The Köppen–Geiger climate classification made simpler (I hope so)

[Azélor]

I would advise to wait before doing the precipitation step because I rushed it a bit. I still need to revise it just in case. Still, comments are welcome.

So, now for the last step..

My first try was a mess and the second one was too complicated.
This one is much simpler but unlike the previous one, it's not 100% automatic. The 2nd one generated all the temporary layers. While technically possible, it would be really complicated to do so now.
So, my best option is to upload a .psd file. The file would include all the dummy/temporary layers set at the right place, in the right order and with already written names. One will just have to fuse his layers with the ones already present.
This should guarantee that everything will work as intended.

Except that the size of the file might be too small or too big compared to the map, so people will need to adjust that. I made the temporary layers to be in the top left corner, and they are as small as possible (about 80 by 80 pixels).

[mbartelsm]

Hey Azelor, would it be possible for you to still explain how to do the last step manually? for those who don't use photoshop

[Azélor]

Well, I would have hoped that it could still work on Gimp. I'm not an expert, but I saw there could be a plugin maybe?

Anyway, it's possible to do it manually since that's what I need to do to record the script.


Basically, to find the resulting climates, we need to combine the temperature and precipitations layers, the 4 of them.
The main problem is that it result in around 3600 different combinations. It depends on the magic wand to select each combination and depending on the colors used, your likely to push the magic beyond capacity. Eliminating so of the by simplifying the table get the tola to 640. But the color problem is still there.

It was not easy to find colors where 2 combo are sure to be unique. Obviously, the basic colors are nice but really useful since they cover the whole color spectrum, I had to convert them. I tried different things and although there might be other options, this on is working well. I can't simply change the colors in the tutorial because it would really not be intuitive, as explained below.

Here what I did, a summary:

January temperature: from hottest to coldest, I replaced the colors with saturated blues, from bright to dark
July temperature: from hottest to coldest, I replaced the colors with saturated reds, from bright to dark
Combine the 2, and it give purple-ish colors. 100 combos

Here, we can isolate the combinations for tundra and eternal ice since the precipitations are not important to determine these. They are the 25 combos in the bottom right.total-tundra/ice= 100-25 = 75
We can simplify the table because now the temperatures are reversible: a cold January/ hot July is the same as a hot January/ cold July. Which further limit the possible combos. (75-5)/2=40
So we have 40 combinations for the temperatures. Some are extremely unlikely but still possible, especially on an alternate-earth.

Winter precipitations: blue, from bright to dark
Winter is in January in the north but in July in the south. So the precipitation map for January should have the top blue and bottom red. The opposite for July (Top:red bottom:blue)
Summer precipitations:red from bright to dark

Now, I realize that this part is not necessary because, the resulting combos can be simplified. Form 25 to 16. It's based on some mathematics: example 2 temperate cities will have the same climate even if they receive 1000mm and 20000mm of rain annually. After a certain pint the extra rain does not affect the climate (not based on the categories we are using). I decided to convert the remaining combinations in green. So I end up with the 3 primary additive colors: red, blue, green.

Then, we need to combine the temp and precipitation layers to get the final combination. And by using the right color key, we can identify all the specific climates.


I will provide more images later.

[Mysterious Mapmaker XXIII]

Nice to hear we're finally getting to determining the actual climate.

So, are you still doing the "automation" step or not? I'd like to know.

[Azélor]

Nice to hear we're finally getting to determining the actual climate.

So, are you still doing the "automation" step or not? I'd like to know.

Yes, It's almost working. It's getting it right til the 3/4 and messes everything else.



@Charerg I will look at your world as soon as I can make the script work.

[Charerg]

The tutorial is shaping up nicely. So far I've managed to follow it pretty easily, though I guess it might help that I went through Pixie's earlier tutorial previously, so I'm somewhat familiar with the subject.

I thought I'd post my "steps" that are part of the tutorial here, since that way it's a bit easier to get feedback. There's also a WIP thread that includes some of the tutorial-related maps (currents and atmospheric pressures), but I'll throw them into the attachments so anyone reading this won't have to dig through the WIP thread to find the relevant maps.

So far I've completed the steps of the tutorial for January temperatures and precipitations. There's some room for improvement in the details, but I think it will be easier to just fine-tune the eventual climates if something seems a bit off, rather than trying to perfect the temperature and precipitation maps. Anyway, criticisms and suggestions are welcome.

January temperatures:
Aduhr jan temp.jpg

January precipitation:
Aduhr jan precipitations.jpg

[Charerg]

I went and finished the process manually, and here is the resulting climate map (including the climatologically relevant oceanic currents, for easy reference):

Köppen Climates.jpg

For purposes of comparison, here is a prior version of the climate map I did back during the summer:

ERMap.png

Overall, I think it's a pretty big improvement over the previous map, which was partly done based on Pixie's tutorial and mostly done by just looking at Earth's Köppen climates on similar latitudes, or with making "educated guesses".

Although it has to be said that the final climate map shown here has been heavily "corrected" manually at places. It turned out that I screwed up with the precipitations, and Mediterranean climates were practically nonexistant, so I had to put them in manually. Also, I narrowed down the temperature combinations into 20 (from the maximum 100 possibilities) after merging the temperature maps, removing any weird "fringe areas". I didn't bother with refining the merged precipitation map, since I essentially did this manually by first selecting a temperature category (from the merged temperature map) and then going through the different precipitation areas within said temperature category. Theoretically, this would still result in 720 possible combinations, though not nearly all actually turned up.

Still, it took a long time. In hindsight, it might've been better to reduce the temperature categories still more, theoretically you only need 10 to do the different climates (I think). Merging the precipitation maps also resulted in a lot of "fringe areas" with weird combos, especially in mountainous areas. If I were to do this manually again, I'd consider maybe refining the merged precipitation map a little before defining the climates. In any case, what this tells to anyone doing this process via script: don't expect a perfect result, there will probably be a lot of weird fringe areas that will require manual corrections. Also, you need to be pretty careful when doing the precipitation maps, since at least I found it's pretty easy to rush them, and the accuracy of the climate map suffered as a result.

What I also found out in the process is that for purposes of fantasy mapping it might actually be worth it to expand the Köppen classification to include dry polar climates (in my map, I added ES and EW, steppe tundra and polar desert). Present day Earth doesn't really have these climates (as far as I know), but during the Ice Ages they would have been pretty widespread. I think the reason why dry polar climates are marginal nowadays is because all glaciated landmasses are practically islands, with no major non-glaciated areas. This means that tundra climates only occur in maritime areas, without extending far inland.

That said, it's common in fantasy maps to include landmasses that are only partially glaciated (essentially similar circumstances to Ice Age Eurasia). In these kinds of circumstances, you'd probably have an Ice Cap climate (EF) surrounded by polar desert (EW), transitioning into steppe tundra (ES).

EDIT:
It's also worth mentioning that I didn't quite realize until comparing with Earth's Köppen climates that the influence of the Winter Westerlies extends all the way to the Hindu Kush and Pamir mountains, resulting in "dry summer" climate types. It might be worth it to add a footnote about this into the precipitation section. At least I seriously underestimated how far inland the effect of the Winter Westerlies extends.

[Azélor]

I took your maps and used them to run the script. It should be functional now. The result is at the bottom.

---

The Mediterranean climates do exist in my version of the map, they cover maybe a quarter of what you have on your modified map (in number of pixels not in km2).

Yea, your right, some combination exist but they are really unlikely.

Still, it took a long time. In hindsight, it might've been better to reduce the temperature categories still more, theoretically you only need 10 to do the different climates (I think). Merging the precipitation maps also resulted in a lot of "fringe areas" with weird combos, especially in mountainous areas. If I were to do this manually again, I'd consider maybe refining the merged precipitation map a little before defining the climates. In any case, what this tells to anyone doing this process via script: don't expect a perfect result, there will probably be a lot of weird fringe areas that will require manual corrections. Also, you need to be pretty careful when doing the precipitation maps, since at least I found it's pretty easy to rush them, and the accuracy of the climate map suffered as a result.

It might be a good idea to reduce it further but it make the distinction between the arid, semi-arid and humid areas even less accurate. There are 10 different climate base solely on temperature: A, Ca, Cb. Cc, Da, Db, Dc, Dd, tundra, Ice. (without getting too much into details, we can't really identify Cc fro Cb with only 2 temperature maps)

The arid climates are defined by the aridity level, which depend on the evaporation (based on temperatures) and the precipitations. This means that although I have 16 temperature combinations for A,each combo will have a different aridity level and some will be steppes or deserts.

Fringes areas where pretty limited in y test. One thing I will mention (I mentioned it already but it was several months ago) is that the transition zone between the deserts and the humid climates is often missing. This happens often when the aridity changes rapidly due to a combination of a higher temperature and lower precipitations, and the fact that there are 4 layers with more or less random borders. You might end up with artefacts, or small odds colors.

Weird climates are almost to be expected in mountainous areas because, as mentioned earlier, the temperature and precipitations can change a lot over a small area. Like in the Andes.

What I also found out in the process is that for purposes of fantasy mapping it might actually be worth it to expand the Köppen classification to include dry polar climates (in my map, I added ES and EW, steppe tundra and polar desert). Present day Earth doesn't really have these climates (as far as I know), but during the Ice Ages they would have been pretty widespread. I think the reason why dry polar climates are marginal nowadays is because all glaciated landmasses are practically islands, with no major non-glaciated areas. This means that tundra climates only occur in maritime areas, without extending far inland.

By following the default classification, tundra and ice caps are only defined using the temperature. It's so cold and the evaporation rate so low that they can't be dry, or barely even if they receive a small amount of precipitations. I think, but the formula I'm using for the aridity tend to give weird results at low temperatures.

The dry tundra might be interesting though. Even if, form a demographic point of view, the population density on the tundra is really low and would be even lower with an arid climate.

tundra climates only occur in maritime areas, without extending far inland.

And in mountainous ones, especially in Tibet. But the reason they occur mostly on the coasts is because the temperature gradient is smaller than inland. Large landmass get hotter especially in summer making it too hot for a polar climate and even to hot for a tundra as in Yakutsk for example.

It's also worth mentioning that I didn't quite realize until comparing with Earth's Köppen climates that the influence of the Winter Westerlies extends all the way to the Hindu Kush and Pamir mountains, resulting in "dry summer" climate types. It might be worth it to add a footnote about this into the precipitation section. At least I seriously underestimated how far inland the effect of the Winter Westerlies extends.

Western disturbances as Indian climatologists calls it. I decided to include it with the orographic lift effect since it only rain at higher altitude (for this area and others at similar latitudes).

---

finale copie.png

The original I took on the forum were a bit blurry so there's a lot of unwanted small dots here and there, but we can have the general picture. I haven't made any transition areas on this, it's just as it came out after running the script. That partly explain why the desert are about twice as big.

Other things worth mentioning:

No Da: Da requires cold and below in winter and hot and above in summer. The only places with these temperature are too dry (either steppes or deserts)
No Dd: requires at least mild in summer and not warmer than deadly cold in winter.

Other than that, it looks plausible but I haven't looked at the temp and precipitation maps to see if they contain errors.

[Azélor]

Step 7, the script/the climates (manual version at the bottom)

Instruction in order to use the script. Explanations to do it manually will come later.

• MAKE SURE YOUR NOT USING THE ORIGINAL OF YOUR LAYERS, KEEP THEM IN ANOTHER FILE BECAUSE THE SCRIPT WILL MODIFY THEM.
• ALSO, DON'T PUT OTHER LAYERS THAN THE 4 MENTIONED ABOVE.
• Make sure your layers were made in an RGB file (it should be the default color mode if you never touched it). If your not sure, you can check the colors to see if they match.

Possible problems while running the script:

• Some areas are not selected and there are no climates: make sure your colors matches those I used in the tutorial. And make sure that you are only using these colors on the layers.

INSTRUCTIONS

• Download the zip file
• Open climates.psd
• Expand the Canvas size in order to have the same size as your map. Expand it toward the right and bottom to make sure the default layers stays at the top left: the position of these layers is important.
• Place your 4 layers (2 for the temperatures and 2 for the precipitations) in the file. Position them so your not missing a part.
• Fuse each of these layer with the corresponding already named layer. Make sure to keep the default names and the same order (for the layers).
• Fill the missing Background part with a pure black
• Make sure all the layers are made visible
• Load the script: go to window, script, click on the symbols with a triangle with three lines on the right side on the window, select load a script
• click on play to activate the script

koppen colors.png

Improving the climate map:

• Cfc climates are not really accurate. The script cannot differentiate them properly from Cfb. Therefore, most of the Cfc are actually Cfb. Cfb is much more common than Cfc. Normally the poleward climate progression should be either Cfb,Dfb,Dfc... if the climate is more continental or Cfb,Cfc,Tundra is the climate is more oceanic. Inland we could have Dfb climates right next to Cfc but never north of it.
• Some steppes are missing. When you see a desert and a humid climate without a steppe in between, it needs to be added manually. Expand the steppe inside the desert. Expand a lot if it's a plain and just a thin line if there is an elevation.

Azelor climates.zip
For the manual version:

I hoped it's not too badly written, English is my second language and PS doesn't have a spellcheck.

climates - explications copie.png

Also, these are the files I used for the script at their real size. I would recommend using these instead on the ones in the example above since they are smaller. And the larger one might contain some errors.

[Charerg]

Great work getting the script working!

Unfortunately I'm using GIMP, and I don't think there's a way to make *.atn files work with it (as far as I know). Anyway, if it's not too much trouble, would you mind running a 2nd "test run" of the script for me?

I refined the precipitation maps considerably, and they should now be much more accurate than the previous admittedly somewhat rushed versions. Also, I put these maps in *.png format and left the graticule out, so that should take care of the random dots showing up. The temperature maps are unchanged for now, since I think they're acceptable, and it should be interesting to see how much the results differ with just better precipitation data. There won't be any Dd climates without changing the temp maps, but since Dd is very rare in any case, I don't think it's too much of a flaw if it's missing.

Regarding dry tundra: I suppose the sort of Ice Age-era "steppe tundra" might be represented with just "cold steppe" using the default classification, it's after all more or less the same thing. I'm not sure if anyone has attempted to use Köppen classification to re-construct Ice Age climate, but as far as I know the steppe tundra is thought to have supported a surprisingly rich fauna, inlcuding large herds of herbivores like horses, saiga antelopes, bison, mammoths and so forth. All in all, in terms of human population, it may have been a better environment than modern "wet tundra" because of this. Indeed, perhaps even better than the taiga, at least for a hunter-gathering culture. The dry climate would have also been an advantage in winter, leading to less thick snow.