Sorry to summarize so late. Firstly, with my thanks to Digby
Millikan, Isobel Clark, Andrew J. Lister, Jens Nieschulze, Donald E.
Myers, Brian Gray, especially to Jens Nieschulze and Andrew J.
Lister, who are so kind to send the references/articles to me.
Original Question: Is a 50m*100m forest plot with 1623 trees enough
to geostat species distribution?
1 There are different opinions about area, Isobel thinks it is number
and spatial distribution of samples rather than area coverage, Jens
holds the view that it depends on the forest (trees density?) and the
plot shape should be square, Andrew thinks there are enough trees.
2 On the methods, Donald gives the general rules in detail, however,
most of them advise me to adopt indicator variogram and indicator
kriging, Digby sets different values for different species, Andrew
sets 0, 1 for the absence/presence of one species. Brian suggests
that it is a matter of multivariate normal assumptions more than a
multivariate binominal one.
Since I am a newer in geostat, there must be some misunderstanding,
please feel free to correct it.
Institute of Botany,School of Life Sciences,
Zhongshan University,Guangzhou,Guangdong Province,
Zip Code: 510275