A volcano erupts somewhere in the tropics. Sulfate aerosols blast into the stratosphere. Within months, the monsoon over South Asia falters and harvests fail. That much has been documented for years. What was not understood until now is exactly how that same eruption simultaneously drains rainfall from northern East Asia, thousands of kilometres away, through an atmospheric pathway that has been operating for at least a thousand years without anyone knowing it was there.
A study published in Nature Communications by researchers from Wuhan University and the University of Tokyo has traced that pathway in full. Large tropical volcanic eruptions do not just weaken the monsoon over the Indian subcontinent. They trigger a chain reaction high in the atmosphere that reaches across the entire width of Eurasia, producing coordinated, simultaneous droughts over two of the most densely populated agricultural regions on Earth, often within the same summer.
The mechanism runs through a circulation pattern called the circumglobal teleconnection, known as the CGT. This is a wave of atmospheric energy that travels along the upper-tropospheric jet stream, circling the Northern Hemisphere in a slow, undulating pattern. Meteorologists have long recognised it as a major driver of summer rainfall variability across Eurasia. What this paper establishes is that tropical volcanic eruptions can force that wave into a specific damaging configuration from the outside, independent of ocean temperatures or internal climate variability.
The process starts with aerosols. When a large tropical eruption sends sulfate into the stratosphere, it raises the planet’s reflectivity and reduces how much solar energy reaches the surface. The land and ocean cool, the temperature contrast that powers South Asian monsoon convection weakens, and the towering cloud systems that carry rainfall inland lose intensity. Precipitation drops across the subcontinent.
That weakening of convection removes a major heat source from the atmosphere. Deep tropical convection normally releases enormous amounts of energy as water vapour condenses into cloud and rain. When it shuts down, the surrounding atmosphere loses that input and responds by generating a Rossby wave, an undulation that propagates outward and eastward along the jet stream. That wave settles into the negative phase of the CGT, placing anomalous northerly winds over East Asia high in the troposphere. Those northerly winds drive strong downward air motion over northern East Asia. Subsiding air suppresses cloud formation and rainfall. The monsoon rains that normally water northeastern China and surrounding regions arrive weakened or fail to arrive at all.
Three independent lines of evidence point to the same conclusion. Tree-ring records extending back to 1657 were used to reconstruct how the CGT behaved in summers following major eruptions. The mean anomaly recorded in the first post-eruption summer sat at minus 0.74. Climate model simulations from the Community Earth System Model Last Millennium Ensemble, covering 850 to 2005 across 13 ensemble members and 91 eruption events, produced a mean anomaly of minus 0.59 in the same window. Idealised experiments using a linear baroclinic model, where diabatic heating over South Asia was artificially reduced, generated the same wave train structure within 25 days.
To rule out El Niño and the Indian Ocean Dipole as the real drivers, the team isolated eruptions that occurred during neutral conditions for both ocean patterns. The negative CGT response and the coordinated drought pattern appeared regardless. After statistically removing those ocean signals from the data entirely, the volcanic fingerprint remained intact. The wave train is not a product of ocean state. It is a direct atmospheric response to reduced tropical heating.
Tree rings tell the historical side of this story in ways that model runs cannot. The Tambora eruption of 1815, the largest in recorded history, produced documented famines across the Indian subcontinent and severe drought in northern China. The 1991 Pinatubo eruption left measurable rainfall deficits across both regions. Across every large tropical eruption in the dataset going back to 850 CE, the same pattern appears. Two distant regions dry out together, in the same summer, driven by the same upper-atmospheric wave.
The drought signal peaks in the first boreal summer after an eruption and fades through the second year. Model simulations show a slight extension into year two, attributed to the thermal inertia of the ocean-atmosphere system. Tree-ring signals in the second and third years remain statistically weak, likely reflecting biological memory in tree growth rather than any sustained atmospheric response.
Previous attempts to forecast post-volcanic drought leaned heavily on ENSO-based pathways, which are inconsistent across events and sensitive to initial ocean conditions. The CGT pathway operates regardless of what the Pacific or Indian Ocean is doing at the time of an eruption. When El Niño or the Indian Ocean Dipole do coincide with an eruption, the drought anomalies reflect a straightforward addition of both signals rather than any interaction between them. The volcanic component remains identifiable and consistent.
South Asia and northern East Asia together feed a large share of the global population. Simultaneous crop failures across both regions in a single summer is not a remote possibility. According to a millennium of proxy records and climate model output, it has happened repeatedly, driven each time by the same chain of atmospheric events now laid out in full.
Future large tropical eruptions will interact with a jet stream already being reshaped by warming ocean temperatures and shifting aerosol concentrations. That evolving background state could push the CGT response in directions that historical patterns do not anticipate. Above The Norm News will be watching closely and report if needed.
Source:
Nie, W., Xia, J., Kino, K., She, D., & Oki, T. (2026). Tropical volcanism triggers pan-Asian monsoon droughts via circumglobal teleconnection. Nature Communications, 17, 2701. https://doi.org/10.1038/s41467-026-70710-x
