Dear Habib,

For patch sampling, you can consider each patch as sampling unit.

In the case of different size of patches you can select a particular area for sampling. You can also have replicates inside of each patches for larger patches (you can get average or use as replicate but with some statistical concern).

You can also read our recent articles for patches effects bahalkeh er al 2021.

Best wishes,

Mehdi

Thank you Mehdi.

Regards.

Dear Habib

I think that one of the principal measures of interest in your problem is the area of the patch, and that it may be difficult to build a sampling frame enumerating each of the patches.

You could do a probabilistic sampling producing coordinates (x,y) that you can find on the field. Then a patch that has an area of the double than other, has the double probability of being selected by this method. Then you can calculate the probability of each sampled patch as proportional to its area. When you determined the selection probability of each patch you can produce estimators weighted by the probability of selection. The sampling method would be weighted simple sample with restitution. The rest of the variables on or under the patches should use the same weighted estimator.

I hope this help.

Guillermo E. Ramos

Thank you very much Guillermo,

Regards,