Is there any lower limit in the number of sites to be meaningful? Not really, but few sites will usually result in large standard errors of the indices and trends.

Not really, but few sites will usually result in large standard errors of the indices and trends.