In the vast catalogue of atmospheric phenomena, few are as visually dramatic or dynamically complex as the Fujiwara effect — the interaction between two tropical cyclones that come close enough to influence each other's motion. Named after the Japanese meteorologist Sakuhei Fujiwhara, who first described the phenomenon in 1921, the Fujiwara effect occurs when two cyclones approach within approximately 1,400 kilometres of each other and begin to rotate around a common centre, like two dancers in a slow, destructive waltz. The interaction can result in one cyclone absorbing the other, both cyclones being deflected from their original tracks, or the two merging into a single, larger storm — outcomes that are among the most difficult to forecast in tropical meteorology and that have produced some of the most unusual and destructive storm tracks in history.
TL;DR: The Fujiwara effect occurs when two tropical cyclones approach within approximately 1,000–1,400 km of each other and begin interacting through their mutual circulation. The cyclones rotate around a common centre (the centroid of their combined vorticity), and the interaction can result in orbit (mutual rotation without merger), partial stripping (the larger storm absorbs moisture and vorticity from the smaller one), complete merger (the two systems combine into one), or elastic interaction (the storms influence each other's track before separating). The effect was first described by Sakuhei Fujiwhara in 1921 and remains one of the most challenging forecast scenarios in tropical meteorology.
1,400 kmMaximum distance at which the Fujiwara effect can begin
1921Year Sakuhei Fujiwhara first described the interaction
~1–3/yearAverage frequency of Fujiwara interactions in the western Pacific
12–24 hrsTypical duration of a complete Fujiwara merger
Fujiwhara's Discovery: Vortices in a Water Tank
Sakuhei Fujiwhara, working at the Central Meteorological Observatory in Tokyo in the early twentieth century, first observed the interaction between paired vortices not in the atmosphere but in a water tank. By creating two vortices in a controlled experimental setup, he observed that they did not simply coexist independently but interacted — rotating around each other, approaching, and eventually merging when close enough. He published his findings in 1921, noting the analogy to the behaviour of typhoons in the western Pacific, where multiple storms occasionally formed simultaneously and appeared to influence each other's paths.
Fujiwhara's insight was that the interaction between two vortices is not a minor perturbation but a fundamental change in the dynamics of both systems. Each cyclone generates a circulation that extends far beyond its visible cloud structure — the outer flow of a mature tropical cyclone influences the atmosphere hundreds of kilometres from its centre. When two such circulations overlap, each cyclone finds itself embedded in the flow field of the other, and its motion is no longer determined solely by the large-scale environmental steering flow but by the combined influence of the environment and the neighbouring cyclone. This mutual influence is the essence of the Fujiwara effect: neither storm moves independently when the other is nearby.
Fujiwhara identified several possible outcomes of the interaction, which subsequent research and observation have confirmed and refined. The simplest is mutual orbit — the two cyclones rotate around their common centroid (the centre of mass of their combined vorticity) without significant changes in intensity or structure. If the cyclones are of unequal strength, the smaller one orbits more rapidly around the larger one (just as a small mass orbits closer to a larger mass in gravitational interactions). More complex outcomes — including partial absorption, complete merger, and track deflection — depend on the relative sizes, intensities, distances, and environmental conditions of the two storms.
The Mechanics: How Two Storms Dance
The Fujiwara interaction is governed by the same fluid dynamics that describe the interaction of any two vortices in a rotating fluid. Each tropical cyclone is a large-scale atmospheric vortex with cyclonic (counterclockwise in the Northern Hemisphere) rotation. The outer circulation of each cyclone creates a flow that advects (carries along) the other cyclone. When two cyclones are separated by a distance comparable to their outer circulation radii — roughly 500–1,400 kilometres, depending on the size of the storms — this mutual advection becomes the dominant force controlling their motion.
The centroid of the interaction — the point around which the two cyclones orbit — is determined by the relative vorticity (rotational strength) of each storm. If the two storms are of equal intensity, the centroid is midway between them, and both orbit at equal rates. If one storm is significantly stronger, the centroid shifts toward the stronger storm, and the weaker storm orbits around the stronger one while the stronger storm remains relatively stationary — analogous to the way a planet orbits a star rather than both orbiting a common centre at equal distance. This asymmetry has important forecasting implications: the stronger storm's track is less affected by the interaction, while the weaker storm's track can be dramatically altered.
As the interaction progresses, the two cyclones typically spiral closer together. The mechanism for this inward spiral involves the mutual deformation of each cyclone's wind field: the shearing effect of one cyclone on the other disrupts the outer circulation, reduces the angular momentum of the system, and allows the storms to approach. If the storms approach close enough — typically within 300–500 kilometres — one of three things happens: the weaker storm is torn apart and absorbed by the stronger one (the most common outcome), the two storms merge into a single, larger cyclone (less common but observed), or the interaction causes one or both storms to accelerate away from the interaction zone as the complex dynamics of the paired vortices produce an outward trajectory (the least common outcome, sometimes called elastic interaction).
Notable Fujiwara Events in History
The western Pacific, which produces more tropical cyclones than any other basin, is the primary theatre for Fujiwara interactions. The region's prolific storm production means that multiple typhoons frequently coexist in the basin, and interactions occur an average of one to three times per year. Typhoon Pat and Typhoon Ruth in 1994 provided one of the most clearly observed Fujiwara interactions: the two storms orbited each other for approximately 36 hours before Pat, the weaker system, was absorbed into Ruth, which subsequently intensified and struck Japan.
In the Atlantic basin, Fujiwara interactions are less common but not unknown. Hurricane Iris and Tropical Storm Karen in 1995 interacted in the central Atlantic, with Karen being absorbed by the larger circulation of Iris. More dramatically, the 2017 Atlantic season produced a near-Fujiwara interaction between Hurricanes Irma and Jose, which tracked through the same region of the Caribbean and western Atlantic within days of each other. While the two storms did not undergo a classic Fujiwara merger (their separation was generally too large for full interaction), the outer circulation of Irma influenced Jose's track, steering it northward and away from the U.S. coast — an outcome that, paradoxically, may have prevented a second catastrophic landfall.
One of the most unusual Fujiwara events in recent history occurred in the western Pacific in 2009, when Typhoon Parma and Typhoon Melor interacted in a complex dance near the Philippines. Parma, which had already struck the Philippines once, was deflected back toward the islands by the approach of Melor, causing a second landfall that compounded the devastation from the first. The interaction was poorly forecast by operational models, which struggled with the complexity of two strong cyclones interacting in a constrained geographic environment — demonstrating the forecasting challenges that Fujiwara interactions present even with modern numerical weather prediction.
The Role of Ocean Temperature and Atmospheric Conditions
The outcome of a Fujiwara interaction depends not only on the properties of the two storms but on the oceanic and atmospheric environment in which the interaction occurs. Sea surface temperatures determine the energy available to both storms during their interaction: warmer waters provide more latent heat fuel, which can sustain or strengthen both systems during a prolonged interaction that might otherwise weaken them through mutual disruption. In environments with marginal sea surface temperatures, the energy demand of two simultaneously intensifying cyclones can exhaust the available heat content, weakening one or both storms even without the mechanical disruption of the interaction itself.
Vertical wind shear — the change in wind speed and direction with altitude — affects how each storm responds to the proximity of the other. In high-shear environments, the disruption of the eyewall structure by the neighbouring storm's circulation is compounded by the tilting effect of the environmental shear, making weakening or dissipation of the smaller storm more likely. In low-shear environments, both storms can maintain their vertical structure more effectively during the interaction, making prolonged mutual orbit and eventual merger (rather than rapid absorption) more likely. The environmental context transforms the Fujiwara effect from a simple two-body problem into a multi-factor interaction whose outcome depends on variables that extend well beyond the two storms themselves.
Forecasting Challenges: Why Two Storms Are Harder Than One
Fujiwara interactions represent one of the most challenging forecast scenarios in tropical meteorology. The difficulty arises from several compounding factors: the interaction adds degrees of freedom to the forecast problem (the motion of each storm depends on the other's position and intensity, creating a coupled system), small errors in the initial position or intensity of either storm are amplified by the interaction (the sensitivity to initial conditions is enhanced), and the numerical models that predict storm motion must resolve the structure and circulation of both storms simultaneously — a requirement that stretches the resolution and physics of even the best modern models.
Track forecast errors during Fujiwara interactions are typically 50–100 percent larger than for isolated tropical cyclones at the same forecast lead time. The errors are not random but systematic: models tend to either underestimate the degree of interaction (predicting that the storms will pass without significant mutual influence) or overestimate it (predicting merger when the storms instead separate). The threshold distance at which the interaction begins, the rate at which the storms spiral together, and the outcome of the interaction (merger versus separation) are all sources of uncertainty that compound with the normal uncertainties of tropical cyclone forecasting.
Intensity forecasting is further complicated because Fujiwara interactions can either weaken or strengthen the component storms depending on the specific dynamics. The shearing effect of one storm's outer circulation on the other can disrupt the eyewall and weaken the storm, while the merger of two systems can concentrate vorticity and produce a post-merger storm that is more intense than either predecessor. The 2012 interaction between Typhoons Saola and Damrey in the western Pacific illustrated this complexity: Saola weakened during the interaction phase but re-intensified after absorbing moisture and vorticity from Damrey's remnants — a sequence that the operational forecast models handled with varying degrees of success.
The Fujiwara effect is not limited to tropical cyclones — it is a fundamental property of interacting vortices in any fluid, and examples can be found across scales from laboratory experiments to galactic dynamics. In the atmosphere, extratropical cyclones (mid-latitude low-pressure systems) can interact in Fujiwara-like fashion, with two lows orbiting each other or merging into a single, deeper low. These interactions are common in the North Atlantic and North Pacific storm tracks and are routinely handled by numerical weather prediction models, though they contribute to forecast uncertainty in the same way that tropical Fujiwara interactions do.
In the ocean, mesoscale eddies (rotating water masses tens to hundreds of kilometres in diameter) undergo Fujiwara-like interactions as they drift and encounter each other. Two eddies of the same rotation sense (both cyclonic or both anticyclonic) will orbit each other and potentially merge, while two eddies of opposite sense will propagate along each other's streamlines without merging — a distinction that has no direct analogue in tropical cyclone interactions (since tropical cyclones in the same hemisphere always rotate in the same direction). The study of ocean eddy interactions has contributed significantly to the theoretical understanding of vortex dynamics that underlies the analysis of the Fujiwara effect.
In the context of climate change, the frequency and characteristics of Fujiwara interactions may change as the tropical cyclone climatology evolves. If global warming produces more intense but fewer tropical cyclones (as some projections suggest), the frequency of interactions might decrease while the consequences of the interactions that do occur — involving stronger storms — might increase. Alternatively, if the geographic distribution of tropical cyclones shifts, new basins might experience Fujiwara interactions that were previously rare. The research community has only recently begun to examine these questions, and the answers remain preliminary and uncertain — but the fundamental physics of the Fujiwara effect will not change, even as the storms that experience it evolve.
The Fujiwara effect — the interaction between two tropical cyclones rotating around a common centre — produces some of the most complex and difficult-to-forecast storm behaviour in tropical meteorology.
Key insight: The Fujiwara effect reveals that tropical cyclones are not isolated weather systems moving through an independent atmosphere — they are vortices embedded in a fluid, and when two vortices come close enough, they interact as fundamentally as two magnets or two gravitational bodies. The outcome of the interaction — orbit, absorption, merger, or deflection — depends on the relative size, intensity, and distance of the storms, but the interaction itself is inevitable once the circulations overlap. The atmosphere is not a stage on which storms independently perform; it is a medium through which they interact, and the Fujiwara effect is the most dramatic manifestation of that interaction.
The forecast paradox: Tropical cyclone track forecasting has improved dramatically over the past fifty years — average 72-hour track errors have decreased by approximately 70 percent since the 1970s. But Fujiwara interactions remain one of the situations where forecast skill deteriorates most sharply, because the coupled dynamics of two interacting storms amplify the initial condition errors that are inherent in any forecast. The better we become at forecasting single storms, the more conspicuous our remaining weakness at forecasting interacting ones. The Fujiwara effect is where the limits of predictability are most painfully visible.
Understanding the Fujiwara effect:
The Fujiwara effect begins when two tropical cyclones approach within ~1,000–1,400 km of each other
The storms rotate around a common centre — like binary stars orbiting their centre of mass
The most common outcome is absorption — the larger storm tears apart and absorbs the smaller one
Complete merger can produce a storm more intense than either predecessor
Track forecasts during Fujiwara interactions have 50–100% larger errors than for isolated storms
The western Pacific averages 1–3 Fujiwara interactions per year — the most active basin for this phenomenon
In summary: The Fujiwara effect — the mutual interaction of two tropical cyclones that come close enough to influence each other's motion — is one of the most fascinating and challenging phenomena in atmospheric science. Named after Sakuhei Fujiwhara, who first described it from water-tank experiments in 1921, the effect produces a range of outcomes from gentle mutual orbit to violent absorption and merger, each determined by the specific dynamics of the interacting storms. The forecasting challenges posed by Fujiwara interactions are significant — errors increase, models struggle, and the outcomes are sensitive to initial conditions in ways that push the limits of predictability. But the phenomenon also reveals something fundamental about the atmosphere: it is a fluid, and the vortices within it interact according to the same physics that governs spinning objects from laboratory experiments to galaxies. The next time two storms appear on the same satellite image, spinning their respective spirals within a few hundred kilometres of each other, what you are watching is not merely weather — it is fluid dynamics on a planetary scale, a demonstration of the universal tendency of vortices to interact, influence, and ultimately transform each other.
