Fact-Check Analysis

Claim: Dai Ato's model had an error of one to two parts per million by 2022 when using only ocean temperature as a predictor.

Evaluation:

Based on my research of Dai Ato's published study in Science of Climate Change (Vol. 4.2, 2024), the claim is essentially accurate with some important context:

• Dai Ato's study used multiple regression analysis with sea surface temperature (SST) and human emissions as explanatory variables¹²
• The study found that "the independent determinant of the annual increase in atmospheric CO₂ concentration was SST, which showed strong predictive ability"¹
• The smallest error in 2022 was 1.45 ppm, achieved using the HAD-SST dataset with data from 1959 onward²¹
• This 1.45 ppm error falls within the claimed "one to two parts per million" range¹²
• The model showed an extremely high correlation with actual CO₂ measurements (r = 0.9995, P < 3e-92)²¹

Important context:

• The 1.45 ppm error was specifically for the best-performing model configuration (HAD-SST from 1959). Other SST datasets or shorter time periods produced larger errors (e.g., 21.1-30.8 ppm when using data only since 1979)¹
• While the regression technically included both SST and human emissions variables, the analysis demonstrated that human emissions were statistically irrelevant, making SST effectively the sole predictor²¹

Overall validity: 90%

1. [PDF] SCC-Ato-Multivariate-Analysis-Vol.4.2.pdf - Science of climate change
2. Dai Ato: The Sea Surface Temperature Rules