Global warming is causing a decline in marine oxygen, driven by changes in ocean stratification and oxygen solubility. This decline has significant implications for marine ecosystems, including impacts on biological productivity, habitat sustainability, and the ocean's carbon sequestration capacity. Oxygen minimum zones (OMZs) are expanding rapidly, affecting nitrogen cycling and greenhouse gas emissions. The North and Tropical Pacific, with its vast OMZs, accounts for a substantial portion of this oxygen loss. Multi-decadal climate oscillations, such as the Pacific Decadal Oscillation (PDO) and the AMO, are known to modulate oxygen concentration trends, but instrumental records are too short to fully understand their long-term impact. This study leverages annually resolved sedimentary records to reconstruct past changes in deoxygenation, thereby assessing natural variability and improving future prediction capabilities. The Gulf of California, located within the North Pacific OMZ, offers annually varved sediments ideal for high-resolution reconstructions of denitrification, a key indicator of deoxygenation.

Previous research has linked the decline in oxygen and increase in denitrification in the North Pacific OMZs to a shift in the PDO. Similar trends are observed between oxygen anomalies in the Eastern Tropical North Pacific (ETNP) and the AMO index. However, these instrumental records are limited, hindering a comprehensive understanding of the relationship between multidecadal climate oscillations and marine deoxygenation. While the PDO has been suggested as the primary driver of natural oxygen variability in the Pacific OMZ, primarily based on model simulations, the potential role of the AMO has been largely ignored. Studies have explored various mechanisms by which the PDO might influence mid-depth oxygen, including isopycnal movements, trade wind shifts, and large-scale changes in oxidant demand. However, these model simulations have often neglected the potential contribution of the AMO.

The study employed annually varved sediments from the Gulf of California (GoC) to reconstruct denitrification changes over the last glacial-interglacial cycle. Ten sections, each spanning ~200 years, were subsampled at annual resolution and analyzed for nitrogen isotopes (δ¹⁵N), a proxy for denitrification intensity, and biogenic silica, an indicator of productivity. Spectral analyses (Blackman Tukey and Maximum Entropy methods) were performed on both filtered (removing periodicities >70 years) and unfiltered denitrification records to identify dominant periodicities. Wavelet analysis was used to determine non-stationary periodicities. Additionally, biogenic silica accumulation rates were reconstructed and analyzed to assess the relationship between productivity and deoxygenation. Finally, modern sea surface temperature (SST) and subsurface circulation data were regressed against positive AMO and PDO phases to investigate the mechanistic links between climate oscillations and oxygen supply to the North Pacific OMZ.

Spectral analyses of the annually resolved denitrification records revealed dominant periodicities of 25 years (similar to the PDO) and a secondary periodicity of 10 years (similar to the NPGO). Analyses of unfiltered records identified periodicities around 83 and 31-35 years, consistent with the AMO. These AMO-like periodicities were most prominent during warm climatic periods (Holocene and Bølling-Allerød). During cold periods (Last Glacial Maximum, Heinrich Event 1), different modes of multidecadal variability were observed, suggesting a fundamental difference in the AMO during cold phases. Cross-correlation and coherence analyses showed no significant relationship between denitrification and biogenic silica accumulation, indicating that local upwelling-induced productivity does not primarily control oxygen levels in the GoC. Band-pass filtering analyses further supported this finding, showing that longer-period (PDO- and AMO-like) oscillations accounted for a greater proportion of denitrification variability compared to shorter-period (ENSO- and NPGO-like) oscillations. Reanalyses of SST and subsurface circulation data revealed a strong correlation between the AMO and the Equatorial Under Current (EUC) velocity. Positive AMO phases correlated with a slowdown of the eastward EUC, reducing oxygen transport to the eastern Pacific. The PDO showed a less clear impact on the EUC, though it might act synergistically with the AMO to modulate oxygen levels. A long-term comparison between AMOC proxy, denitrification and EUC further supports the finding that AMO has a greater influence on deoxygenation than previously thought.

The findings demonstrate that the AMO and PDO are significant drivers of deoxygenation in the ETNP, with a minor influence from local biological productivity. The AMO's impact on oxygenation is greater than previously recognized. The AMO influences the EUC strength and hence the oxygen transport to the ETNP. The PDO, while showing marginal impact on the large-scale zonal subsurface circulation, can act synergistically or antagonistically with the AMO. The study highlights a potential global control on marine oxygen at multidecadal timescales originating from the North Atlantic, linking changes in the AMO with changes in the Indian monsoon, atmospheric circulation, and upper ocean circulation in the Pacific. This inter-basin teleconnection underscores the complexity of factors influencing oxygen levels and the need for a holistic perspective in future predictions.

This study reveals the significant role of the AMO in controlling deoxygenation in the North Pacific, exceeding that of the previously emphasized PDO. The AMO's influence on the EUC and atmospheric circulation patterns impacts oxygen supply to the ETNP. Future research should focus on refining models to incorporate the interactive effects of the AMO and PDO on subsurface circulation and oxygen transport, considering the implications of ongoing climate change and the potential for synergistic effects with global warming to exacerbate deoxygenation.

The study focuses on a specific region within the North Pacific OMZ and the findings may not be generalizable to other OMZs. The interpretation of δ¹⁵N as solely reflecting denitrification intensity during all periods might be an oversimplification, especially in the cold periods where other processes could influence nitrogen isotopic compositions. Finally, while the study provides strong evidence for the link between AMO and deoxygenation, establishing direct causality requires further investigation.