Geostatistical R Package - Board

Is there a way of applying Conditional Geostatistical Simulations across locally disaggregated sediment patches - i.e. if the species being surveyed is only found on specific habitat types, and these areas are not contiguous across the survey domain?

Hello there
If I understand clearly your problem, you would like to perform one or several simulation(s) over a habitat which is split into separate non contiguous parts over the field. Moreover let me add that you want each simulation to remain consistent when comparing the outcome on different parts.

To solve this problem, let me simply illustrate it by saying that you would like to perform a simulation over the whole field ... but to make it visible ONLY in the parts corresponding to the habitat of your specific species. This assumes that the different parts, although separate, belong to the same continuous field.
Am I correct?

If this is correct, the solution is very easy!!! And already available!

It suffices to create a selection (mask) which only keep the "valid" parts of the habitat. This selection must be applied on the output Db (say the output grid). Finally if you run the simulation process using this grid as output (dbout), the selection will be performed over the whole field but written in the output file only within the patches.

This should solve your problem...

Please tell me if I was right.

Didier Renard wrote:Hello there
If I understand clearly your problem, you would like to perform one or several simulation(s) over a habitat which is split into separate non contiguous parts over the field. Moreover let me add that you want each simulation to remain consistent when comparing the outcome on different parts.

To solve this problem, let me simply illustrate it by saying that you would like to perform a simulation over the whole field ... but to make it visible ONLY in the parts corresponding to the habitat of your specific species. This assumes that the different parts, although separate, belong to the same continuous field.
Am I correct?

The short answer would be that I am not sure! I suppose, however, if one is managing a resource in such a way that these non-contiguous habitat patches are all assumed to belong to a single population/area, then the assumption that they are parts of a continuous field makes sense.

Didier Renard wrote:It suffices to create a selection (mask) which only keep the "valid" parts of the habitat. This selection must be applied on the output Db (say the output grid). Finally if you run the simulation process using this grid as output (dbout), the selection will be performed over the whole field but written in the output file only within the patches.

This should solve your problem...

Please tell me if I was right.

I understand your solution, and it makes sense in the context of the above assumption. Thank you!