The dividing line is simple; it's when the average changes. You can't objectively measure this, but you can get certainty bounds by using many data points. The smaller the change, the more data points you need.
A single huge drought in a place prone to droughts is not a very strong data point by itself. Unless it literally doesn't end, we'll need many more years of rain measurements to know if the baseline has changed.
If you go global it's a lot easier to take independent measurements and show climate change exists at all. But to ask whether the level of rain in California is from climate change simply can't be answered as quickly.
Actually, an X-year event means that an event of equal or greater size has a 1 out of X chance of occurring per year which is distinctly different from for Y >> X => Y/X occurrences.
For example, a 100 year flood has a probability of occurring (or being exceeded) of 1% per year which is distinctly different from "a 100 year flood is expected once every 100 years".
The law of large numbers implies your statement but that isn't what the term X-year event means.
Also, minor nit, "X year event" doesn't mean the event should happen every X years, but that over a span Y >> X, the event happens Y / X times.