Event risk: Uncertain future events that may or may not occur. I'd say this includes severe weather events like floods and droughts.
Non-event risk:
Variability type - Uncertainty about some key characteristics of a planned event or activity or decision. Certain weather looks like it probably could be in this category - in terms of unseasonal conditions (of course, this is in the pmbok). Hurricane or a drought is an event risk though. But unseasonably warm or cool temperatures are variability-type non-event risk. There is no “risk” that it will warm up during the summer... we know that will happen, but if it gets hotter than any record temperature, that’s non-event variability risk. Per the pmbok, this type of risk is managed through Monte Carlo analysis, so we are limited to planning based on probabilities of outcomes. I would suggest that we are limited to Monte Carlo because there is not expert judgment that could help.
Ambiguity type: a lot like event risk, but more limited in terms of the unknown condition. So if we have to interpret a statute to implement a plan, and there is uncertainty or ambiguity in the statute, we take a risk that get it wrong, and maybe have a contingency plan for alternative interpretations. The pmbok recommends we deal with this risk through expert judgment or benchmarking. This one is not uncertainty about a planned event or activity or condition. I think that is important, but I am having a hard time articulating how.
