Analysis of AMOC weakening trends and threshold uncertainties

The primary data for this analysis is derived from the RAPID array at 26N, providing a 20-year continuous record from 2004 to 2023. This dataset is supplemented by statistical early warning signals from SST fingerprints (Ditlevsen 2023) and physics-based indicators (Westen 2024).

The observed trend in the Atlantic Meridional Overturning Circulation (AMOC) shows a decline of -1.19 Sv/decade (p=0.0008; $R^2=0.4718$; computed from the full 2004–2023 RAPID record using linear deseasonalised fit). McCarthy et al. (2025) report -1.0 Sv/decade (90% CI: -0.4 to -1.6 Sv/decade) over a comparable window; the difference is attributable to fitting period endpoints and deseasonalisation method, not a substantive disagreement. The evidence suggests a direction of weakening, with the 2022 minimum recorded at 14.6 Sv (monthly processed value; published annual means for the same period are approximately 15.2 Sv).

However, several critical gaps and systematic uncertainties must be documented to ensure the integrity of this assessment:

First, the quantitative trend rests entirely on a single measurement array. Volkov 2024 demonstrated that a correction to the Florida Current component alone reduces the observed trend by approximately 40%, bringing it to marginal significance; a second correction of similar magnitude would eliminate statistical significance entirely.

Second, the temporal resolution of the 20-year RAPID record is insufficient to distinguish forced anthropogenic decline from multi-decadal natural variability. Because the record is shorter than the variability cycles found in paleoclimate proxies, the observed decline could represent a phase of natural variability rather than a permanent shift.

Third, the temporal window for a potential tipping point is highly uncertain. The Ditlevsen 2023 estimate provides a 70-year range from 2025 to 2095. This span is too wide to distinguish between a threshold crossing in five years and one in 75 years, meaning the current data confirms directionality but not urgency.

Fourth, the mechanistic chain relies on Greenland melt as the primary driver, yet observational studies indicate that deep convection in the Labrador Sea - the primary AMOC pump - is more sensitive to local wind patterns and the North Atlantic Oscillation (NAO) phase.

Finally, the projected catastrophic consequences, such as 3-8°C European cooling, are derived from abrupt collapse scenarios in climate models. These do not necessarily follow from the gradual weakening currently observed. A slow weakening at current rates would produce a much smaller warming offset of approximately 0.5-1.5°C.

While the direction of the signal is consistent across multiple studies, the magnitude of the trend and the precise timing of any threshold crossing remain unquantified due to these instrumental and temporal limitations.