Occupancy estimation for multiple seasons

questions concerning analysis/theory using program MARK

Occupancy estimation for multiple seasons

Postby zpearson » Wed Aug 02, 2017 10:37 am

I hope whoever reads this will have a little patience.... it will be needed. I have been glued to this forum and other articles trying to somewhat learn MARK.

I am trying to use mark to look at occupancy and detectability of Montezuma quail as part of a master’s thesis. A little about my study:

I conducted point count surveys at 60 locations during the summers of 2015 and 2016. At these locations, during each visit I recorded presence (1 or 0) and a string of survey specific covariates (temp, pressure...etc.). Once per summer I recorded site specific micro and macro habitat variables (herbaceous height, % forb cover, Bunchgrass density...etc.)

In year 2015 I visited each location 6 times and in 2016 7 times. (2015- 00101. 2016- 1001010)
In addition to modeling individual covariates and combinations of covariates I would like to model occupancy by year. What is the best way to do this? I have attempted to use a robust design occupancy and also using the occupancy estimation module where I created a group for year 1 and year 2. In both scenarios. I have sixty encounter histories for each year for a total of 120 rows.

Question 1: In a robust design occupancy, when building models with covariates does it use total occupancy for year 1 and year 2? For instance if I am interested in comparing herbaceous cover from year 1 to occupancy, is this comparing herbaceous cover_1 to occupancy from only year 1 or the combined occupancy from year 1 and year 2. I have the same set of covariates for year 2 that I would like to compare with only year 2 occupancy. If I input a year 1 covariate the model still uses n=120, I would expect n=60.

Question 2: If this is completely off track, what methods are recommended to compare year1 occupancy to year1 covariates and year2 occupancy to year2 covariates? (grouping, creating year variable, robust design vs. occupancy estimation, running yr1 and yr2 from different inp)

Thanks for your patience I am a newbie with MARK and have been mostly self taught.
Posts: 1
Joined: Wed Jul 26, 2017 3:10 pm

Re: Occupancy estimation for multiple seasons

Postby BeckyRuzicka » Sun Aug 13, 2017 8:45 pm

Hi Zach,

For starters, it sounds like you need to modify your inp file.

zpearson wrote:n year 2015 I visited each location 6 times and in 2016 7 times. (2015- 00101. 2016- 1001010)

If you have 60 unique sites you should have 60 rows. Following your example the encounter occasions for a single site should be combined: 001011001010 (I'm assuming you visited the same sites each year). In the robust design, your primary occasions are years and secondary occasions are the site visits within years. In your case, you have 2 primary occasions (or years), 6 secondary occasions within the first year, and 7 secondary occasions within in the second year. If you click on the 'easy robust design times' button in the analysis specifications window it will help you set this up. And if you haven't re-downloaded the MARK book in awhile, take another look - there is now a chapter on occupancy! It has an example similar to this one.

In general, handling time-varying individual covariates can be accomplished by adding a covariate for each interval (in your case year, it sounds like) to the encounter histories (example: 001011001010 1 0.3 0.9; where the last two cols may be herbaceous veg in 2015 and 2016) and then referencing the covariate in design matrix in the appropriate row. Within the DM you will have one column for herbaceous veg that will reference both covariates. I don't know that I can specify any more than that without being more familiar with the parameterization of the model you are wanting to use (I'd likely lead you astray!)

Hope this helps.

Posts: 1
Joined: Fri Aug 11, 2017 3:52 pm

Return to analysis help

Who is online

Users browsing this forum: No registered users and 1 guest