Read on Twitter
#Elniño is trending for obvious reasons, but let me inject a note of caution... 
The conventional wisdom pays attention to the ENSO forecasts collated at IRI, which come in two flavors, dynamical and statistical. iri.columbia.edu/our-expertise/…
The initialized dynamical forecasts project a very strong El Niño in the fall/winter this year (Nino3.4 > 1.5). However, the statistical models are far less bullish (predicting Nino3.4 < 0.5). The consensus forecast splits the difference more or less. But that's a big spread!
There is another method tho - w/a similar level of skill - the analog method using existing climate model simulations! The idea is to look over hundreds of yrs of simulation & pick out the years that look most like today, & predict using the average of what happens in those runs.
No initialization shock, no climate model drift, and no new model runs. Almost too good to be true!
(the method is described in Ding et al, 2019; agupubs.onlinelibrary.wiley.com/doi/full/10.10…).
(the method is described in Ding et al, 2019; agupubs.onlinelibrary.wiley.com/doi/full/10.10…).
And what does this method suggest? Basically, it looks like Nino3.4 ~ 1 by the end of the year. An El Niño event, but nothing huge. psl.noaa.gov/forecasts/seas…
I quite like the analog method (for basically aesthetic reasons), but while it doesn't have quite the track record as IRI, it should serve as a note of caution to those predicting Thermageddon next year.
For reference, a Nino3.4 index of 1 (statistically) in the NH winter is associated with 0.06ºC extra in GISTEMP in the year following. Enough for a bump, and maybe a record, but not an imminent breaching of the 1.5ºC level
• • •
Missing some Tweet in this thread? You can try to
force a refresh
