrame(y = extractL50d, numPrimary=2)

fmL50d <- colext(~1, ~1, ~1, ~1, data=UMFL50d)

fmL50d.sum <-summary(fmL50d)

Example Result summary table of model (from input below):

Call:

colext(psiformula = ~1, gammaformula = ~1, epsilonformula = ~1, 

    pformula = ~1, data = UMFL50d)

Initial (logit-scale):

 Estimate    SE      z P(>|z|)

   -0.158 0.201 -0.786   0.432

Colonization (logit-scale):

 Estimate   SE     z  P(>|z|)

       -4 1.04 -3.84 0.000123

Extinction (logit-scale):

 Estimate    SE        z P(>|z|)

 -0.00272 0.295 -0.00921   0.993

Detection (logit-scale):

 Estimate    SE    z  P(>|z|)

     3.26 0.464 7.02 2.21e-12

AIC: 262.684 

Number of sites: 100

optim convergence code: 0

optim iterations: 42 

Bootstrap iterations: 0 

Example data:

> extractL50d

        Visit1.Year1 Visit2.Year1 Visit1.Year2 Visit2.Year2

Site1              1            1            1            1

Site2              0            0            0            0

Site3              1            1            0            0

Site4              0            0            0            0

Site5              0            0            0            0

Site6              1            0            0            0

Site7              0            0            0            0

Site8              0            0            0            0

Site9              0            0            0            0

Site10             1            1            1            1

Site11             0            0            0            0

Site12             1            1            0            0

Site13             1            1            0            0

Site14             1            1            1            1

Site15             1            1            1            1

Site16             0            0            0            0

Site17             1            1            0            0

Site18             0            0            0            0

Site19             1            1            0            0

Site20             1            1            1            1

Site21             0            0            0            0

Site22             0            0            0            0

Site23             0            0            0            0

Site24             0            0            0            0

Site25             0            0            0            0

Site26             1            0            0            0

Site27             0            1            0            0

Site28             0            0            0            0

Site29             1            1            1            1

Site30             0            0            0            0

Site31             0            0            0            0

Site32             0            0            0            0

Site33             1            1            0            0

Site34             0            0            0            0

Site35             0            0            0            0

Site36             1            1            1            1

Site37             0            0            0            0

Site38             0            0            0            0

Site39             1            1            0            0

Site40             0            0            0            0

Site41             0            0            0            0

Site42             0            0            0            0

Site43             0            0            1            1

Site44             0            0            0            0

Site45             1            1            0            0

Site46             1            1            1            1

Site47             0            0            0            0

Site48             0            0            0            0

Site49             0            0            0            0

Site50             1            1            0            0

Site51             1            1            1            1

Site52             0            0            0            0

Site53             1            1            0            0

Site54             1            1            1            1

Site55             0            0            0            0

Site56             0            0            0            0

Site57             1            1            0            0

Site58             0            0            0            0

Site59             1            1            1            1

Site60             1            1            0            0

Site61             1            1            0            0

Site62             0            0            0            0

Site63             0            0            0            0

Site64             1            1            0            0

Site65             0            0            0            0

Site66             1            1            1            1

Site67             1            1            1            1

Site68             1            0            1            1

Site69             1            1            1            1

Site70             1            1            1            1

Site71             0            0            0            0

Site72             0            0            0            0

Site73             1            1            1            1

Site74             0            0            0            0

Site75             1            1            0            0

Site76             0            0            0            0

Site77             0            1            0            0

Site78             0            0            0            0

Site79             0            0            0            0

Site80             0            0            0            0

Site81             1            1            0            0

Site82             1            1            1            1

Site83             0            0            0            0

Site84             0            0            0            0

Site85             0            0            0            0

Site86             1            1            1            1

Site87             1            1            1            1

Site88             0            0            0            0

Site89             0            0            0            0

Site90             0            0            0            0

Site91             0            0            0            0

Site92             1            1            1            1

Site93             1            1            0            0

Site94             0            0            0            0

Site95             1            1            1            1

Site96             1            1            1            1

Site97             0            0            0            0

Site98             0            0            0            0

Site99             1            1            0            0

Site100            1            1            0            0

[Marc Kery]

Dear Cheryl,

unmarked computes and reports the estimates of these probability parameters on the logit scale. Hence, you need to apply the inverse-logit transformation to these estimates before you can compare them to your input. For the output in your mail, here are the corresponding estimates on the probability scale:

> plogis(-4)

[1] 0.01798621

> plogis(-0.00272)

[1] 0.49932

These would appear to be about right ?

Best regards  --- Marc

---

From: 'cheryl brehme' via unmarked <unma...@googlegroups.com>
Sent: Wednesday, September 22, 2021 03:30
To: unmarked <unma...@googlegroups.com>
Subject: [unmarked] Strange Results with col/ ext probabilities

 

--
You received this message because you are subscribed to the Google Groups "unmarked" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unmarked+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/unmarked/cfc074e9-bdcb-4960-abaa-aaf2c6ba97adn%40googlegroups.com.

[cheryl brehme]

Hi Marc,

Of course!  I am SO embarrassed.  I really need to get more sleep.

I will use predict to get the parameter estimates and CI's.

I really appreciate your quick response.

If there is any way to delete this post, please do! 

Thanks so much,

Cheryl

[Jim Baldwin]

If there's a way to delete some of the stuff I've posted in the past (and future), count me in!

Jim

To view this discussion on the web visit https://groups.google.com/d/msgid/unmarked/38114c23-56ee-41bb-ba87-143712230834n%40googlegroups.com.

[Marc Kery]

Same here ! But, alas, the internet never forgets, NEVER....

---

From: unma...@googlegroups.com <unma...@googlegroups.com> on behalf of Jim Baldwin <jbald...@gmail.com>
Sent: Wednesday, September 22, 2021 18:48
To: unma...@googlegroups.com <unma...@googlegroups.com>
Subject: Re: [unmarked] Strange Results with col/ ext probabilities

 

To view this discussion on the web visit https://groups.google.com/d/msgid/unmarked/CAJCKiwzsgrs9KQJ4H4XLqHhXLaCh_DRmmuPoC9uUNRBJEKFDxQ%40mail.gmail.com.

[Cheryl]

Ha ha! Oh well! Hopefully it will get buried in the depths soon :)

--
You received this message because you are subscribed to a topic in the Google Groups "unmarked" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/unmarked/C6qWZ0JCuvA/unsubscribe.
To unsubscribe from this group and all its topics, send an email to unmarked+u...@googlegroups.com.

To view this discussion on the web visit

https://groups.google.com/d/msgid/unmarked/ZR0P278MB08695D4B5C0C9A0AA46FEC16EBA29%40ZR0P278MB0869.CHEP278.PROD.OUTLOOK.COM

.

[Rob Robinson]

and at least you have shown to the rest of us we are not alone in missing something we later kick ourselves about, so thanks for that :)

And good luck with the rest of your models...

*************** Learn about Britain's Birds at www.bto.org/birdfacts  ******************

Dr Rob Robinson, Associate Director - Research (he/him)
Hon Reader: Univ East Anglia | Visiting Researcher: Swiss Ornithological Institute
British Trust for Ornithology, The Nunnery, Thetford, Norfolk, IP24 2PU
Ph: +44 (0)1842 750050       T: @btorobrob
E: rob.ro...@bto.org      W: www.bto.org/about-bto/our-staff/rob-robinson

======== "How can anyone be enlightened, when truth is so poorly lit" ========

To view this discussion on the web visit https://groups.google.com/d/msgid/unmarked/431523408.175624.1632331982669%40mail.yahoo.com.

[Chris Sutherland]

And also proven, yet again, that this is a very friendly and helpful group!

To view this discussion on the web visit https://groups.google.com/d/msgid/unmarked/CAGcXcEZQqOZpJYTE-8Gwz9ukWCTVxPqnEZbkoz6zLj1oK%2BQw_w%40mail.gmail.com.