Separate genuine cointegration from a measurement artifact

Determine whether the apparent cointegration between global mean sea level and global mean surface temperature anomalies during 1947–1967 reflects a genuine episode of tighter climate co-movement or a measurement artifact associated with the documented discontinuity in sea-surface-temperature observations.

Background

The generalized-infimum Engle–Granger statistic identifies 1947–1967 as the window containing the strongest evidence of cointegration between global mean sea level and global mean surface temperature anomalies. This interval coincides with a documented wartime change in sea-surface-temperature measurement methodology centered on 1945.

The paper notes that the observed mean-reverting residual over this interval may therefore arise either from a genuine period of tighter co-movement between the climate variables or from the observational discontinuity itself. The reported statistics locate the relevant interval but do not resolve which interpretation is correct, leaving the distinction as an unresolved empirical question.

References

These two readings cannot be separated with the statistics reported here.

Cointegration by Parts: Locating Cointegration in Time  (2609.10020 - Kvist et al., 9 Sep 2026) in Section 5, Application: Temperature and Sea Level