A cloud is less an object than it is an event.
A cloud is better described by a proliferation of verbs than with a singular, stolid noun. It is a blossoming, boiling, building affair with no beginning or end.
A cloud is ephemeral, emergent. It occurs at many different scales at once: in the interaction between a few molecules of water vapor, but also in the rearing of the thunderhead standing miles high, heavy and dark with downpour. As it rains, the cloud shreds itself like wool over the fine teeth of a carding comb.
A cloud becomes and becomes and becomes, and at some moment, the last waft of moisture wicks off into the atmosphere, and the sky is left open and unmarked.
Clouds are visible traces of our nonlinear, dynamic atmosphere. Their behavior depends on both the internal mechanisms of the weather system they comprise as well as the conditions they encounter as they move over the face of the Earth. The looping feedback between these inputs produces both regular features and surprising jumps. Clouds happen in characteristic patterns we name cirrus, cumulus, or stratiform; and yet they are new every time.
In the 1960s, the mathematician and meteorologist Ed Lorenz described the atmosphere as a realm of “deterministic chaos,” neither purely rational nor purely disordered. In other words, cause leads to effect, in that past atmospheric conditions determine future ones, but approximations of these conditions fail to approximate that future beyond a relatively short timeframe. The “butterfly effect,” which Lorenz introduced at a conference in 1972, refers to the possibility that very small, even undetectable perturbations can compound into dramatic differences over time.
To grapple with deterministic chaos requires negotiation of scale. In the atmosphere, interactions occur at many different levels at once, from the water molecule to the thundercloud. Each scale requires its own standard of measurement. Do you measure to one degree, a tenth of a degree, a millionth of a degree? No matter how precise the measurement, minute differences between the true system and its approximation persist. Lorenz calculated that even the most sophisticated weather forecast will fail to reliably predict the future more than two weeks ahead, as the real world and the forecast drift ineluctably apart.
Climate, which is defined by 30-year averages, bears a probabilistic relationship to weather. Yet this statistical smoothing does not eliminate the problem of chaos so much as disguise it under a veil of regularity. Over 30 years, stable-looking averages can be constructed, but beneath that surface, spiky, chaotic systems ceaselessly swirl and change.
In forecasts of climate change, clouds represent a significant unknown: They are capable of amplifying warming, by trapping heat, or reducing it, by reflecting sunlight. In the most basic terms, the higher up in the atmosphere a cloud occurs, the more heat it will trap. Meanwhile, the thicker a cloud is, the more reflective it becomes. Since a shift in either direction could amplify or dampen ongoing climate processes, the question rapidly spirals into overwhelming uncertainties.
A drop of water becomes a cloud. A cloud begets rain. Rain churns the earth and gathers in rivers, lakes, and oceans. From these simple transitions, an infinite array of futures emerge.
As the mathematician Benoit Mandelbrot once observed, in a statement equally self-evident and profound, “clouds are not spheres, mountains are not cones, and lightning does not travel in a straight line.”
No one but a mathematician or a climate modeler is likely to mistake a cloud for a sphere, and yet natural forms’ simple defiance of standard geometry is indeed remarkable. Clouds, mountains, and lightning look jagged and uneven, yet they are not without their own species of order. Mandelbrot coined the term “fractal” for these shapes that seem to subvert Euclidean lines while concealing deeper kinds of symmetry. The forces governing the formation of these natural fractals are often very simple—the random wandering of a particle or the repetitive force of a drip of water—but repeated iteratively, they can form patterns of near-infinite complexity.
Fractals are self-similar, which is to say, their parts resemble the whole. Imagine each letter in the word “fractal” was formed by a smaller line of text spelling out the word “fractal.” Each letter of those small words would be made up of a further “fractal,” and on and on. The closer you look, the more you see: an infinite lacework repeating and repeating and repeating itself. Thus, the branches of a tree resemble smaller trees, each containing further branchings and copies.
“It seemed there were limits to what we could know about the world, independent of our ability to observe or compute. The blank spaces of the future emerge out of the dynamics of these systems themselves, their jagged geometries.”
Amelia Urry, writer
Many chaotic patterns and dynamic systems contain such self-similarities across scale, whether in time or in space. Waves, for instance, may break when inches high or 50 feet tall, but they follow the same template: the rear, the curl, the bright splintering of foam.
Like the atmosphere, the ocean is a realm of deterministic chaos. Predictable, periodic patterns can be interrupted by irregularities, complex interactions, and rogue events; meanwhile, what looks like randomness may yield surprisingly rigid rules. The turbulent movements of water are evident in the wavelets that stir the ocean surface and the monstrous, hidden currents roiling the deep sea. Waves may break semi-reliably in certain favored surf spots depending on wind patterns, swell, and underwater topography, but they can also appear unexpectedly, out of the blue. These so-called rogue waves or freak waves may be more than twice as high as waves in the surrounding sea. They can occur in a string of waves or on their own, and have been known to swamp ships and oil rigs. They emerge out of the chaos, a deadly signal.
The mid-20th century articulation of chaos theory, and the discovery of fractal patterns within systems as widely separated as ocean waves and financial markets, revised expectations about the predictability of the physical world in complicated ways. On the one hand, fractal rules could help imitate the real outputs of these dynamic systems. On the other, this insight failed to lead to greater predictability beyond the near-term. Our best guesses remained statistical, not specific.
It seemed there were limits to what we could know about the world, independent of our ability to observe or compute. The blank spaces of the future emerge out of the dynamics of these systems themselves, their jagged geometries.
In the cold, water ceases to behave like a fluid. As busy hydrogen and oxygen atoms slow, they lock together, forcing nearby molecules into place. A crystalline grid emerges among the formerly disordered jumble of liquid water. Ice begins to form—rigid as glass, and as clear.
But ice does flow. Where it gathers in great enough quantities, it begins to move under its own weight like a frozen river. The mass grinds against its bed, churning up sediment, gouging a channel as it goes. It gathers speed as it descends. Faster-moving sections fragment into a chaos of broken blocks, tall as buildings, tumbling over each other. Nearby, slow-moving ice rests deep and placid under a velvet of new snow.
Like water, like air, the physics of glaciers are at once straightforward and complex. They can be modeled with some basic parameters: temperature, slope, the new accumulation of snow, and the ice whisked away by the evaporative force of the wind. Balance these factors just right and you get a glacier. Play the simulation forward: Watch it move.
In the early 2000s, a thousand square miles of floating ice known as the Larsen B ice shelf in Antarctica disintegrated in a matter of weeks—a result not accounted for in any existing ice sheet models. Glaciologists were faced with a sudden, dramatic illustration of the limits of their predictions. As the Intergovernmental Panel on Climate Change began drafting its fourth assessment report, the authors of the chapter on ice concluded that scientific understanding was so uncertain that it would be better to omit any projections of dynamic loss from Antarctica altogether. The result was a stable-looking prediction of gradual change, which disguised the increasingly alarming possibility of rapid and unpredictable changes. The ice sheet modeler Philippe Huybrechts commented, with a whiff of defeat, that “it may well be the weather of the ice sheets that no model can ever be expected to predict accurately.”
The weather of the ice sheets looms blackly in our future, like a vast storm. If glacial dynamics generate feedback loops that accelerate collapse, then global sea level may rise over generations rather than centuries. But exactly how the storm will break is still impossible to say. Many things move within the ice. Much that we cannot see may prove decisive. The sound of a new crevasse opening into the ice, like the pop of an ice cube fracturing in a drink. The millennia-deep weight of the ice sheet grinding its path to the sea.
Over 100,000 years ago, after an ice age had waxed and waned over the Earth, after the seas had overspilled their shores and the world ran thick and hot with flood, after all this had come and gone, a skin of ice formed over a shallow sea in what is now West Antarctica. The ice thickened, fed by the glaciers spilling out from high continental masses on both sides. The sea retreated as a lens of ice grew in its place, until its belly scraped the bottom of the basin in which it sat. To us, it looks permanent as a mountain. But in geological time, its existence is as transient as a cloud. Over millions of years, the ice has flickered in and out of existence. Never as fast as this, perhaps. Or perhaps faster.
The glacier, like the world, is an entangled swarm of cause and effect, where each action carries its consequence back to the starting point.
We know what weather looks like, but we still cannot predict the shape of the storm we have stirred out of this chaos. Until it looms above us. Until it breaks.
Special Thanks Department of Geology and Palaeontology at the Natural History Museum, Vienna
This story first appeared in Atmos Volume 11: Micro/ Macro with the title, “On Fractals and Forecasting.”
Biome
Join our membership community. Support our work, receive a complimentary subscription to Atmos Magazine, and more.