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

[Shaetano]

Hi and thanks for the reply. Wow.. I must say, that is way more than I expected. Thank you for your help.

Most of your points helf me very much, I even started to redo the map as an Equirectangular projection. It isn't THAT much work anyway..

There are just one thing I want to add and two questions about your work. First - the big blue blob in the north-east which you identified as a part of the ocean is in fact just a blue part of the cold climate of the climate map. But I don't think that would affect the currents too much, so I'd just let them travel around the bigger landmass.

One thing is a bit unclear for me. The pressure/wind maps you posted - do they follow the guide for one of the seasons, or did you just but them together ignoring the current month as a little "overview". The tip about not including the continental pressure systems is great. They just add to my confusion and if their influence isn't that big, I think I won't inclue them either.

The 2nd question - how the hell do you draw these nice arrows?

[Charerg]

Hi and thanks for the reply. Wow.. I must say, that is way more than I expected. Thank you for your help.

Most of your points helf me very much, I even started to redo the map as an Equirectangular projection. It isn't THAT much work anyway..

There are just one thing I want to add and two questions about your work. First - the big blue blob in the north-east which you identified as a part of the ocean is in fact just a blue part of the cold climate of the climate map. But I don't think that would affect the currents too much, so I'd just let them travel around the bigger landmass.

One thing is a bit unclear for me. The pressure/wind maps you posted - do they follow the guide for one of the seasons, or did you just but them together ignoring the current month as a little "overview". The tip about not including the continental pressure systems is great. They just add to my confusion and if their influence isn't that big, I think I won't inclue them either.

The 2nd question - how the hell do you draw these nice arrows?

The pressures and winds are just an overview, they're basically intended to demonstrate what the basic pattern should look like (like I mentioned, it does vary a bit between the seasons). As to the arrows, they're just hand drawn. In GIMP, you can use the "Smooth stroke" option and adjust the settings to draw nice and smooth lines.

[Shaetano]

Hey!

Like I said I started from the beginning again. With the help of your pictures I went through the guide again and hope I did a little bit better this time.

What I am unsure about are parts of the pressure systems. I followed the guide more or less, but don't know if it would be better / more realistic to change the size/position of the systems from season to season more than I did.

Here are my maps from step to step. The last one, after the climate script, isn't edited, as it takes really much time and I want to see if I have to change anything before I start with that.


General Map with heights & currents:
currents.png

Winds & pressure January:
wind_pressure_january.png

Winds & pressure July:
wind_pressure_july.png

Temperatures January:
climate_january.png

Temperatures July:
climate_july.png

Rain January:
rain_january.png

Rain July:
rain_july.png

Climates:
climate_zones.png

[Charerg]

A bit of news to revive this thread (btw, shouldn't this be in the tutorial section now that it's finished?):

I made a bit of a modification for Azelor's tutorial, ramping up the precipitation categories to 8 (from 6). Additionally, I changed the scheme for categorizing the climate types (precipitation-wise, the temperatures remain unchanged). As most who have used the tutorial are probably aware, the current version tends to have a few problems (namely relative lack of BS areas in some cases, and a tendency for Af->BW or Aw->BW transitions), although it is still by far the best and most advanced climate-generation tutorial available. I think I've managed to identify and largely eliminate these problems, although this solution does complicate the tutorial slightly due to a greater amount of precipitation categories.

But before I get into the modified version, let's take a look at the changes I made to the climate categorization scheme:

A) Defining the precipitation patterns of A climates separately from C and D
A climates use different criteria when determining their type (f, m, w or s) than C/D. For example, it's entirely possible for a climate that would otherwise be classified as 'Cf' to be classified as Aw or As. Meanwhile, only the very rainiest climates (no month with less than 60 mm rainfall) qualify as Af.

I used the following scheme to classify the climate types, differentiating Af, Am, Aw and As from C/D f, s and w:

Prec Combos.PNG

The table displays the different average precipitation values I used for the different precipitation categories, and the Mean Annual Precipitation (Pann) calculated for each combo (this is used to determine whether a climate is steppe or desert). Note that 's' or 'w' climates would qualify as Aw/As (but as you can see, some 'f' climates also qualify as Aw/As!).

B) Determining the aridity threshold
Köppen uses different aridity thresholds depending on Mean Annual Temperature (MAT), and whether 2/3 or more of the precipitation occur in winter or in summer. These classes are quite different from the climate categories s, w and f. So, I used a different table to determine which aridity threshold should be used:

Prec Values for PTh Calc.PNG

Here, areas below the purple line can never be arid (they are too rainy), areas below the yellow line can never be desert (they can still be steppe), and areas above the red line are always desert.

Finally, here is the table I used to determine the MAT value of the various temperature combinations (Tann=mean annual temperature), as well as the associated aridity thresholds (areas with Mean Annual Precipitation (MAP) below the threshold are classified as steppe, and areas with precipitation less than half of the threshold are desert):

Temp Values.PNG

The red lines denote the boundaries of different temperature-based climate types, the yellow line denotes the boundary between Da and Da2 (an extreme version of Da, this climate doesn't appear on Earth). Note that although I used 'f', 's' and 'w' here to signify the various aridity thresholds, these refer to the aridity category, not the actual climate category!

C) The Results
With that out of the way, let's get into the testing of the modded version. First of all, I used data from WorldClim to generate temperature and precipitation maps for Earth (these should be almost identical to those used by Azelor in his earlier post). The only major difference is that these are in the Equirectangular Projection, and the precipitation categories are different:

January Temp:
Temp Jan.png

July Temp:
Temp Jul.png

January Prec:
Prec Jan.png

July Prec:
Prec Jul.png


And then, the final result. This is in the more commonly used "wikipedia colour scheme". Note that as no data was available for Antarctica, it has just all been painted as EF (I think some tundra exists there as well):

KG Climates.png

This is completely "script-generated" (actually manually done as I use GIMP), although I did shuffle some categories around to make it match actual Köppen maps of Earth as closely as possible. In general, it matches extremely closely, though there are three major divergencies:

1): Equatorial Africa is a bit messed up. This is caused by the "data gap" from having data only from January and July. As a result, there are weird instances of BS and As popping up, and Af and Am do not cover quite as much area as they should.

2): I classified any B climate with Cool (0-10) winters or colder as BSk/BWk. As a result, some areas like Northern Arabia have Bk climates instead of Bh. However, as drawing the Bh/Bk boundary at Cold (-10 to 0) would classify large areas of Central Asia and the North American prairies as Bh, I decided that this was the lesser of two evils.

3): The third difference is of course the distribution of the Cc climates. An old issue, and it can't really be fixed with the temperature categories we have withouth eliminating Cc climates completely (as pointed out by Azelor previously).

That said, I have yet to test this on a fictional world. The good news is that while this probably does sound rather complicated, it doesn't actually make the tutorial much harder to implement (you just have to use those 8 precipitation categories instead of the original 6). The bad news is that as I don't have PhotoShop, there is no possibility of scripting this, so this will be strictly manual-implementation-only. I'll try to put a "supplement for Azelor's instructions" together detailing how to implement this, in case someone is interested in trying this out (though I'm mostly making this for my own use ). I should note that this is perhaps intended more for the "advanced user" who is already familiar with the tutorial (and has preferably already tried the manual implementation before), since otherwise the extra complication will probably be of little benefit.

[Azélor]

I'm not sure if I misunderstood but A climates are different of C because of the temperatures, not the precipitations.
I'm also unsure why you said that the precipitation categories should be different for A, C and D. What do you mean by that?

[Charerg]

I'm not sure if I misunderstood but A climates are different of C because of the temperatures, not the precipitations.
I'm also unsure why you said that the precipitation categories should be different for A, C and D. What do you mean by that?

I mean that Af has specific criteria (no month with less than 60 mm) that are different from Cf/Df, as do Am (annual precipitation more than 25*(100-Pmin), where Pmin is the precipitation of the driest month). Likewise Aw and As are determined based on whether the month with less than 60 mm occurs in summer or in winter.

These criteria are much looser (in the case of Aw and As) than the criteria used to define Cs/Ds (3 times as much rain in the rainiest winter month than the driest summer month) and Cw/Dw (10 times as much rain in the rainiest summer month than the driest winter month). This is particularly true in our case perhaps, since we only have data from two months (so, effectively 3* the rain in winter->s and 10* the rain in summer ->w). I'm not sure if you took these into account, but this does result in the possibility that some climates with the same precipitation pattern could be classified as either Aw/As or Cf/Df depending on which temperature group they belong to.

The "Prec Combos" table I posted demonstrates this quite well. You can see that some precipitation combinations are classified as "f OR Aw" or "f OR As". One fringe case is the W4+S4 (avg. 50mm + avg. 50mm), which results in an even precipitation pattern, but is nevertheless too dry to be classified as either Af or Am. In most cases, this would actually fall into BS (because of the relatively low overall precipitation), but not in all. I chose to classify this as Aw, since it's more common than As, but there isn't really a category within A that represents this particular climate well.

[Azélor]

I took these numbers in consideration but since you used more categories, you had to redo the process anyway.

[Charerg]

I took these numbers in consideration but since you used more categories, you had to redo the process anyway.

Yup, although I noticed there are actually some cases which should be Aw but are classified as Af in the standard version. Here's an example from Shaetano's recent map:

January rain:
rain_january.png

July rain:
rain_july.png

Climate zones:
climate_zones.png

Here the area outlined in red has category 2 rain (25-50mm) in January. Since this falls below the "no month with less than 60mm" threshold these areas should actually be classified as Aw and Am (rather than Af). Since I think Shaetano just used the script to generate the climates, this is probably a bit of a flaw in the "data processing" of the script. I assumed that you had lumped all the 'f climates' together into one category (since these would be classified as f climates if they were C or D), but I guess some combinations might be misplaced instead if you took the differences with the classification of the A climates into account?

[Charerg]

*cue Jurassic Park theme song*

As a bit of a pseudo-scientific intellectual curiosity, I've been thinking of attempting to do a map of the Köppen climates for Pangaea using the tutorial (should also be a fairly good test of the modified version). Since fantasy maps feature huge continents fairly regularly, this might also be somewhat informative for fantasy cartography in general. I chose to use 179 Mya (Toarcian) as the date, since at this date the majority of the Pangaean landmass was sitting straight on the Equator (an arrangement that is roughly comparable to the continent of Eocidar in my own con-world of Aduhr). Another reason for the choice of exact date is that there's good quality data available for this date from the PALEOMAP Project (CR Scotese has uploaded a pdf portfolio that has all sorts of maps for this date, you can check that out if you have an academia.edu account (actually, I think you can just use a google account for that)).

So, first off, the elevations:
180 Mya - Toarcian Elevations.png

This is my very rough adaptation based on the original map from the PALEOMAP project. You can find the original in the pdf I linked, or alternatively you can check out Mr. Scotese's Youtube channel (this vid at 180 mya shows about the same map in Mollweide projection). I forgot to include the elevation key here, but it's the usual one:

Dark Green: 0-250m
Green: 250-500m
Light Green: 500-1000m
Yellow: 1000-2000m
Orange: 2000-3000m
Brown: 3000-4000m
Dark Brown: 4000-5000m
Grey: 5000m-6000m

Note that there's a lot of guesswork involved in the elevations (like I mentioned, I did this fairly quickly, so it's meant to be merely "good enough", not necessarily accurate when it comes to the minor details). Still, I believe the elevations should be more-or-less accurate (like I mentioned, I based this on the PALEOMAP reconstruction), at least when it comes to the distribution of highlands/lowlands and their relative elevations. It's of course extremely debatable how tall the Appalachians might have been, for example, during the early Jurassic. However, we can make a decent guess based on modern geological features that the highest average elevations probably occurred in the Altai-Sayan region, with the then-recent collision between the Cathaysian-Amurian and Eurasian plates.

Regarding the climates, the main climate indicators we have that can help to reconstruct the Pangaean climates are lithological indicators (coal deposits, bauxites, evaporites and so forth) and floral indicators (leaf shapes of fossilized plants). The portfolio of PALEOMAP includes a basic map of the broad climate zones based on lithological indicators (again, Mr. Scotese also has a youtube vid that includes this). For the floral indicators, there's also data available, along with model-generated climate zones on the website of the Paleogeographic Atlas Project. Those will come in handy when this progresses into the stage where temperatures and precipitations are determined.

And finally, before I get started with the actual climate-related stuff, I should note that I'm assuming that Earth is in the same phase of the Milankovitch cycle as today (same axial tilt and so forth). With the intro out of the way, lets start with the currents:

180 Mya - Currents.png

Here, I've lazily drawn both neural and warm currents as red, and cold currents as blue (these represent relative rather than absolute temperatures, btw). I've put the approximate locations of the semi-permanent oceanic pressure centers on the map along with the currents (the actual locations will vary seasonally, this is just for visualizing the general pattern). I should note that although the portfolio does include a map of simulated oceanic currents, I've relied on my own intuition here (although it agrees with the simulated stuff for the most part). It should be noted that Mesozoic climate simulations tend to be highly inaccurate, and hence I think we're better off relying on the supercomputer of the human brain here (not that this is a super-serious reconstruction, more like an intellectual experiment).

So, that's it for the preview. I'll try to get the atmospheric pressure systems/dominant winds maps together next (this is when stuff gets complicated).

[Charerg]

And the Pangaean "climate study" continues:

In the past I've often advocated largely ignoring the continental pressure centers, and concentrating on the oceanic pressure centers. In general, I think this is good advice because the oceanic pressure centers are largely anchored in place by the oceanic currents, and hence their position is much easier to determine. When it comes to continental pressure centers, they vary much more seasonally (and based on the topography), and also there is a relatively complicated interplay between continental and oceanic pressure centers, that can be hard to figure out (and not just *can be*, it is hard to figure out).

Anyway, this time, since this is an actual map of Earth (albeit 180 million years ago), I've decided to attempt to model the whole circulation. I used the general locations of the pressure centers in the currents map (see my previous post) as a starting point, with seasonal fluctuation taken into account. Since the Pangaean climate is believed to have experienced highly monsoonal behaviour, I've generally used modern northern hemisphere as a basis for both hemispheres (see this site for info about the Indian Monsoon on present-day Earth).

That said, the development of the Asian Monsoon is believed to be connected to the uplift of the Tibetan Plateau. So, I've generally assumed the Pangaean Monsoon to be somewhat less extreme particularly in the high latitudes, due to the absence of a similar formation.

So, the pressure and wind maps:

January:
180 Mya - Jan At.png

July:
180 Mya - Jul At.png

As I mentioned, I generally used the behaviour of oceanic pressure centers in the N. Pacific and N. Atlantic as a basis for both hemispheres. So, during northern summer, we see the N. Panthalassic high pressure center intensifying in response to the formation of a major low pressure center over Pangaea, and vice versa in the southern hemisphere. Similarly the Polar High Pressure Centers generally form over land in winter (cooling up faster than the ocean), and they expand equatorwards when high-elevation areas are present (Antarctic and Siberian highlands). During the summer, the Polar HPCs disappear, while the Subpolar Low Pressure Centers weaken accordingly and retreat to the continents (land warming up faster than the ocean), but during the winter they move back over the oceans and intensify significantly in response to the formation of the Polar HPCs over the continents.

So, in essence:
Winter-> the Polar cell intensifies, the Ferrel and Hadley cells retreat south
Summer->the Ferrel and Hadley cells intensify and expand north, the Polar cell becomes very weak

This is behaviour comparable to the northern hemisphere today, and as the closest present-day equivalent, I believe should be about right. Another thing worth pointing out is the direction reversal of the cross-equatorial trade winds. This is an important phenomenon when it comes to Tropical climates (if trade winds originating in the N. or S. hemisphere cross the equator, the Coriolis force inverts their direction from easterly to westerly flowing).