For decades, climatologists have sought to determine whether the prominent interdecadal oscillations in sea surface temperatures (SSTs) found across the Atlantic and Pacific Oceans are forced internally by climate dynamics or externally through greenhouse gas emissions and volcanic eruptions.
This dilemma is examined in a new paper by researchers at Florida State University. Using Rotated Low‑Frequency Component Analysis, they applied it to very large ensembles of climate models, in conjunction with observational data, to extract externally forced, or “forced,” signals attributable to external drivers from patterns generated internally by natural variability.
The AMV was found to be largely forced across multiple climate models (CESM1, CESM2, GFDL‑CM3, E3SMv2). So, near loss of memory means that over long time scales, changes in Atlantic temperatures can be largely understood as externally forced by increasing concentrations of greenhouse gases or volcanic eruptions everywhere but the North Atlantic Ocean here. Indeed, observational records also confirm this; the AMV is practically entirely forced externally.
The Pacific, however, tells a different story. The multidecadal variability of the Pacific Decadal Oscillation (PDO) is dominated by its substantial internal variability, suggesting that PDO fluctuations are driven more by intrinsic than forced dynamics of the ocean‑atmosphere system.
Analyzing how climate change is affecting plant and animal life in the Atlantic
Assistant Professor of meteorology Michael Diamond said, “We know that important sources of natural variability in Earth’s climate system exist, and our ability to distinguish between these natural and human-forced sources of temperature variability is key to projecting future temperatures and their related impacts on society.”
This will sharpen attribution of past climate changes and inform future projections. Because of its more forced variability, the Atlantic may be highly predictable for some emissions scenarios. While the Pacific will continue to make seasonal predictions more difficult, due to its internal variability continuing to introduce prolonged uncertainty in climate outlooks.
El Niño and La Niña are common climate patterns in the tropical Pacific that oscillate on a timescale of a few years. In contrast, the Pacific Decadal Oscillation (PDO) has an orders-of-magnitude slower phase-shift timescale of 20 to 30 years.
To explore this long‑term climate cyclical behavior, the authors leveraged a Python programming approach and an innovative statistical technique, Rotated Low‑Frequency Component Analysis (RLFCA).
This method relies on a technique developed to detect temperature change trends by their rate of evolution. The RLFCA moves these patterns toward externally forced reference modes, isolating signals driven by external forcing (e.g., greenhouse gases or volcanoes) from those generated by fluctuations in the intrinsic climate system.
Using RLFCA and a hundred years of climate model data (1920–2025), the research team created new, detailed records on how long-term Pacific temperature patterns evolve.
FSU meteorology graduate alumnus Anthony Freveletti said, “Since human emissions build up in the atmosphere over many years, the temperature changes they cause develop gradually over time. In contrast, natural fluctuations driven by factors such as ocean currents, wind patterns, and air pressure occur more rapidly. Our analysis effectively separates these forced and unforced changes within those data trends by identifying which patterns are fast-evolving and which are slow-evolving.”
Freveletti built on that approach with a “rotational” step in which the identified patterns are reordered by known external influences, as calculated by climate models, allowing us to better understand the causes of temperature variability.
The analysis, however, indicated that what had previously been viewed as a natural variability in the Atlantic Ocean was actually an overlap of two forced phenomena: greenhouse gas emissions generating global warming and anthropogenic aerosols shading and cooling the surface ocean.
“Our results show a complex interplay of air pollution and greenhouse gas emissions is responsible for historical temperature patterns in the Atlantic Ocean that led to various weather phenomena, such as a spike in hurricane frequency since 1990,” Diamond said. “We should not expect to return to an inactive hurricane era by chance alone; the future of human emissions will be the most important driver of Atlantic temperatures going forward.”
Journal Reference:
- Anthony S. Freveletti, Michael S. Diamond, Robert C. J. Wills. Multidecadal SST Variability Assessed as Primarily Forced in the Atlantic and Internal in the Pacific Using Rotated Low-Frequency Component Analysis. Geophysical Research Letters. DOI: 10.1029/2F2025GL121516
