"Weathering the Storms" (Dragon #137) and the Köppen Climate Classification System

The second major inspiration for what would later become THE ALMANAC of FANTASY WEATHER was Lisa Cabala's article "Weathering the Storms" in The Dragon #137.

The central insight of the article was to present twelve climate templates, templates that both tracked the different climate tyoes of our own world as well as (presumably) any half-way "realistic" fantasy world. The climate templates were these:

1. Desert
2. Tropical Savanna
3. Steppes
4. Equator
5. Warm and Rainy
6. Monsoon
7. Warm with Dry Summer
8. Warm with Dry Winter
9. Cool and Rainy
10. Cool with Dry Winter
11. Tundra
12. Polar

Each type was defined by its temperature ranges within the four standard seasons as well as the daily precipitation chances within those seasons. Those two variables and how they changed or varied (or didn't) from season to season constituted the basics of any climate type, and simply by presenting twelve of those types you could track the richness and diversity of all the earth's climates, or so it seemed.

For example, much of Western Europe was Warm and Rainy (5):

Unlike some other climate types, Warm and Rainy had a simple precipitation rate scheme. Daily chance of precipitation was a constant 40% throughout the year.

That's it.

I found this to be an improvement over what I saw as the more fiddly mechanism for determining temperature presented by David Axler in his "Weather in the World of Greyhawk" (The Dragon #68). There, the referee would, among other things, constantly be checking latitudes. Cabala's scheme also seemed more true to life. While latitude obviously affects temperature, in our own world wildly different climate types co-exist at identical latitudes. And in turn, very similar climate types appear at quite different latitudes.

So, the original design for my fantasy weather algorithm was to enter the climate data for the twelve climate types and then choose the appropriate one for each geographical location I wanted to model. If the average temperature for one or more seasons came out significantly different than that shown by the data for the location, I would simply make the appropriate adjustments.

However, in my quest for greater "realism," I soon became a bit dissatisfied with the list. As I found myself making more and more "adjustments" it became clear to me that the twelve templates offered worn't really enough. There were hot deserts and cold deserts, for example.

Somewhere along the line I discovered the Köppen Climate Classification System. In the late 19th century, the Russian-German climatologist Wladimir Köppen had also separated the world into different climate types. After the system was modified and refined by others, there were a resulting thirty climate types (or thirty-two, depending on how you counted them). Here were hot deserts and cold deserts (and hot semi-deserts and cold semi-deserts), among other things.

My inner weather nerd (yet another inner nerd this outer nerd never knew it had) became excited.

Again, one of the things that is fascinating about the climate types (if you're into this sort of thing) was how they didn't track latitude, at least as much as you would think. The world map above sort of gives you a sense of that, though it's more apparent if you focus on a smaller area. Such things as warm and cold ocean currents, wind patterns and so on are often much more important.

This is of course a commonplace for anyone with even a minimal knowledge of weather and climate, but it was useful and interesting to be reminded. 

I suspect Cabala had riffed off of Köppen but had wisely chosen to simplify it for her Dragon article. Thirty templates would have been too many for a scheme that required one to consult charts and roll dice. But if I were designing a relatively complex algorithm from scratch, starting with thirty sets of data wasn't that much more difficult than starting with twelve.

I would later make things even more granular to model different areas within each type - London versus Paris, for example - in order to make each location modeled truly "its own." And in the end, the classifications themselves would fall out as I settled on basing each of my own climate templates on the raw climate data for the particular location modeled, as opposed to its Köppen classification (which used the same sort of data but in a more general manner).

So, for example, I used Athens, Greece as a model for my internal Köppen template of Csa - Temperate, Dry and Hot Summer:   

But when creating my own Climate Template 101: Temperate South, which I roughly modeled on the climate of Jerusalem (also classified as Csa), I entered an entirely new set of data:

Would anyone have known that Temperate South, which I claimed to have roughly modeled on Jerusalem, was in fact really roughly modeled on Athens?

Yes. I would have. God would have.

Okay, even God probably wouldn't have cared, but still.

THE ALMANAC of FANTASY WEATHER: Volume 1, Swords & Sorcery was intended to simulate Middle-Eastern, Central Asian, Robert Howard-ish, Fritz Leiber-ish, "Swords & Sorcery"-ish sorts of climates. And, dammit, Athens didn't really fit. Entering the data for a second Csa climate type only took an extra hour or so.

And now, that one weather nerd out there who knows that it never would get down to -11 F in Jerusalem can't complain.   

But readers of THE ALMANAC of FANTASY WEATHER will notice that each of my own climate templates is also labeled with the Köppen type that it fits.

Tomorrow, the excitement level ramps up. Yes, I'm talking about temperature.