Problems with GENPRES multi-season simulations

announcements (new versions, changes, bugs, installation problems...) related to program PRESENCE

Problems with GENPRES multi-season simulations

Postby mrodtka » Tue Nov 13, 2012 5:35 pm

Hi,

I've been attempting to run simulations of the multi-season model in GENPRES but PRESENCE crashes fairly regularly if I deviate from the default values, especially when I change values in the '# surveys per season' box. Have tried various approaches to inputting this information including recommendation of Sept 09, 2010 post by Eurycea (i.e. entering '5' or '5,5,5,5,5' when I have five surveys per season). Have also tried reinstalling. Anyone else having problems?

Am running latest versions of PRESENCE (5.3) and GENPRES (4) on Windows 7 Professional.

Thanks

Mike
mrodtka
 
Posts: 21
Joined: Tue Nov 13, 2012 4:22 pm
Location: Rocky Mountain House, AB, Canada

Re: Problems with GENPRES multi-season simulations

Postby jhines » Wed Nov 14, 2012 6:54 am

Mike,

One thing that needs to be done with the multi-season model in GENPRES is to make sure the colonization and extinction values match up with the number of surveys per season. For example, if the number of surveys per season is 2, then the 2nd, 4th, 6th,... columns for eps and gam should be non-zero and the others zero. Can you tell me what values for all of the parameters you're using which caused the crash?
jhines
 
Posts: 599
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: Problems with GENPRES multi-season simulations

Postby mrodtka » Wed Nov 14, 2012 5:38 pm

Thanks!

After reading your reply and rereading CH 13 of Donovan and Hines (2007) I suspect I'm inputting my parameter values incorrectly but can't identify where exactly I'm going wrong. Desperation is setting in!Here's what I am trying to simulate:

Assessing occupancy at 22 sites, 5 surveys of each site per season, repeated for 5 seasons (i.e. 550 individual surveys total). PSI: 0.47, P: 0.5 throughout, I planned on varying extinction and colonization rates to simulate various declines in psi but was starting with the defaults (0.2,0.1) each season. From your post I think I should be entering eps and gam values in the 5th (and multiples thereof) column and all other columns should be zero.

Is there an explicit example of how this information is entered into the GENPRES interface for the Multi-season model that I've overlooked?

Mike
mrodtka
 
Posts: 21
Joined: Tue Nov 13, 2012 4:22 pm
Location: Rocky Mountain House, AB, Canada

Re: Problems with GENPRES multi-season simulations

Postby jhines » Thu Nov 15, 2012 8:36 am

Mike,

GENPRES simulates data in the same manner it generates expected values, by recursively following the number of sites through each survey. This means that it generates every possible detection history and once you get over about 12 surveys, the number of possible histories gets big. (For 25 surveys, that would be ~33.5 million histories.) For that reason, I added an option to simulate by individual site, which only generates as many histories as sites. To select this option, go to the 'Options' menu and select 'simulate by individual'.

Obviously, I need to modify GENPRES to automatically switch to this option, or warn you instead of 'blowing up'.

Jim
jhines
 
Posts: 599
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: Problems with GENPRES multi-season simulations

Postby mrodtka » Thu Nov 15, 2012 1:31 pm

Good to know. Looking forward to giving this a try. Thanks Jim!
mrodtka
 
Posts: 21
Joined: Tue Nov 13, 2012 4:22 pm
Location: Rocky Mountain House, AB, Canada


Return to software problems/news

Who is online

Users browsing this forum: No registered users and 14 guests

cron