Incorparating survey covariates

[Mauricio]

Dear members
I am pretty new researching with occupancy models and also using presence. I am researching frogs and I set 100 sampling units which were visited 4 times each one , twice in the morning and twice in the afternoon. I have a doubt about incorporating 2 covariates into presence design matrix. One of them is weather condition which I divided into 4 categories sunny, partialy sunny, cloudy and rainy.

The other covariate is sampling schedule which I codified as 1= morning and 0= afternoon
I was trying to incorporate sampling schedule as follow:
Site 1: 0 0 1 1 (afternoon - afternoon - morning - morning)
Site 2: 1 0 0 1 ( morning - afternoon - afternoon - morning)
Site 3: 0 0 1 1 ......and so on.

About weather condition I was thinking to assign
1 0 0 0 for sunny
0 1 0 0 for partialy sunny
0 0 1 0 for cloudy
0 0 0 1 for rainy
But I have 4 visits and I dont know how to design the matrix

II was reading and doing all the exercises in "Donovan, T. M. and J. Hines. 2007. Exercises in occupancy modeling and estimation" from chapter 1 to 9 and I was checking all post in this forum but I havent found an answer yet.
If someone could help me I would be really grateful.
In advance thank you very much
Mauricio
Ps: I apologize because my english

[darryl]

It's similar to what you've done for morning/afternoon, but now you're going to need a series of covariates, 1 for each level of the weather covariate.

So if you recorded weather conditions as (using the obvious abbreviations):

Site 1: s ps c r
Site 2: r c r s
Site 3: s s ps s
...

The the first covariate for sun (or not) will look like:
1 0 0 0
0 0 0 1
1 1 0 1
...

the next one for partially sunny will be:
0 1 0 0
0 0 0 0
0 0 1 0
...

etc.

Does that help?

[Mauricio]

Dear Darryl
thank you very much for your answer, I am very grateful. Let me check if I understood. According to your answer I should divide weather condition in 4 covariates and entering as 4 sampling covariates in the presence design matrix: one covariate for cloudy , one for rainy , etc.

My question now is , can I enter all the weather conditions as a 1 sampling covariate assigning values like
Rainy= 0
Cloudy= 1
Partialy sunny= 2
Sunny= 3

and thus the data in presence would look like
site 1 : 0 3 3 0 (r -s-s-r)
site 2: 0 1 1 0 (r -c- c-r)
Site 3: 1 1 1 2 (c -c -c -p)

I ask this because I was reading a suggestion in this forum (Survey covariate format by berghsm » Mon Jul 06, 2009 10:28 am ) where jim suggested that Disturbance could also be entered as a 'continuous' covariate( even though disturbance was not measured as a continuous covariate):

<pre>
dist=0 if disturbance=none,
dist=1 if disturbance=low,
dist=2 if disturbance=medium,
dist=3 if disturbance=hi
</pre>

In advance thank you very much for you kind help.

[darryl]

You could do it that way, but that is now assuming that there is some consistent change as the weather changes from sunny to rain (weather is now a ordinal categorical covariate). My previous suggestion would allow for inconsistent changes and presupposed no one set of conditions would be best. As an ordinal categorical covariate either sunny or rain would be best, it can never be cloudy or partial sunny. Probably ok for some species, but not all. You're free to use whatever you feel is best.

[Mauricio]

Thank you very much Darryl I really appreciate your advice
I think I will use weather as a 4 covariates keeping it as a categorical ´factors because I can not assume that there is some consistent change in them.
Thanks again and best regards