Error with running RMark -- Dimension in GETCMD exceeded.

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

Error with running RMark -- Dimension in GETCMD exceeded.

Postby lachrich » Thu Oct 07, 2021 5:58 pm

Hi all, I am fairly new to RMark and MARK in general. I am trying to run a nesting success model though am having issues with the following error. Whenever I run the model, I receive the following error --- "Dimension in GETCMD exceeded".

I have tried to ensure my data frame is in the exact same format as the examples (all the required variables are there, and they are the same class as the example mallard data set). I have omitted extra covariates from my data set in case they were causing the issue but it has not resolved the error. I run my model with the following code:

Code: Select all
```
run_mark_function <- function(){
Null <- mark(data = mark_data, nocc = 4604, model = "Nest",
                    model.parameters = list(S=list(formula = ~1)))

return(collect.models())
}

mark_results <- run_mark_function()

```


Note I have also tried setting up the function as follows with no luck.

Code: Select all
run_mark_function <- function(){
Null <- mark(data = mark_data, nocc = 4604, model = "Nest")

return(collect.models())
}


The error is thrown by R, and also appears at the end of the .OUT file.

One possibility I wondered is if I have set up my 'nocc' argument incorrectly. I have observations of nests over a number of years with varying intervals between observations. I assigned a variable of day 1 (the first day there was an observation) through to the final day there was an observation (4604). Similarly, I have my 'FirstFound', 'LastPresent', and 'LastChecked' variables set according this set up of observation days (the day between day 1 and day 4604).

Any help or guidance would be greatly appreciated!
lachrich
 
Posts: 3
Joined: Wed Oct 06, 2021 9:03 pm

Re: Error with running RMark -- Dimension in GETCMD exceeded

Postby Rotella » Thu Oct 07, 2021 6:33 pm

If the nesting season is constrained to a period within each year, e.g., a 100-day period sometime within each year, you might consider using the number of days in that period as your number of occasions, which would let you have a much smaller number of occasions. You could then incorporate information about which year the data were from using groups (different years = different groups) or individual covariates (e.g., indicator variables for each year) if you want to build models with annual differences in daily survival rates. For example (thinking of a possible Austral nesting season), if the earliest day that you started any nest in any year was October 1, you could treat day 1 as October 1 and reference your dates from that date out until the end of the given year's nesting season. Of course, if you're working with a species that nests continually throughout the year, this might not work so easily. Good luck!
Jay
Rotella
 
Posts: 72
Joined: Mon Jun 09, 2003 11:32 am

Re: Error with running RMark -- Dimension in GETCMD exceeded

Postby lachrich » Thu Oct 07, 2021 6:52 pm

Hi Jay,

Thanks for the reply! Unfortunately I have a species that nests all year round. Though your suggestion may be the only way around my issue. I could set-up my days as 1-365 with year as a grouping variable. However, there are a different number of surveys each year (some have 4-5 others only have 1 or 2), so I feel using year would not be appropriate. Though I would love to hear your thoughts on this. Would it be appropriate to number the surveys using 1-365, though exclude year from the formula?

Thanks!
lachrich
 
Posts: 3
Joined: Wed Oct 06, 2021 9:03 pm

Re: Error with running RMark -- Dimension in GETCMD exceeded

Postby Rotella » Thu Oct 07, 2021 7:07 pm

I think that numbering from 1-365 could work for your situation, but the answer likely depends on what questions/patterns you most want to evaluate. If you are interested in temporal patterns within the year (e.g., dry season vs wet season), then being able to set up a design matrix with a row for every day of the year would allow you to explore a variety of temporal patterns over the course of a year. And, you could imagine also (perhaps) including intercept adjustment terms for different years. Perhaps others have additional or other ideas. Good luck!
Jay
Rotella
 
Posts: 72
Joined: Mon Jun 09, 2003 11:32 am

Re: Error with running RMark -- Dimension in GETCMD exceeded

Postby jlaake » Fri Oct 08, 2021 1:41 pm

I did some searching and this is not an R function, so I ran your nocc with the example mallard data and came up with

ERROR -- Dimension in GETCMD exceeded. Tell Gary.

The Gary is Gary White and this is in MARK. It is occurring when it tries to read in the 4000+ time intervals. I'll let Gary know and see what he thinks unless you have decide to go with annual periods.

--jeff
jlaake
 
Posts: 1417
Joined: Fri May 12, 2006 12:50 pm
Location: Escondido, CA

Re: Error with running RMark -- Dimension in GETCMD exceeded

Postby lachrich » Mon Oct 11, 2021 7:38 pm

Thank you Jeff and Jay for your advice.

Using the annual numbering fixes the issue. Though I realise now that the value specified for "nocc" is what is used to determine the probability of surviving the study. I am interested in the probability of surviving the nest period. Would it therefore make the most sense to use the nest age/day of the nest period instead (i.e. a nest period is 145 days, so I would number 1-145). Then I can include day of year/month/year as potential additional grouping variables.

Cheers,

Lachlan
lachrich
 
Posts: 3
Joined: Wed Oct 06, 2021 9:03 pm


Return to RMark

Who is online

Users browsing this forum: No registered users and 9 guests

cron