This text comes out of a conversation about the question:

» What if climate change affected a transition zone region (for example somewhere on the border between a rain forest and an arid region) such that the region switches from being a 'wet' to 'dry' climate, but in the process gains the remote potential for more extreme storms than those previously possible? Would these characteristics of the transition show up if we were monitoring regional change through statistics defined by Extreme Value Distributions (EVD)? «

Here we create a deliberately contrived situation where this is not the case using the very useful Python lmoments3 library.