ENSO events, which include El Niño and La Niña phases, influence weather patterns around the globe. While these events typically peak around December and can usually be forecast with confidence around six months in advance, predicting them earlier has long been a challenge. To address this challenge, Australian scientists at CSIRO and the University of Melbourne analysed more than 7,500 historical ENSO forecasts from climate models developed in nine countries.

This meta-analysis focused on the relationship between the starting state of the Pacific Ocean and the subsequent accuracy of the forecast one year later. The study found that forecasts initiated during La Niña conditions consistently performed well even when predicting ENSO events a year ahead. In contrast, forecasts beginning from neutral conditions were far less reliable.

Co-author Dr. Mandy Freund from the University of Melbourne and researcher with the Australian Research Council Centre of Excellence for 21st Century Weather, stated that the findings provide a new way to assess confidence in long-range climate forecasts. She noted that the consistent performance of one-year forecasts starting from La Niña suggests there is something about these conditions that makes the tropical Pacific system more predictable. This insight allows forecasters to identify which specific predictions deserve greater confidence, potentially enabling earlier warnings for El Niño events.

Lead author Dr. Nandini Ramesh, Senior Research Scientist at CSIRO and Adjunct Research Fellow at UNSW Sydney, explained that while it has been known for some time that starting conditions affect how well model forecasts perform, the study shows how drastic those differences can be. Dr. Ramesh emphasized that the researchers can now use this information to identify which forecasts deserve greater confidence. This capability could help improve how early warnings are communicated to the public, allowing for more effective preparation for climate impacts.

The researchers also found that different forecasting approaches may be suited to different climate conditions. Traditional physics-based models struggled when forecasts began from neutral conditions, whereas machine learning and statistical models may have potential to improve performance in these situations. The findings highlight opportunities to strengthen early-warning systems, although further research is needed before the approach can be incorporated into operational forecasting systems.

Dr. Ramesh noted that they are still a few steps away from applying the findings to the models used in practice by forecasting agencies. However, the results hold promise for being able to communicate early warnings for future events more effectively. The research offers new insights into the drivers of ENSO predictability and could ultimately help improve preparedness for droughts, floods and other climate impacts associated with El Niño and La Niña events.

Subject of Research: Predictability of El Niño-Southern Oscillation based on initial Pacific Ocean conditions

Article Title: Scientists find key to predicting El Niño 12 months ahead

Article References: Scientists find key to predicting El Niño 12 months ahead. Original publication (Bibliographic details unavailable.)

Image Credits: AI Generated

DOI: Not provided

Keywords: El Niño, La Niña, ENSO, Climate Forecasting, Pacific Ocean, CSIRO, University of Melbourne, Weather Patterns, Science Advances, Climate Models, Scientists, find

The research, led by Robert S. Powell of the University of Cambridge with co-authors at Harvard University, Tulane University, and several other institutions, quantifies the spatial pattern of sea-level change, known as a sea-level fingerprint, produced by mass loss from the Greenland and Antarctic ice sheets and from glaciers worldwide through the year 2100. The team’s central finding is stark: for Small Island Developing States in the Pacific and Indian Oceans, the ice-melt contribution to local sea-level rise exceeds the global average by up to roughly 24 percent. Crucially, this amplification emerges regardless of whether the meltwater originates from Greenland, from Antarctica, or from the global glacier system.

To understand why the location of an island matters as much as the volume of melting ice, it helps to revisit a counterintuitive principle of geophysics. The ocean is not a bathtub that fills uniformly. Ice sheets and glaciers are so massive that their gravitational pull draws water toward them, piling the sea into a subtle bulge in their vicinity. When the ice melts, that gravitational attraction weakens, and water flows away from the shrinking ice mass toward the far side of the planet. The result is a distinctive pattern: sea level actually falls close to major ice losses even as it rises much faster than average thousands of kilometers away.

This redistribution is captured by what scientists call gravitational, rotational, and deformational effects, often abbreviated as GRD. Gravity, as described above, drives the migration of water away from melting ice. Rotation matters because moving water shifts the distribution of mass on Earth’s surface, nudging the planet’s rotation axis and altering the shape of the sea surface through changes in the centrifugal force. Deformation enters because the solid Earth itself responds: the crust flexes under the changing loads of ice and water, and this shifting topography further reshapes the ocean basin. Together, these three mechanisms produce a highly structured map of sea-level change, with some regions experiencing far more than the global mean and others far less.

The fingerprints are normalized in the study, meaning the researchers express local sea-level change as a ratio to the global mean value for each melt source. A normalized fingerprint greater than one indicates a location where that particular ice loss raises local sea level faster than average. By computing these fingerprints for Greenland, Antarctica, and the world’s glaciers, and then applying ensembles of ice-sheet and glacier mass-balance projections, the team could estimate how the ice-melt component of sea-level rise at specific island locations would evolve through the end of the century under the projected melting of each source.

What makes the new result remarkable is its robustness. One might expect that the geography of amplification would depend heavily on which ice reservoir dominates future melt. Greenland’s fingerprint, for instance, is strongest in the far Pacific, while Antarctic loss tends to raise sea level most dramatically in the Northern Hemisphere. Yet the analysis shows that Small Island Developing States in the tropical Pacific and Indian Oceans fall within regions of amplified sea-level rise for all three melt sources simultaneously. The tropical oceans, lying far from all major ice masses, occupy a zone where the fingerprints of Greenland, Antarctic, and glacier melt all exceed unity. The amplification is therefore source-independent: whatever combination of ice loss unfolds over the coming decades, these islands are locked into a geography of above-average rise.

The implications for the affected nations are profound. Small Island Developing States, a group recognized by the United Nations and including countries such as those across the Pacific and Indian Ocean basins, are among the world’s most climate-vulnerable nations, with low-lying infrastructure, economies tied closely to the coastline, and limited capacity for engineered defenses. A systematic excess of up to about 24 percent over the global mean means that planning exercises based on global-average sea-level projections would systematically underestimate the hazard these countries face. Coastal inundation frequencies, saltwater intrusion into freshwater lenses, and the viability of critical infrastructure all scale with local rather than global numbers, so the difference between an average projection and an amplified one can translate into materially different adaptation decisions.

Methodologically, the study represents a synthesis of two strands of climate science that are often pursued separately. Ice-sheet and glacier models produce projections of mass balance, the balance between accumulation and loss of ice, expressed as ensembles that capture uncertainty in future emissions, ice dynamics, and model structure. Sea-level fingerprinting converts those mass changes into spatially varying patterns of ocean height through the GRD framework. By combining ensembles of mass-balance projections with fingerprint calculations, the researchers could propagate uncertainty from the ice models through to the local sea-level estimates at island locations, producing a probabilistic picture of the amplification rather than a single deterministic value.

The persistence of the hotspots through 2100 is another key element of the finding. Because the fingerprint pattern depends primarily on the location of the ice loss rather than its magnitude, the normalized amplification at a given island remains comparatively stable even as the absolute rates of sea-level rise accelerate. In other words, the islands’ disadvantage is not a transient feature of early-century melt but a durable property of the planet’s response to ice loss. As total meltwater volumes grow, the absolute excess experienced by these nations grows with them, compounding the exposure of coastlines that are already on the front line of climate change.

The study also carries a broader message about how sea-level projections should be communicated and used. Global mean sea level is a convenient summary statistic, but it is a poor guide to local risk. The gravitational, rotational, and deformational physics that produce fingerprints are now well established, and applying them routinely to national and regional assessments would sharpen the picture of who is exposed and by how much. For the Small Island Developing States of the Pacific and Indian Oceans, the new analysis confirms that their geography places them in a systematic and source-independent zone of amplified rise, a finding that turns a subtle geophysical principle into a concrete matter of climate justice and adaptation planning. As the world’s ice continues to retreat, the seas will not rise evenly, and the places that contributed least to the problem will, once again, feel more than their share of it.

Subject of Research: Sea-level fingerprints of global ice loss and their amplified impact on Small Island Developing States

Article Title: Global ice loss drives source-independent amplification of sea-level rise at Pacific and Indian Ocean Small Island Developing States

Article References: Powell, R. S., Cederberg, G., Dangendorf, S., Lucas, E., Bao, X., Creveling, J. R., Oelsmann, J., Tsuji, S. R. J., Tsuji, L., & Mitrovica, J. X. (2026). Global ice loss drives source-independent amplification of sea-level rise at Pacific and Indian Ocean Small Island Developing States. Communications Earth & Environment. https://doi.org/10.1038/s43247-026-04132-4

Image Credits: AI Generated

DOI: 10.1038/s43247-026-04132-4

Keywords: sea-level rise, sea-level fingerprints, Small Island Developing States, Greenland ice sheet, Antarctic ice sheet, glacier melt, gravitational effects, climate change, Pacific Ocean, Indian Ocean, ice loss, coastal vulnerability

The two patterns at the heart of the study are the monsoon’s fingerprints of the El Niño-Southern Oscillation, or ENSO, in the Pacific, and the Indian Ocean Dipole, or IOD, in the Indian Ocean. El Niño events, marked by unusually warm sea surface temperatures in the equatorial eastern Pacific, have long been known to suppress Indian rainfall, while La Niña conditions tend to enhance it. The IOD, a see-saw of sea surface temperatures between the western and eastern Indian Ocean, similarly modulates the monsoon, with positive dipole events generally favouring above-normal rain. What has remained poorly understood is how these two influences combine, particularly during the recent decades of accelerating ocean warming.

To disentangle the two signals without imposing any preconceived filter on the data, the team applied a data-adaptive statistical technique called multi-channel singular spectrum analysis, or MSSA. The method decomposes daily rainfall and satellite-measured outgoing longwave radiation into space-time patterns, extracting oscillations and persisting structures directly from the observations. Because the rainfall data cover only India’s land area, the researchers first identified the seasonally persisting modes in outgoing longwave radiation, which spans both land and ocean, and then projected the century-long rainfall record onto those modes. This procedure yielded two clean components, dubbed M-ENSO and M-IOD, each retaining the same sign of anomaly through most of the June-to-September season but varying from year to year.

The results are striking. The sum of the two modes correlates at 0.83 with the observed seasonal rainfall anomaly across the full 1901 to 2021 period, meaning that the combined influence of the Pacific and Indian oceans explains the bulk of the year-to-year variation in India’s summer rain. Individually, the ENSO-related mode correlates at 0.55 and the IOD-related mode at 0.6 with the total anomaly. Crucially, the two modes sometimes reinforce one another and sometimes cancel out. In 1983, both modes were positive and the season ended with a strong rainfall anomaly of 0.89 millimetres per day; in 1972, both were negative, producing a weak monsoon at minus 1.62 millimetres per day. In 1989 and 1997, the modes opposed each other and the seasonal rainfall came out near normal.

The year 1997 stands out as a dramatic illustration of this destructive interference. That year brought the strongest El Niño of the twentieth century, an event that on its own would likely have produced severe drought across India. Instead, a coincident positive IOD event counteracted the Pacific influence, and the seasonal rainfall anomaly finished at just minus 0.08 millimetres per day, essentially normal. Over the full record, the two modes were jointly positive in 32 years, yielding an average anomaly of 0.48 millimetres per day, and jointly negative in 37 years, averaging minus 0.68 millimetres per day. Years of opposing phases, 52 in total, produced near-normal seasons regardless of which mode dominated.

Beneath this interannual drama, the study uncovers a longer-term reorganisation of the monsoon system. Splitting the record into two epochs, 1901 to 1960 and 1961 to 2021, the researchers found that sea surface temperatures in both basins showed little or no warming trend in the first epoch but steep warming in the second. The Indian Ocean warmed faster than the Pacific, consistent with the assessment of the Intergovernmental Panel on Climate Change that the Indian Ocean and western Pacific have warmed faster than the global average. The variance of the Niño-3.4 index, a standard measure of ENSO, increased from 0.29 to 0.42 square kelvin between the epochs, while the variance of the dipole mode index rose from 0.08 to 0.12 square kelvin, signalling more energetic variability in both oceans.

The rainfall itself followed suit. Nonlinear trend analysis, performed with ensemble empirical mode decomposition and independently confirmed by the MSSA method, reveals that the all-India monsoon rainfall was dominated by strong positive anomalies before 1960 and by predominantly negative anomalies with an almost flat trend after 1960. The two ocean-driven modes shifted in opposite directions: the ENSO-related rainfall mode, which had an increasing trend in the first epoch, settled into a near-flat but positive regime in the second, while the IOD-related mode flipped from positive anomalies in the first epoch to a pronounced negative trend in the second. The relative strength of the two modes, rather than the behaviour of either alone, therefore largely determines the long-term character of the monsoon.

The spatial fingerprints of the two epochs also differ in revealing ways. Correlation maps show that the ENSO-monsoon relationship has remained essentially stable across the century, with the familiar Pacific pattern of sea surface temperature anomalies linked to Indian rainfall in both epochs. The IOD-monsoon relationship, by contrast, was weak and ill-defined before 1960 but became stronger and better organised after 1960, particularly when the IOD and ENSO influences opposed each other. In the second epoch, positive rainfall anomalies became associated with a clear positive dipole pattern in the Indian Ocean, whereas in the first epoch the corresponding dipole signature was weakly negative. The authors attribute this reorganisation to the anomalous warming of the Indian Ocean, which has significantly altered the intensity and phases of the IOD mode and its associated rainfall over the past six decades.

The implications reach well beyond the archive of historical gauges. India has issued seasonal monsoon forecasts for more than a century, first with statistical regression models that ultimately showed little skill and more recently with dynamical global models that still perform poorly. The authors argue that a central reason for this shortfall is the inadequate representation of the oceans in prediction systems. Their findings suggest that reliable seasonal forecasts and long-term projections demand coupled ocean-atmosphere models capable of correctly simulating sea surface temperatures over both the Pacific and the Indian Ocean, including the Arabian Sea, and of capturing the constructive and destructive interference between the two basins’ influences.

The study also leaves open a compelling question for future research: why did the epochal shift occur around 1960, and how will the monsoon respond as the differential warming of the two oceans continues through the twenty-first century? With ocean heat content expected to keep rising, and with the livelihoods of more than a billion people hanging on the answer, the framework developed by Krishnamurthy and Shukla provides a unified way to read the monsoon’s past and a sharper lens for anticipating its future. What emerges is a monsoon not governed by a single oceanic puppeteer, but by a delicate duet between two warming basins, one whose harmony, and discord, may now be changing with the climate itself.

Subject of Research: Long-term variability of Indian summer monsoon rainfall linked to ENSO and the Indian Ocean Dipole under differential ocean warming

Article Title: Long-term variability of Indian monsoon rainfall related to warming of Indian and Pacific oceans

Article References: Krishnamurthy, V., & Shukla, J. (2026). Long-term variability of Indian monsoon rainfall related to warming of Indian and Pacific oceans. Climate Dynamics, 64(11), Article 451. https://doi.org/10.1007/s00382-026-08409-4

Image Credits: AI Generated

DOI: 10.1007/s00382-026-08409-4

Keywords: Indian monsoon, ENSO, Indian Ocean Dipole, sea surface temperature, ocean warming, climate variability, seasonal prediction, Climate Dynamics, monsoon rainfall, Pacific Ocean, Indian Ocean, climate change