Poisson assumption in partial duration series models challenged by new study
The study looked at how well the Poisson process fits data from 26 gaging stations in Great Britain. The researchers found that the Poisson assumption doesn't always hold true for all the data, but it can be acceptable in some cases. They suggest that if there is a relationship between data points, it's better to look for it in the process that generates the data, rather than in the data itself.
