top of page

Consequences in Alternate History: Whither the Weather

By Tom Anderson

Masters of the near miss, (“‘Constant sunshine’? Three-foot drifts!”) Masters of the half-truth, (“Anybody ‘ere seen my roof?”) Masters of understatement, (“Sandbags, boats, insurance payments!”) So, we talk about, we talk about, we talk about the weather...

-Ian McMillan and Martyn Wiley on weather forecasters

A commonly invoked principle in alternate history (AH) is the ‘butterfly effect’, although this is often misunderstood or defined differently. These Consequences articles have looked at the idea that a tiny change can have a direct consequence upon seemingly unconnected people and events; but though rapidly gaining huge complexity, this is fundamentally a deterministic principle. The butterfly effect is also defined as a quantum effect in which a small change anywhere will effectively instantly ‘reset the dice being thrown’ everywhere else in the universe—but though this is probably the most scientifically rigorous approach to AH, it also feels disappointingly non-common-sensical to the human mind wanting a connected narrative.

However, we often forget that the original source of the name stems from meteorology, the study of weather and weather forecasting. In 1961 the great Edward Lorenz was using an early digital computer to predict the weather, but found to his surprise that tiny changes to his initial conditions—as small as whether a number was rounded to six or three decimal places—produced completely different weather predictions for a few days later. This was the beginning of what is called chaos theory, and the term ‘butterfly effect’ was coined in 1969: to quote the title of a 1972 paper, “Does the Flap of a Butterfly's Wings in Brazil Set Off a Tornado in Texas?” That tiny change to the initial conditions meant that even the most well understood complex system could not be accurately predicted over a long period. To this day, no matter how extensive our satellite networks and how sophisticated our computer models, the weather cannot be accurately predicted beyond a chaos event horizon of, at most, two weeks. This is an inherent property of nature and is unlikely to ever be bettered—though we can spot and communicate changes earlier than in the past.

It is interesting to note that from a human perspective, it is natural to group all ‘natural disasters’ under one heading. It matters little whether your house is destroyed by a hurricane, an earthquake, a volcanic eruptions or a flood—seemingly. Yet from an AH perspective—providing the AH in question is written by someone who knows what they’re doing—these categorise quite differently. Hurricanes and floods (triggered by rainfall) are fundamentally chaotic effects. It is natural to ask questions like “If John Kerry won the US presidential election in 2004, how would he handle Hurricane Katrina’s devastation of New Orleans in 2005?”

But that question is almost certainly meaningless. If Kerry wins in 2004, something would have had to change, our hypothetical butterfly has flapped its wings, and by the time the hurricane season of next year comes around, things would be quite different and there might be no Hurricane Katrina at all. Or one might be given that name, but be less intense, or make landfall far from New Orleans. While it’s natural for us to assume that, well, ‘natural’ disasters will stay the same in the background as we swap out our protagonists, that’s not how the world really works. Of course, this also ignores the fact that floods are also directly affected by human activity such as how strong a dam is (or indeed the levées breaking in the case of New Orleans). This is aside from the general impact on climate caused by human activity, of course, which would also change matters (e.g. if a different and less or more polluting engine is popularised, if a war causes oil prices to rise, and so on).

Conversely, earthquakes and volcanic eruptions are more or less predetermined and fixed by comparison: they are affected by external factors, but the timescale is so much longer by comparison over the span of human history. We can usually assume that there will always be a San Francisco earthquake in 1906—whether San Francisco is called St Francis, Cometa, or doesn’t exist at all in this timeline. Vesuvius will always erupt in 79 AD, even if Pompeii was the cowed colony of a victorious Carthage dominating the former Roman lands. There will always be an earthquake on Boxing Day in 2004 that devastates Southeast Asia—albeit via the aftereffect of a tsunami, so one could regard an earthquake-triggered tsunami as (paradoxically) a weather-type effect that is far more deterministic than a hurricane or typhoon. This comes with the caveat that it is possible for human activity to affect earthquakes (as in the case of mining or drilling in the wrong place at the wrong time) but generally speaking this distinction holds. Asteroid and cometary impacts also fall into this category of being unaffected by transient events; in “The Twilight’s Last Gleaming” I attempted to come up with a pure chaos theory justification for how an asteroid that did not hit the Earth in our timeline (OTL) did in another one, all stemming from the random radioactive decay of a single atom. I did this because I always felt unsatisfied by the idea that one can simply invent or move asteroid impacts for the sake of a story, even if it gives an excellent story such as “The Loud Blast That Tears The Skies”, one of a number of pieces of AH fiction that posit a different impact site for the Tunguska Event of 1908.

"Lancaster Sands" by J.M.W. Turner. The reddish tint of the skies in his pictures of this era is thought to be influenced by the effects that caused The Year Without A Summer

Because of this, in grand-scale timelines such as my “Look to the West” I feel the correct “hard” AH approach is to assume that impacts, volcanoes and earthquakes will take place on the same dates and in the same places as our timeline, but to immediately discard any temptation to feature an iconic storm, flood or harsh winter from our timeline. Such events should take place, but on almost randomised dates (aside from the constraints of the hurricane or monsoon season of an area and so on). This is the appropriate outcome of the chaos model that Edward Lorenz discovered. One rare exception is the Year Without a Summer (1816) which caused crop failures and mass deaths in Europe and America (and elsewhere) which in turn meant many Americans fled New England to settle the Midwest, having a huge effect on American history. This event would always take place regardless of TL, because it stems from the ‘nuclear winter’ like effect of the eruption of the volcano Tambora in modern Indonesia. A smaller similar event happened in 1883 with the eruption of Krakatoa. By contrast, a crop failure caused by disease (such as the Irish potato famine of the 1840s, which also affected many parts of Eastern Europe) could have happened at any time, earlier or later, more or less severe, and indeed I have this happen earlier and less severely in “Look to the West”.

Of course, this article has only considered major disasters. Weather can have far more subtle impacts on history than simply destroying towns and slaying people. In the age of sail, winds would determine whether ships could give battle, could give chase, could catch others. The fates of many countries were determined by sail-based naval battles which might have happened differently. The Spanish Armada of 1588 was defeated more by weather effects than English actions, hence why the medal issued to the English sailors was inscribed with ‘God Breathed And They Were Scattered’. In a fascinatingly parallel piece of history, the Mongol invasions of Japan a few centuries earlier were also defeated by what the Japanese called a ‘divine wind’—the origin of the term Kamikaze, later applied poetically to suicide pilots in World War II.

Land battles, too, could be averted or diverted by rainfall in the wrong place. Until the late nineteenth century, rainfall could make guns useless, and in earlier centuries bows as well. Even today, an unexpected rainfall can drastically change the shape of a battle. Cloud cover can interfere with visual satellite cover to an extent.

There is much more that could be said about the effects of weather on history, but for now—just remember that not all natural disasters happen on a set, deterministic calendar regardless of human activity!



bottom of page