swanrandom: Generate Communities as Random Poisson from Swan Expectations

In jarioksa/natto: An Extreme 'vegan' Package of Experimental Code

Generate Communities as Random Poisson from Swan Expectations

Description

Poisson random numbers lose species as they introduce new zeros for low expected values. They are also unable to compensate this by generating new "unseen" species. To compensate this, we first use Swan transformation with one pass to generate above-zero expected values for unseen species. After generating the Swan probabilities for missing species we standardize all species to their original totals; this is a similar idea as the species coverage in the Good-Turing model (Good 1953). Finally, a community is generated as a Poisson random variate of adjusted observed abundances and Swan probabilities. The process maintains average species richness and diversity, but Poisson distribution probably underestimates the real sampling variance.

Usage

swanrandom(x)

Arguments

x	community data of integer counts

Value

Randomized community matrix.

Author(s)

Jari Oksanen

References

Good, I.J. (1953) The population frequencies of species and the estimation of population parameters. Biometrika 40, 237–264.

jarioksa/natto documentation built on March 28, 2024, 12:45 a.m.