Need help with population estimate using mark recapture data

[Mariana]

Hi I am a honours student at James Cook University, Australia. I am working with feeding ecology of juvenile green turtles at Green Island. And I am having some problems estimating the population of juvenile turtles at a given area utilizing mark recapture data. Since the main aim of my project is not to estimate population I dont have much time to spend on this. BUt I think it would be very valuable for my project to at least have a robust estimate .
I have been trying to run the analysis utilizing jolly Seber, however when I run the program it says that a numerical convergence was never reached. I am not sure what this means.
I would be very please if anyone could help me out .
Thanks
Mariana F.

[Guest]

Hi. I've been trying to use the Jolly Seber function as well and have received the same error. I also get the following error: Error number 2 from VA09AD optimization routine. I attempted to run the data with the other optimization routine but it was also unsuccessful. The data is correctly formated to my knowledge, I have no problem running CJS models with the same input file. Do input files for JS differ from CJS, and if so, how? Also, I am running an old version of MARK, v2.1, which I find to be more stable than the newest version, however there might be a problem with running JS on this version - if anyone knows about this, please let me know. Thank you very much.

[Susanne Meyer]

Hi there,

I have the same problems with JS - always shows me "numerical convergence never reached with max. G= 3828780" ??? What does that mean? Tried POPAN instead - it worked, but my population size is (I think) too small. I captured 220 individuals at one pond and POPAN tells me I have a population size of 222. My capture efficiency isn't that good.
Also Robust Design shows me exactly the number of individuals I've captured per session. How come?
Can anyone help?

Thanks
Susanne

[RCulloch]

Hi,

just to join in with this one, i am trying to run several pradel survival and lambda models, but two models don't work (phi(t)p(t)lambda(t) and phi(.)p(t)lambda(t)) In both instances an error message with numerical convergence never reached followed by max G = 895.4198 and max G 311.8782 respectively for the models are displayed. If someone could explain why this might be occuring and how, if possible to solve it then i would be most grateful. I normally follow the link to the old forum in the FAQ section of this forum but it seems to be gone!

Thanks and best regards,

Ross Culloch

[adelong]

Hi

I, too, need help with estimating population size using MARK. Previously, I was using the CJS model to obtain phi and p, but discovered that the multi-strata models might work better for my organism. I'm still in the process of learning them and although more time consuming than CJS, making really great progress. However, I've been asked to get population estimates for a report due this summer. As this was not my initial intent for using MARK, I have not explored how to do this. Is there a quick and relatively easy way to obtain population estimates?

Angie~

[cooch]

Hi

I, too, need help with estimating population size using MARK. Previously, I was using the CJS model to obtain phi and p, but discovered that the multi-strata models might work better for my organism. I'm still in the process of learning them and although more time consuming than CJS, making really great progress. However, I've been asked to get population estimates for a report due this summer. As this was not my initial intent for using MARK, I have not explored how to do this. Is there a quick and relatively easy way to obtain population estimates?

Angie~

Estimating population size in open populations is not trivial, and is (more often than not) not overly precise. The standard approach is to use the Jolly-Seber model (in one of its various incarnations). There are two commonly used versions in MARK which might generally be worth pursuing: the POPAN model, or Link-Barker. See the MARK helpfile.

If the population is closed, then a closed estimator is the preferred approach. Lots of ways to handle these sorts of estimates in MARK.

At present, there is no way in MARK that I'm aware of to derive estimates of N in a MS context.

Alternatively, you might choose to resist the siren song of abundance as a metric of interest, and focus instead on rates of change: you can estimate realize lambda with some precision (in several ways within MARK), without ever estimating abundance.

[cooch]

Hi,
I normally follow the link to the old forum in the FAQ section of this forum but it seems to be gone!
...

FYI.

Some pages from the old forum can be found in the Google cache, but, the old forum had several security issues which forced its elimination.