How do I determine my K value in pcount()

[Brian Whyte]

I'm using pcount() to estimate abundance of primate groups from point transect survey data. The methods I used are the same as other studies that eventually used n mixture models. As I analyze my data though, using the same code shown in the pcount() help example, I notice that every value I use for K dramatically changes the results. A K of 3 gives ~40 monkeys estimated (very likely for the survey area I sampled in). A K of 4 gives ~ 60, and a K of 50 gives 1000... and this continues, where every value I give for K just makes the estimated abundance larger and larger...

In the Royle 2004 paper that pcount() is derived from, Royle shows an application of the n mixture methods where K = 200, but everything from K=20 up to 200 didn't change estimates that much. The data(mallard) example set provided by unmarked also shows this pattern, where at some point changing K doesn't influence estimates that much. With my data being different, does this indicate a problem with my data fitting the assumptions of the n mixture methods?

For reference, attached is my .csv data and the code I've been using to analyze it. Could anyone please let me know why K is functioning differently for me, or if my data seems fit for n mixture methods such as pcount()?

Thank you very much

B

 

[Richard Chandler]

Hi Brian,

If the estimates are highly sensitive to K, it suggests that detection probability is very low. In general, it is hard to estimate or model abundance when p is low. 

I suggest trying to explain variation in abundance and detection probability using covariates. If the estimates don't stabilize after that, you might have to give up on estimating abundance, unless you have some other data, such as distance sampling or capture-recapture data.

Richard

--
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.
For more options, visit https://groups.google.com/d/optout.

--

Richard Chandler

Assistant Professor

Warnell School of Forestry and Natural Resources

University of Georgia

706-542-5815

Lab website

[Kery Marc]

Dear all,

I echo Richard's comments on K. It is good practice to check sensitivity to K for any analysis of binomial N-mixture model, e.g., by comparing the parameter estimates obtained with unmarked's default (100 more than the max count) and with, say, 200 more than the max count. If estimates change, increase K and repeat until estimates stabilize. If you don't get stabilisation, then there is likely to be an estimability problem. All this applies especially for NegBin Nmixture models !

Best regards  --- Marc

---

From: richard....@gmail.com [richard....@gmail.com] on behalf of Richard Chandler [rcha...@warnell.uga.edu]
Sent: 10 July 2016 13:37
To: unma...@googlegroups.com
Subject: [unmarked] How do I determine my K value in pcount()