initial.ages of long-lived species

posts related to the RMark library, which may not be of general interest to users of 'classic' MARK

initial.ages of long-lived species

Postby awan » Thu Mar 07, 2019 9:46 am

Dear all,

I suspect I am running into limits of what may be possible in Mark/RMark, but I just wanted to ask your advice in case I am missing something obvious.

I am studying a long-lived species (up to 43 years), and because of historic metal-ringing etc., in many instances we can identify the minimum age of an individual, which ranges from 0 to 39. From a previous post, ages can be automatically converted to an age group and init.age

Code: Select all
data$MinKnownAge=as.factor(as.integer(data$MinKnownAge))
init.ages=as.integer(levels(data$MinKnownAge))
ms.pr=process.data(data = data, begin.time = c(1983), groups = "MinKnownAge",
initial.age = init.ages, age.var = 1,
model = "MSLiveDead", strata.labels = c("A", "D", "N","W"))


The design data took a long time to create, and no wonder why, the survival design data is 104,000 lines because it includes a line for every initial age, stratum and year (25 years, 1983 - 2018).Psi is 300,000 lines

Is there a way to include a broad array of initial ages without creating a group for every init.age? I have run simpler analyses with just four age groups, but I am investigating senescence and thus it would be great if I could include the true(ish) age of the individual. I later use add.design.data to create agebins, but I would prefer individuals to enter the agebin at the correct time.

Many thanks in advance.

Andy
awan
 
Posts: 16
Joined: Fri Feb 10, 2017 10:52 am

Re: initial.ages of long-lived species

Postby jlaake » Thu Mar 07, 2019 2:54 pm

A couple of choices here. One is to bin the initial ages in the data prior to grouping to cut down on number of groups. For example, 0-2,3-5,... and then use midpoint as the age here 1,4,.... A little kludgy but if the species is very long lived this can cut down on the groups with some inaccuracy. A better alternative is to ignore the age variable in the design data and use a time-varying individual covariate for age. This will necessitate creating 43 age values (age1,age2,,,age43 assuming time values are 1,2,..43) and assigning the age of the critter at each occasion. For occasions prior to first capture you can use any value but 0 makes the most sense but those data prior to first capture are not used in the MSLIVEDEAD model. Then you would use age in the formula but to avoid confusion I would suggest using a different name or use all caps AGE so the code won't confuse age in design data with individual covariate. There are advantages/disavantages of group variables vs individual covariates - see section in Workshop notes on this.
jlaake
 
Posts: 1417
Joined: Fri May 12, 2006 12:50 pm
Location: Escondido, CA


Return to RMark

Who is online

Users browsing this forum: No registered users and 10 guests

cron