Skip to content

Plant biomass

1 message · yann clough

#
Dear John,

there is no way to provide a definite answer to your question without 
actually seeing the data and knowing where they come from.
Try to give some thought to what generated the data (e.g. you may have a 
mixture of distributions, for example because processes generating the 
zeros may be different than those generating the biomass data, and/or 
due to experimental treatments or environmental variables).
That being said, the Gamma distribution (see dgamma etc. in R) is the 
continous counterpart of the negative binomial , so that might be a hot 
candidate. Consider also the lognormal - or a mixture distribution. I 
recommend Bolker 2008, Ch.4, for a discussion of distributions and their 
uses in modelling ecological data.

Best wishes

Yann

r-sig-ecology-request at r-project.org schrieb: