Using an arguably more robust dataset (fixed in time, space and landscape type with less bias in selection and less opportunity for evasion) there’s still no apparent link between landscape type and puritanism.

For thoughts about why not, have a look at the Conclusions.

Here’s a link to a map of the Great Ejection in Dorset where you can see the data laid out across the different landscapes. “C Events” are ejections recorded in Calamy – the dataset I’ve used.