Thanks for your question. The 65/66 and 82/83 maps were lifted out of the second graphic in this previous blog post (https://www.climate.gov/news-features/blogs/enso/united-states-el-ni%C3…). That graphic includes all El Niño events of the past 50 years, including the 57/58 that you mention. For this article, we lifted out two strong events with opposite precipitation anomalies becuase we wanted to emphasize the point that in any given year--even a strong El Niño--other random climate influences can overrule the classic wet signal.

If you look at that graphic, it shows that out of the 7 strong events that have occurred since 1950, 6 of them led to wetter than average winters of various magnitudes in CA. But in 65/66, despite the strong El Niño, CA was drier than average. El Niño is about increased probability, not certianty. That was the main point we were tring to convey by only showing the two examples.

I'm not one of the scienitsts blogging here, simply the editor, but i know from previous blog posts that while "analog" forecasting--trying to predict what will happen based on seeimingly similar past events--isn't completely meaningless, it's usefulness has significant constraints. Even seemingly insignificant differences between one event and another can allow the climate system to evolve along two radically different pathways, leading to different outcomes.

And statististically speaking, when you try to pull out a single year as being "most like" the current conditions and try to infer the future based heavily on that one case, you're essentially trying to draw statistical inferences from a sample size of 1. It would be like trying to infer what all dogs look like based on having seen a single dog.