MSJollySeber Multi-state Jolly Seber analysis

[keevilm]

I have been unable to find much information on the multi-state Jolly Seber analysis. I am currently working with a dataset consisting of 17 years of mark recapture at 3 sites with occasional movement of individuals between them. I have fitted a handful of multistrata models from RMark but I would also like to estimate abundance. Unfortunately, I have not been able to find an example of an MSJollySeber analysis to get a better idea of how to set up the pi and pent parameters in the multistate case. Here is the information I can find is only the article in the Mark GUI help, this thread viewtopic.php?f=1&t=4329&p=14391&hilit=multistate+jolly+seber+popan#p14391 and a mention in with the new version announcement viewtopic.php?f=21&t=4345 (and also in the documentation zip file)

RMark 2.2.9 (13 Sept 2021)

NEW FEATURES

* added model MSJollySeber with example of simulated data courtesy of Gary White

Try as I might, I just can't find the example mentioned and I would love to take a look at it. Thanks.

[cooch]

Because the new chapter in the 'Gentle Introduction' covering this model hasn't been approved for 'public release yet' by its author (Bill Kendall), who developed the model (conceptually), and worked with Gary on implementing it in MARK. I don' believe Jeff has added it to the roster of MARK models that are accessible through RMark.

[keevilm]

Thanks, that explains why I can't find a description. Having tried it, I can say that it does work work from RMark, or at least I can say that it runs and returns results with fitted values for pi and pent parameters. I will either work from the GUI or take another approach to estimate abundance for now.

[jlaake]

That model is in RMark as you discovered. Not sure why example is not showing up. I'll have to look into it. Pent is probability of entry over time like in popan models. I believe pi is probability you enter into each particular state. These should be defined in MARK help files. Abundance will be derived parameters which are stored in results$derived within model object.

[keevilm]

Thanks for looking into it. The help has verbal descriptions of the general meaning of those parameters but I've get two groups and three sites and I am a bit confused in applying constraints and interpreting the output so a worked example would be useful.

[jlaake]

I just posted an update to version 3.0.5 on my google drive that now includes an example for ?MSJollySeber and one for ?HidMarkov which were both promised in version 2.2.9. Not sure why they were not posted. I also fixed the pim header construction for pi parameter in MSJollySeber. For MSJollySeber I used an example from Gary White with 3 states and only setup the model he used to generate the simulated data so I could be assured it was working correctly. I also created another example with dummy values a group variable with 2 groups that you mentioned you have. I didn't use the group field in the model because the data were not simulated with a group variable but if you still need help you can post here and I can show you how to include the group variable in the formula. It does show what happens to the derived abundances with 2 groups.

This new version in not on CRAN but if you are on Windows you can download the following file and install from local zip.

https://drive.google.com/file/d/1-ioCyxtkjrPR9prhsnT7N7t58fjbHNlS/view?usp=sharing

If you are on Mac or Linux you will need to install from source using the following file.

https://drive.google.com/file/d/1-d1ox8BJ-3zfOii2j3XVmUoNuR6jhG_X/view?usp=sharing

Let me know if you encounter any problems. --jeff