R help - Jun 2014 - fitting grid-based models

Hello!

I'm trying to generate grid-based landscape mimicking observed data. For
this purpose I'm using simecol and adapting the "CA" (stochastic
celular
automaton) model included as example in vignette. I think this example
could be nice given observed data has spatial autocorrelation. My observed
data is the percentage of forest cover in a cell, ranging 0.0 to 1.0. I'm
also including an extra parameter (nini) which is the number of initial
points where the CA model starts (in your case you use
init(CA)[18:22,18:22].
So, I would like to fit my three parameters (pbirth, pdeath & nini) to my
observed data. My observed data is also a matrix. I have checked in help
but I haven't seen an example fitting parameters to observed data using a
gridModel. Could you help me?
An extra posibility is to take summary statistics from simualtions (let's
say, autocorrelation or sum of values in the landscape) and try to fit this
statistics with others calculated from observed data.

Thank you very much for your help!
Javier

	[[alternative HTML version deleted]]

Hi,

I have no example at hand, but the usual way could be indeed to
calculate a time series of an adequate statistic (e.g. spatial
statistic) from both, the observed and the simulated data and then to
apply standard model fitting.

Thomas