Why does the year 1998 keep showing up? Well, I know an answer, but it’s not pretty—it’s because it was unusually warm year for the ocean surface (less cold water coming up from below), and thus is a common target for cherrypicking. Every time you pick an outlier as the end of your ranges, you insert bias one way or another—by getting an abnormally high warming rate from “1978-1998” or an abnormally low warming rate from “1998-present” (both used in this post).
This same problem shows up (if your presentation is right) with Easterbrook’s claim to have found a sinusoidal cycle in the ocean—a periodic cycle should not end on an unusually warm year for the ocean, it should end on an average year! But 1998′s high temperature means you can draw nice straight lines through it as an “elbow” in the graph, so of course it’s 1998.
To allay my concerns I would have also liked to see a discussion about what kind of bias is introduced by the hand-picked intervals that start/end at 1998, or even better using a presentation method like running means that doesn’t rely on hand-picked intervals.
Why does the year 1998 keep showing up? Well, I know an answer, but it’s not pretty—it’s because it was unusually warm year for the ocean surface (less cold water coming up from below), and thus is a common target for cherrypicking. Every time you pick an outlier as the end of your ranges, you insert bias one way or another—by getting an abnormally high warming rate from “1978-1998” or an abnormally low warming rate from “1998-present” (both used in this post).
This same problem shows up (if your presentation is right) with Easterbrook’s claim to have found a sinusoidal cycle in the ocean—a periodic cycle should not end on an unusually warm year for the ocean, it should end on an average year! But 1998′s high temperature means you can draw nice straight lines through it as an “elbow” in the graph, so of course it’s 1998.
That would be more shocking if the OP hadn’t specifically mentioned this:
To allay my concerns I would have also liked to see a discussion about what kind of bias is introduced by the hand-picked intervals that start/end at 1998, or even better using a presentation method like running means that doesn’t rely on hand-picked intervals.