This graph is from a meteorology student doing thesis work on extreme rainfall events. It's from the famous smethport, Pa 1942 event. The lines are recurrence intervals based on hydrology studies that are done for all the United States. The red dot is the observed event. There seems to be a major disconnect in exponential orders of magnitude. I'm sure part of the issue are standard fat-tail events are weird and smoothing for all points within a specified grid. But my lord, the upper bounds for what can happen at any one point in an area are off the charts compared to our models. Sent from my SM-G970U using Tapatalk