PubPeer

[pukpr]

https://pubpeer.com/publications/398D8BF9028B5861D1340E02AB7911

Swedan has just admitted to recycling the data. The statement "the observed ENSO-related data are used to derive the Oceanic Niño Index" is the telling admission. He's describing a loop and calling it a pipeline: observed record → sets initial conditions + forcing parameters → model runs → produces theoretical ONI → compared back against the same observed record. Each arrow correctly exists yet the problem is that the loop closes. "Derive ONI" implies one-directional flow (data → output), but the actual dependency graph has the observed data entering twice — once as calibration input, once as validation target — which is the textbook definition of non-independence between fit and test set. Transparency about a circular method doesn't stop it from being circular — it just makes it easy for us to point out.

What is to prevent more of these empty -- nearly tautological papers -- from being published? Anyone can prompt an LLM to generate fits to data from non-independent sources, but this does not advance research.

Moreover, the skill level of the fitting does not pass rudimentary test. The charts from the paper are attached

There is no long term coherence here. At best the models are fitted for initial conditions and barely are able to recover a pair of cycles. And these are the only artifacts available (?!?)

As a counter-example, I will take from unpublished work that attempts to do something similar -- Here I will use ENSO-related data (SOI) to derive MJO.

Look closely and see that the model is simply applying a 21-day lag of the SOI data to align with the 140 degree longitude MJO time-series. Moreover this is long-term coherence, covering many erratic cycles. This is not surprising as others have pointed out that ENSO is a precursor to MJO, as described here: https://www.climate.gov/news-features/blogs/enso/catch-wave-how-waves-mjo-and-enso-impact-us-rainfall

The idea that this is "ENSO-related" data is obvious as it is a direct precursor -- SOI is essentially related to ENSO and ENSO forces the MJO, with a lag effect as the disturbance propagates

From the link above:

The S.S. ENSO cruise ship and MJO speedboat making their way across the "harbor" of the tropical Pacific Ocean. The cruise ship represents the stationary ENSO pattern creating steady, rolling waves. The speedboat represents the rapidly moving MJO travelling through the waves created by the S.S. ENSO, altering the wake that the MJO speedboat is producing. The floating person represents the United States that feels the impact from the both the cruise ship's wake and the speedboat's modified wake. Note that United States refers to just a specific region in the U.S. that changes with time. The impacts from the MJO and ENSO wave interference can vary both in space and time. Climate.gov cartoon by Emily Greenhalgh.

The lesson is that using"related data" has it's uses, as with SOI data with ENSO or ENSO data with MJO. But with this reviewed paper, neither the independence nor utility of the approach has been shown.

[pukpr]

The previous comment was rejected by PubPeer!

Someone else must contribute criticisms here, as this paper is egregiously bad.