[Gstat-info] about gaussian simulation: conditioned or not,
and why so slow?
giohappy at gmail.com
Mon Mar 10 15:31:35 CET 2008
Another question from Italy :-)
I'm running gaussian simulations with gstat, for the very first time.
>From the manual, it doesn't appear clear if it's simply Gaussian
Simulation, or Conditional Gaussian Simulation. My observed values
don't adhere at all to the simulated grid they follow in; so I need to
know it this is due to the grid nodes locations, which are not exactly
centered over the observed samples, or to the fact that simulation is
not strictly conditioned on them.
Second question: to simulate over a grid of about 2500 cells and 78
observed values, it took about 10/15 minutes to run a single
simulation (nsim=1).Is it normal?
I would need to run, at least, 100 simulations!!!
I tried both on a Ubuntu box and Windows.
More information about the Gstat-info