View Full Version : The Köppen–Geiger climate classification made simpler (I hope so)

07-23-2014, 01:03 PM
I see that lot of people have a hard time with their climates when they are creating their world. I'm not an expert, but I'm trying to make it simpler when it come to the climatic zones.

I used this as the main source of information : Köppen climate classification - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/K%C3%B6ppen_climate_classification)
But also other pages and some scientific articles. The information is there (most of it) but it's not very easy to understand. The first part I did was to collect the info on the wiki and also elsewhere to get a clear definition for each climatic zone: what it is but also where to place them: where they are the most likely to appear. It ended in something relatively complex and a long document.

note : My main goal was to find if there was a link between the climate and population density. And if so, to find what density would be for one specific climate.
I realized I had a serious issue with a state with similar climate to India and somewhat similar size but the population was about only 9 millions people. I did not make sense and I had no answers how to get population density numbers right.
I also happened to have no clear idea on where to place the climates. That's why I'm doing this, and hopefully I can work on the second part after this one is finished.


07-23-2014, 01:31 PM
I tried to open the file but OpenOffice tells me it's damaged. Does that document come in any other format?

07-23-2014, 01:37 PM
OpenOffice Writer can read and write .doc files just fine. Except for this one.

07-24-2014, 09:42 AM
I've read the whole thing, it is helpful indeed Azelor.

Being a science teacher and an amateur climatologist/geologist (in fact, an amateur world-builder), I can follow all of it pretty easily. Dunno how less science savvy folks will manage it, but it doesn't seem complicated.
I think it is a very valuable add-on to the tutorial I (we) are currently building.

It made me think of one more climate map that would be very useful: a cross-reference between rain pattern and temperature, to make a two-colored map separating areas where evapotranspiration is greater / lower than precipitation. This could perhaps be helpful (do you think it would be helpful?)

Picking up the word usage in my tutorial, what do you think of this?
Lower precipitation than evaporation (DRY seasons)
Very Hot + Moderate/Low/Dry
Hot + Low/Dry
Warm + Dry
Roughly equal precipitation to evaporation (MODERATE seasons)
Very Hot + Wet
Hot + Moderate
Warm + Low
Mild + Low
Cold + Dry
Very Cold + Dry
Higher precipitation than evaporation (WET seasons)
All remaining combos

Corvus Marinus
07-24-2014, 11:32 AM
Breaking down the meaning of each of the letters is really helpful in understanding the details of the system. And I think Part II: Climate Zones will be very useful; it's a lot clearer than Wikipedia. The formulas in Part I, for me (not very science-literate), do not translate into immediate usefulness in worldbuilding; but I can easily take Part II and use it to "proof" my map after it has gone through Pixie's system.

If I am still around when you have a final draft, I will happily check it for spelling/grammar/formatting, if that is something you'd appreciate.

07-24-2014, 06:01 PM
I've read the whole thing, it is helpful indeed Azelor.

Good to know.

Were you considering a map or some info graphic like this one : http://powerfulinfographic.com/wp-content/uploads/2011/10/transparency-11.jpg
I don't know about the idea. There is a relation between temperature and minimum precipitations to avoid desertification. Did you know that the Sahara would need between 3000mm and 6000mm of rain per year to become a moderate climate similar to Spain? That's a lot of water and it's just the minimum.

There is a part in the guide where I talk about yearly precipitation not seasonal. The problem with the classification is that it compares the driest month with the wettest without taking in consideration if the driest month is really dry. Sometimes, it's not the case. It's considered dry only because the wet month receive a lot more rain. So the letters s and w are more or less valuable here. At the equator, at least we know that under 60mm it's considered dry.

edit : this ! http://en.wikipedia.org/wiki/Aridity_index

and this : http://upload.wikimedia.org/wikipedia/commons/c/cc/K%C3%B6ppen-aridity-index.png

I'm now using this from the Trewartha wikipedia page: BW and BS mean the same as in the Köppen scheme, with the Köppen BWn climate sometimes being designated BM (the M standing for "marine"). However, a different formula is used to quantify the aridity threshold: 10(T − 10) + 3P, with T equaling the mean annual temperature in degrees Celsius and P denoting the percentage of total precipitation received in the six high-sun months (April through September in the Northern Hemisphere and October through March in the Southern).
If the precipitation for a given location is less than the above formula, its climate is said to be that of a desert (BW); if it is equal to or greater than the above formula but less than twice that amount, the climate is classified as steppe (BS); and if the precipitation is more than double the value of the formula the climate is not in Group B. Unlike in Köppen's scheme, no thermal subsets exist within this group in Trewartha's, unless the Universal Thermal Scale (see below) is used.

07-26-2014, 01:03 AM
Hey, I got some numbers!

Replacing this :

• If less than 30% of annual precipitation occurs in the summer : Annual precipitation (mm) < 20 × average annual temperature (°C)
• If more than 70 % of annual precipitation occurs in the summer: Annual precipitation (mm) < 20 × average annual temperature + 280
• Else : Annual precipitation (mm) < 20 × average annual temperature + 140

o If annual precipitation is < 50 % of the threshold = BW: desert climate
o If annual precipitation is between 50 and 100 % = BS: steppe climate

by this:

if the annual precipitation (in centimetres)

are Greater than R= humid
are Smaller than R but greater than R/2= semi-arid
are Smaller than R/2= arid

R=2 x T if rainfall occurs mainly in the cold season (s=summer dry)
R=2 x T + 14 if rainfall is evenly distributed throughout the year (f)
R=2 x T + 28 if rainfall occurs mainly in the hot season. (w= winter dry)
(T= mean annual temperature)

07-28-2014, 10:01 AM
Hmm, this complicates things a little bit. But at the same time, it helps. It helps because now we can get a workflow that will yield more accurate climate maps and it complicates because that workflow will be a little more messy now.

I tried to build a humidity map based on this info and on the scheme I mentioned earlier. It doesn't fit with climate predictions made like I suggested in the other thread in some places - namely, areas classified as savanna/monsoonal close to the tropics now seem Arid (desert?) all throughout the year and steppes at higher latitudes now have a properly Humid season, making them maritime/mediterranean.

Azelor, you have definitely raised a point that can't be overlooked. I will need to review my stuff :) Thanks for that (or not! no, seriously, thanks for that)

07-28-2014, 07:34 PM
You also need to be cautious with the numbers. For example, Jaipur is classified a steppe but receive a little more than R. Inside one climate one can see huge differences. Lisbon (Csb) is a Mediterranean climate but R=4 it's pretty wet. While Los Angeles (Csa) is barely above 1. It's not always clear because categories includes a broad range of possibilities.

I like these formulas because they take into account that precipitation have a different impact depending when it fall.
Logically, if precipitation evaporate at a slower rate in winter, the water (or snow) will stay longer in the environment and thus will have a bigger impact on nature. In theory.

Two cities receive the same amount of precipitation for the year. City A is summer dry and City B is winter dry. Over the course of the year, which of the two cities will be the driest?

areas classified as savanna/monsoonal close to the tropics now seem Arid (desert?) all throughout the year and steppes at higher latitudes now have a properly Humid season, making them maritime/mediterranean.
plausible, but having a wet season does not always make the steppes a maritime/mediterranean climate. Only if they are not too far from the water.

07-28-2014, 08:15 PM
I like these formulas because they take into account that precipitation have a different impact depending when it fall.
Logically, if precipitation evaporate at a slower rate in winter, the water (or snow) will stay longer in the environment and thus will have a bigger impact on nature. In theory.

Yeah, that's the basic reasoning I think. That's what made me review the whole process - I am now going combo by combo, it becomes a huge table.
5 january temperature levels x 6 january rain levels x 5 july temperatures x 6 july rain levels.... 900 entries.
The original idea is to simplify climate prediction, 900 entries isn't simplifying. Still a work in progress...

Two cities receive the same amount of precipitation for the year. City A is summer dry and City B is winter dry. Over the course of the year, which of the two cities will be the driest?
Did I say I am a science teacher? This is easy, when most of the rain falls in winter, moisture is available for longer. Thus, the location with rain in the summer is the driest.
However... plants metabolism is very dependent on sunlight, so the location with rain in summer may have more vegetation cover as both factors for plant growth coincide, and a dry hot summer requires plants adapted to drought, which normally means smaller leaves and slower growth rate.

07-29-2014, 08:38 PM
Yeah, that's the basic reasoning I think. That's what made me review the whole process - I am now going combo by combo, it becomes a huge table.
5 january temperature levels x 6 january rain levels x 5 july temperatures x 6 july rain levels.... 900 entries.
The original idea is to simplify climate prediction, 900 entries isn't simplifying. Still a work in progress...

Did I say I am a science teacher? This is easy, when most of the rain falls in winter, moisture is available for longer. Thus, the location with rain in the summer is the driest.
However... plants metabolism is very dependent on sunlight, so the location with rain in summer may have more vegetation cover as both factors for plant growth coincide, and a dry hot summer requires plants adapted to drought, which normally means smaller leaves and slower growth rate.

I think we need to make things complicated in order to understand the simples rules that makes the system. With some advanced statistics, we might be able to find interesting informations. I would like to see the file when it's done if possible.

I got 10 temperature levels, what are your temperature levels?

07-30-2014, 05:41 AM
Table isn't finished, but here's what I mean:

Table is being made in an excel file and the result exported to pdf like you see here. Starting point should be january temperature, then using magic wand on intercept, user would shorten selection with january rain, then july temperature, then july rain... (and, to cover the whole map, repeat that 900 times!)

Any ideas are very welcome at this point.

Also, as you can see, there are lots of combos which I have doubts about or which, even if I apparently don't have doubts I am plainly wrong about.
If you have time, please give it a look - so far I only have these.

07-30-2014, 11:49 AM
The letters s and w represent the dry season, but I wrote that it was the wet season. So, winter dry always require more precipitations.

(f) R=2 x T + 14 if rainfall is evenly distributed throughout the year
(s) R=2 x T if rainfall occurs mainly in the cold season
(w) R=2 x T + 28 if rainfall occurs mainly in the hot season.
(T= mean annual temperature)

Severely hot: 35 °C or more
(Minimal rainfall for 35°C, if the temperature is higher than that, more rain is required)
Very hot: 28 to 35 °C
Hot: 22 to 28 °C
Warm: 18 to 22 °C
Mild: 10 to 18 °C
Cool: 0 to 10 °C
Cold: −10 to 0 °C
Very cold: −25 to −10 °C
Severely cold: −38 to −25 °C
Deadly cold: −38 °C or below

07-30-2014, 12:15 PM
I'm trying to classify the Dsb climate using your classification but I'm not sure how to.

• Precipitations: moderate
o Summers = wet
o Winters= dry
• Average monthly temperature between -25 °C and 28 °C
o Summer: mild to hot
o Winter: very cold to cold

the problem I'm having is that I know how much precipitation are required yearly but not for the specific seasons.
I think I might have an idea, but it could make the numbers above useless.

07-30-2014, 05:25 PM
I'm trying to classify the Dsb climate using your classification but I'm not sure how to.

• Precipitations: moderate
o Summers = wet
o Winters= dry
• Average monthly temperature between -25 °C and 28 °C
o Summer: mild to hot
o Winter: very cold to cold

You're doing it again... Dsb means "dry summer" ;)
I would make it warm summer with arid or semi-arid conditions (low/dry rain patterns) and very cold winter with humid conditions (any kind of rain pattern except "dry" gives humid conditions in a very cold season)
A mild summer would make it Dsc... A hot summer would make it Dsa... this is the sort of reasoning I am making in building the table.

07-30-2014, 06:13 PM
I have another idea. It will make it easier to decide if one climate is winter/summer dry or forever wet.

first we need to set these assumptions:

It’s in the northern hemisphere: july is in summer and january is in winter
January is always the coldest month
July is always the hottest month
January is the driest month of the year (w=winter dry)
July is always the driest month of the year (s=summer dry)

f: precipitation levels are either on the same category, 1 category down or up.
w and s: they are separated by at least one category

example: Pixiland receive 30mm of rain in January and 70mm in July. =f because they are just 1 category apart.
Azelor Town receive 15mm in July but 75 mm in January. It's a dry summer, category 20-40 is separating each seasons.

Both climate could be considered humid, maybe one is more humid than the other but that's not too important.
That way, it's simpler than the : precipitation < 1/3 of the wettest winter month
and the numbers are not very different.

07-30-2014, 06:54 PM
I see what you mean, and that makes it very easy to decide between a s-climate, a w-climate or a f-climate. But, that's 10 levels of rain.

Say.. we keep the 6 levels of temperature as the current system gave pretty matching results in my test with ascanius and add more levels of rain.
Instead of the existing 7, we add two more levels (could we merge the 0-5 with the 5-10?). The current process gives 7 levels, but I ignore the 6th and 7th.
This can be done adding two layers in the present composition of rain patterns and I think it can be done in a few different ways.
This could work, but it now becomes a 6 x 6 x 9 x 9 set of combinations.... 2916 different combos. That's complex enough, but we're getting to the point where it is impractical.

If I may say, Azelor, you are focused on getting accurate at a given point, knowing the exact conditions, whereas I am focused in getting an overall map of the land. What are we trying to reach here?

07-31-2014, 06:24 AM
This post got me thinking about how short is a 6 tier classification for mean temperature... You are using 11 different ratings in this and yet, there are some temperature combos present in more than one kind of climate, which means it becomes useless for classification means if there isn't more information as the same temperature range can be found in two of the climates...

So I thought the way to counter this sort of lists where we get confused is to use a 2-entry table. So I sat down and made one, I went for an 8 tier classification, including your terms "cool" and "severely cold". More than 8 is too much in my opinion.
So this is it. This table would be the source for classifying the entire thing according to temperature (precipitation/humidity) would come at a later stage.


question 1:
what do you guys think of the use of a 2-entry table?

question 2:
and what about the actual key used to fill in each position?

cold deserts and cold steppes are a miss in this table, as their temperature ranges would be more in the tune with a D-climate - they would have to be determined separately.

note 2:
instead of terming the table X/Y axis as july and january, it could be termed as hottest month vs. coldest month, and then the classifications in each position wouldn't have to be symmetrical, allowing more flexibility - but at the same time, doubling the workload for the user, as it meant working the two hemispheres in separate.

07-31-2014, 12:18 PM
I do like your table, it look simple and clean.

The keys look alright to me so far.

I need to go back to what I said initially that was : when climates get colder, less precipitation is require to stay wet.

We have that formula: Precipitation= temperature/2 (+something)
So, every time you move by 2 temperature categories, the minimum rain required move by one
In photoshop, this could translate in having another layer. The original if for the total precipitation, we don't change that. The second and new layer is a modifier added to the original. It take in consideration that more rain is required in hotter climate to stay wet. The first serves only as a reference representing total precipitations and the second represent the ''wetness level'' or ''relative precipitation''?
It could be done by making colder climates appear wetter using the same color scheme as the original.

08-01-2014, 10:00 AM
We have that formula: Precipitation= temperature/2 (+something)
So, every time you move by 2 temperature categories, the minimum rain required move by one

It doesn't. It is a linear relationship, it changes the scale (temperature and precipitation are measured in different units anyway), but not the progression. The graphs you showed earlier (temperature vs. precipitation) had that straight line. Twice as much temperature requires twice as much precipitation for the same level of "wetness".

In photoshop, this could translate in having another layer. (...) . The first serves only as a reference representing total precipitations and the second represent the ''wetness level'' or ''relative precipitation''?

That's what I was trying with the "available humidity" map and the column called "humidity" in that reference table. Still, I admit having only arid/semi-arid/humid is too short to accurately classify climates - it's well enough to determine deserts but insufficient for anything else.

What should we use for precipitation level?

if I use the holdridge precipitation on the right combine with a possible equivalent on the left.
very wet/ super humid
wet/ per humid
moderate/ humid
steppe?/ sub humid
steppe/ semi arid
steppe/ arid
desert/ per arid
desert/ super arid

I'll try to come up with a second 2-entry table adding up mean temperature and precipitation pattern. I mean, if I understand your idea (and if this is it I am for it), we will have three maps:
1. mean temperature
2. precipitation pattern
3. "wetness level" / "available humidity" / "humidity" (pick your preferred denomination, I vote for "humidity")

Climate regions would then be determined by finding particular combos of mean temperature and humidity.

08-03-2014, 05:22 PM
Great effort there, Azelor. That will help a lot in developing that reference table I mentioned earlier. I am still most worried with the sheer number of possible combinations...

But, I thought of a possible solution. This is a question for expert PSP/Gimp users, as well... If there was a way to set up a filter, in photoshop, that would do a look up based on the color of a pixel in different layers and paint the result of the lookup in a separate layer (did I explain myself in a legible way?) - then we would automate a part of the process.

I mean, automating this tedious task:
- if #color in layer "january mean temperature" is X and #color in layer "january precipitation" is Y and #color in layer ".... (etc.) then #color in layer "CLIMATES" becomes ZZ.

08-03-2014, 07:43 PM
Well, if it is a script, a script it is. I might google for a "how-to" later on. But first need to work on that table..

08-09-2014, 08:09 PM
I like your idea.

We would need to have a key "color to climate type", wouldn't we?

I'm off on holidays in a couple of hours, but will get back to this in September. That is, if you didn't fully solve the problem by then.

10-23-2014, 07:08 AM
Oh wow! That's a lot of good work you did to try to figure out climates and how to represent them.

Unfortunately, you already know, I have no idea about it all, so I can only share my absolute wonder on your job. I wish Pixie was around to offer real input.

08-21-2015, 07:24 AM
Any breakdowns on this? I'm redoing the heightmap to my world and felt brave enough for a second to try and do the new climate myself, but I've re-read this thread and wow, it's scary. Any chance of a version for dummies? Is it even possible?

08-21-2015, 02:03 PM
I'm still working on it but it's taking more time than expected.

I might need a guinea pig, are you interested?
Your using photoshop right?

08-22-2015, 04:40 AM
I do. I'm more than willing to give it a try, but keep in mind how very challenged I am with all the climate stuff, so you'd have to be patient with me.

08-23-2015, 10:40 AM
I, for one, am very curious to see what you two come up with. :)

08-24-2015, 01:56 AM
If you still interested Azelor, my new height-map is almost ready. But for the climate stuff I have to redo the winds and the other stuff right (redo is a funny word, since you did it for me the first time)? That'll take me some time to figure out so I hope the thing I'd try for you isn't for this month's challenge.

08-24-2015, 03:23 PM
I made a cleanup in order to start anew, thus I deleted, merged, truncated several posts. The method has changed.

08-25-2015, 01:08 AM
Now, I will start anew. Well, not really since I intend to keep some of the old data and improve it.
In my attempt to work on a method based on Pixie's tutorial I came up with the idea of recreating the climates of Earth using the tutorial to see what needed improvement. In order to create a realistic climate what model could be better than Earth?

The first attempt failed, mostly because my working method was bad and I used unreliable data. I tried to improve it but it failed again.
For the third attempt, I decided to directly use layers from real data instead of trying to make my own based on pictures found here and there. It does look much better as you will see later but there is still place for improvement. Eventually, I might use better precipitation maps but for now, those that I picked will do just fine as the data is the same (but it's lower quality).

The source of the data is include in these pictures.

The elevation map will be useful later. I have several useless layers in the low altitude, but without them, the map looked too boring.
Temperatures for January and July. No data for Antarctica.
Precipitation for January and July. I might use a better quality data later but this is still good, just does not look that nice.

How to determine if one area is dry ? Using Excel

I changed the precedent formula to this from the Trewartha Wikipedia page:

2(10(T − 10) + 3P)

with T equalling the mean annual temperature in degrees Celsius and
P denoting the percentage of total precipitation received in the six high-sun months (April through September in the Northern Hemisphere and October through March in the Southern)

if evaporation is smaller than the average precipitation = humid
if evaporation is bigger than the average precipitation but smaller than the max = steppe
if evaporation is bigger than max = desert

Thus if one area has an average temperature of 35°C : severely hot + severely hot (probably impossible on Earth) the default minimum precipitation required to assure it's humid is 10(35 − 10) + 0 = 250
Then, we add the % of yearly precipitation falling in the 6 hottest months and multiply by 3. If it fall evenly during the year, the % is 50. So ... 250 + 3(50) = 400
But we still need to multiply by 2, to get 800.
So if precipitation are under 800, it's not humid.

The average yearly temperature is based on the fact that January and July are always the coldest/hottest months and that the temperature change at a constant rate at each month (of course, it's not like that in real life). Thus, in order to get the yearly average temperature I make the average of the lowest and the highest. As simple as that.

For the precipitation, in order to have the yearly precipitations, I make the summation of the precipitations for summer and winter (by previously multiplying the precipitation of January and July by 6).

Each temperature combo has a different aridity threshold but many do share the same numbers because it's based on the average temperature. Once we have the individual threshold for each (250 for severely hot + severely hot),

we need to figure how much rain each precipitation combination will give and how they are sprayed during the year.

For example, a very wet area receive between 100 and 200 mm of rain in summer but less than 10mm in winter. I need to find the average precipitation here. To do that : min + max/2

min = (6*100mm) + (6*0mm)
min = 600mm
600+1260/2 = 930mm that's the average yearly precipitation for one of the 36 rain combinations.

When we have all the 36 averages, we need to find out the T from the formula. Average summer precipitation/ total precipitation.
Total= avg summer + avg winter
avg for a particular season = min+max/2

The example above would give : (100+200)/2 + (0+10)/2
or just divide 930 by 6 = 155
Now (100+200)/2= 150mm falling in summer (July)
so 155/150 = 97% of the precipitations fall in the summer

Thus the total of the Threwartha formula is 2(10(T − 10) = 800
and we add 800+3P = 800 + 3(97)
= 1091mm , that is the precipitations required to keep the place humid.

All that give us the requirements for severely hot + severely hot but just for 1 or the 36 rain combo...
When all the combos are done, we need to do it for the other temperature combinations ( some will share the same numbers)

With that table, we compare the numbers with the previous table (the one indicating the total yearly precipitation, to find when a temperature combination is humid, steppe, or desert.
When that is done, it gives me a nice huge table with 3600 little square full of colors!

Next, it's with Photoshop. Ideally this process from here can be automated with actions/scripts. That's the step 7 of Pixie's tutorial. A brief description:

What the script does:

-It start by making some dummy layers and adjustments to the different maps for technical reasons in order to be able to select all possible combinations.
-Select the different temperature combination and create a new layer for each.
-Using the temperature layer, he regroup them to create the major temperature zones from Koppen: A,Ca,Cb,Cc,Da,Db,Dc,Dd, tundra and ice caps. The last two don't need to take in consideration the precipitation, so they are done at this stage.
-Select the different rain combination and paint them in black on one layer each.
-Regroup the rain combo with all that share the same characteristic for when a specific temp combo is arid/or not.
Most of the rainy combo will always stay wet no matter how hot they are, I put them aside on a specific layer
Others are selected one by one (or several if they share same characteristics) and then, intersect them with the temperature combination that are not humid in order to separate them on a new layer. So, the part that was not selected should be made of only humid
climate. When we are done with this, we can put the remaining areas with the always wet layer created earlier.
-Now we know where the desert, steppe and humid climate are. We can separate the cold from the hot arid climate or not because I don't think it's really a big deal.
-Taking the wet climates, we intersect the layer with the temperature groups (A,Ca,Cb...) and create a new layer for each. Or we could just delete the arid part based on the arid layers, but it's the same result.
-When it's done, we still have to figure whether it's summer dry, winter dry or normal. Normal (f) have no dry season and it's when the precipitation for each season is no further than 1 category apart. For example: Category 5: over 200mm and Category 4: 100-200mm.
I made a simple table for this and it's nothing complicated. Winter dry have very low precipitation in their dry season, often lower than the lows of summer dry.
-With these w,s and f layer, we separate the temperature groups into Cfa,Csa.Cwa,Cfb...
Now we have the actual climates
-Some corrections to the map seems inevitable. Some deserts are surrounded by humid climate with no transition. Usually it's because the transition would occur inside the driest precipitation category. But since I'm using the average, there is only one value abd the transition is sometimes impossible. We need to add the steppe generally outside of the desert. I will come back to this later but the process is not perfect and still need some tweeking.

The result, I had to make the map 20% smaller to make it fit:


08-26-2015, 03:24 AM
The result is pretty damn close to the Koppen maps of Earth, almost exactly! So you're definitely on the right track!

I'll wait for your instructions to do the temperature and precipitation layers then. I'm still finishing the redo of the coastal shelves.

08-26-2015, 04:23 AM
I agree completely with groovey. The differences with the real climate map of Earth are minimal. Astonishing!

Do you need another guinea pig? I have height, ocean currents and wind maps ready, but I'll have to redo temperature and precipitation maps. I'd be interested to follow your procedure and see what the result of it is.

Cheers - Akubra

08-26-2015, 01:51 PM
You are clearly aiming at a high level of accuracy. That is awesome but makes me wonder how do you produce temperature and rain maps with that level of certainty for an imaginary world.. can't wait to see the rest of your tutorial.


08-27-2015, 12:19 AM
Can't wait either, it's better be amazing!
The problem is that I started with the last part of the process (your step 7 if I recall).
Now I got to figure how to make the first part fit.

08-28-2015, 12:44 AM
Some temporary notes, the post will be deleted eventually.

Summer north

1- Green is too cold in high latitudes during the hemisphere's summer.
2- yellow is the default temperature on the coasts above 70.
3-Peach become yellow on coasts above 55 if next to a cold current. Inland could remain peach (boundary limit unknown)

4- The North Pacific current is colder than the Gulf stream. It seems to be closer to the pole and also more parallel to the equator. While the Gulf stream moves more form south-west to north-est. Also, the Pacific ocean is larger and give more time to dissipate the energy before reaching the coasts. It would explain why Alaska is colder than Norway is winter and why California (coast) is colder than Portugal. Also, the winds are from different directions.

7- Haven't thought much about it yet but the red colors are for places with either a low cloud cover as in most case (and receive more energy from the Sun), with lower thermal exchange in the Red Sea (and to the opposite of Antarctica, then have hot waters) or with hotter water temperature as with the Philippines or India.

Northern winter

The Gulf stream pushes the yellow area almost to the Polar circle but only affect coast at that latitude. Actually, the winds play a larger role than the water but the winds are taking the energy from the ocean...

5-Green: it looks like area near the water are warmer than inland because the temperature of the water is hotter. If so, only area at high latitudes will be blue. If there is an exchange of water north-south near the coast, it's green. As I mentioned earlier, in winter, the Arctic is trapped in ice and there is less energy exchange. Furthermore, the ice reflects the sunlight and contribute to the lower temperatures. So, I guess the Arctic need to be considered as a continent because of this? (when considering whetter or not I should apply a continental effect)

08-31-2015, 01:35 PM
Realistic temperature map

temperature colors:

Severely hot : over 35 °C
Red: Very hot : 28 to 35°C
Dark orange: Hot : 22 to 28°C
Orange: Warm : 18 to 22°C
Peach: Mild : 10 to 18°C
Yellow : Cool : 0 to 10°C
Green: Cold : -10 to 0°C
Turquoise: Very cold : -25 to -10
Blue: Severely cold : -38 to -25°C
Deadly cold : below -38°C

The different areas of the world can be classified in several categories, all with unique temperature placement for both seasons depending on the influences they have. This categorization changes according to the season mostly because of changes in wind patterns: Hot current, Mild current, Cold current, Normal (no particular influence), Continental, Continental plus

Hot current: Most of the currents between the tropics are hot but have no impact on the temperatures (well yes but I don't want to go into the details)
Hot current have no impact in summer since the land is already hot.
The Gulf stream and the North Atlantic current are good examples of hot currents outside the tropics.

Mild current, make the climate milder. They are usually hot current that have cooled over time or when a cold and a hot current meet such as south of Greenland. The California current and Alaska are makes the west coast almost homogeneous in terms of temperature.

Cold current: come from polar region, flowing toward the equator. Although they are cold, the land around them is usually above the freezing point. The Humboldt current is the coldest one.

Normal: used when there are no particular influence in the region. It serves as the base temperature for comparison. An example would be inland Brazil. It's too far inland to have a maritime influence but it's not large or hot enough to be continental.

Continental: occur on large landmasses and in Australia(?). They have large temperature gradient often because the winds are blowing from deep inland. New England and Japan in winter ae considered continental.

Continental plus: Is an extreme version of the above. I could probably make a plus plus version but you get the point. Larger landmass, and temperature gradient. Being far from the water is a plus and high cold plateaus in winter also help.China in winter and Central Asia have extreme temperatures and are con+.

General temperature placement: should be at 10º N during the northern summer and 10º S during the northern winter.
it seems to work fairly well if the oceanic currents and major winds are done correctly.

*areas between the tropics are not always mentioned
*extreme temperatures are not included

Winter temperature placement: the numbers between the color names indicate the boundary between the 2.

Hot current: 20 orange 30 peach 40 yellow 65-70 green
Mild current: peach 35 yellow 55 green 65-70 yellow
Cold current: dark orange 5 orange 10 peach 35 yellow 55 green
Normal: dark orange 20 orange 25 peach 35 yellow 55 green ...
Con: peach 30 yellow 40 green 45 turquoise
Con+: dark orange 15 orange 20 peach 25 yellow 35 green 40-45 turquoise 50-55

Specification for winter: Watch out for mid latitudes where the hot poleward current encounters the cold descending current. (Newfoundland, Kuril's islands) have a milder climate. it could be for different reasons but just bend the color poleward. Green til 50.

Greenland is not as cold as it should be, because of the water nearby and the presence of extra-tropical cyclones. Ok...Does that mean that the Eastern Greenland current is mild?

Summer temperature placement: red are only guidelines

Hot current: Is considered normal
Mild current: Red 25 dark orange 35 orange 40 peach 60? yellow
Cold current: dark orange 20 orange 35 peach 55 yellow
Normal: Red 30 dark orange 40-45 (med) orange 45-47 peach 65-70 yellow 70? green
Con: Red 35 dark orange 45 orange 55 peach 0=
Con+: Red 45 dark orange 50 orange 60 peach 70 yellow

If one temperature is indicated to be hotter than the area equatoward of it, usually, it make sense to change to temperate to a lower category since places closer to the poles are logically colder.

Antarctica: is a special case.


09-01-2015, 10:56 AM
You're doing an impressive and very detailed work on this Azelor!

09-02-2015, 01:49 PM
You're doing an impressive and very detailed work on this Azelor!

Thanks! Hopefully, it will be simple enough to be considered manageable for those interested.

I'm now working of the altitude vs temperature. It does not give precise results and I don't think it will. Explanation:

I'm using temperature average, not the exact values. The temperature decrease slowly with elevation. If a category includes temperatures between 22 to 28, and the other is between 18 and 22 (for example), the change in the temperature color represents the change from 22 to 21. It's a small change normally. But since I'm using categories, I have no idea what the exact temperature is. Therefore, it's a change between the average of 22-28 (25) and 18-22 (20). From 25 to 20.

It means that the "lowest" temperatures of the first category sometimes appear hotter that what they should be in reality and some areas of the latter category will appear colder. But I have no idea how to do otherwise.
I was thinking I could select an elevation and expand the selection but some area are steep, others are flats, I have no idea what the exact temperatures are and the temperature variation is not the same everywhere.

What I'm going to do is just to use this simple formula: sea level temperature - (6.49 °C/1,000 m)

Using the average of each temperature category, lower the temperature according the the altitude. When the temperature reaches the average of a lower category, lower the temperature to that category.

Unless someone has a better idea?

EDIT: I'm mostly done with the temperature part. I just need to figure how to place some of the most extreme categories and make a final version of the instructions.

09-03-2015, 03:07 PM
This is the result. It still need to patching but it's looks pretty god. I just need to make a final version of the instructions and maybe some adjustment to normalize things.
Then, it's onto the precipitations but I think I will take a break before starting the next step.

temperature maps for January and July


09-04-2015, 12:39 PM
I'm a bit surprised, are the temperature that hot in January, in the south Bolivia/NW Argentina?

09-04-2015, 01:10 PM
Yes but actually, it's hard to find data to confirm it. Several places in Argentina are on the fringe between 27 and 28. Paraguay has a few over 28 but barely (28.2).

Mysterious Mapmaker XXIII
09-10-2015, 11:37 PM

I'm a conworlder looking to decide my conworld's climates some day, and I love what you're doing with this tutorial! I can't wait to see the full, re-worked version! Especially since it'll be the first version I actually see. :P But honestly, this looks like what I've been looking for in regards to actually doing my climates.

Just for the record, are the bits on page 4 actually part of the tutorial, or just notes? Just trying to get a better idea of what's going on here.

09-11-2015, 10:10 AM
It's more like a draft.

Mysterious Mapmaker XXIII
09-11-2015, 01:47 PM
Ah, okay! Looking forward to the full version!

09-18-2015, 01:41 PM
Some details, some are more evident than other but might still be educative:



Rossby wave. While not unique to America, it have a unique characteristic one this continent as explained in this article : http://www.americanscientist.org/issues/feature/2006/4/the-source-of-europes-mild-climate/3

The Arctic is cold and trapped in ice with low evaporation. A high albedo: reflect a lot of energy to space instead of absorbing. This result in low humidity.

In summer:

One things I haven't really figured out is where the influence of the front ends. I believe that it extends farther south that it's counterpart in Europe, but I don't know where it ends.
America is large enough to create his own modest monsoon. Mexico can be pretty rainy. There is the altitude but there is also a low pressure system around. It's not the ITCZ but possibly and extension of it? Anyway, all of central America is really wet.

July: during the cold season, the coasts receive some rain but it's not much and it's mostly if they are at right angle with the winds coming from the sea. The rest is dry except for the ITZC. Again, the rain belt is mostly a straight line. The Horn is still dry because it receive winds from the interior of the continent blowing toward Arabia...

South America:

January: A lot of things are still obscure about this continent. It receive more rain than Africa. Larger landmasses tend to become hotter and become larger low pressure center. This and apparently, the Amazon forest generates 50-80% of it's the region's precipitations. I've found this data on Earth Stackexchange but, I'm not sure the source is good. The forest must help keeping the place more humid but maybe the influence is smaller.

The reason why there is a larger forested area in SA compared to Africa is probably because the variation of the ITCZ is smaller than in Africa, the center of the continent seems closer to the equator. Just some theories.

Aside from that, Eastern Brazil is drier because it receive drier air from the high pressure system. That relation is inversed in winter but I'll talk about it later.

Mysterious Mapmaker XXIII
09-27-2015, 01:12 PM

Is this to be part of the guide, or just ancillary information?

09-28-2015, 08:44 PM

Is this to be part of the guide, or just ancillary information?

I don't know yet.

I might have figured something out. Basically, we can estimate the quantity of rain mostly based on the air pressure and direction of the winds. And with a combination of other factors such as the distance from the water bodies.

10-06-2015, 03:41 PM
This is a rough version of the tutorial I'm working on. The rest will follow in another post.

Prerequisites to this guide

• I made this guide for earth-like planets.
• This tutorial is mostly a follow-up of the tutorial made by Pixie : http://www.cartographersguild.com/regional-world-mapping/27118-wip-sort-tutorial-climates-applying-geoffs-cookbook-detail-some.html
• To get a better understanding on climates you really need to read this, the climate cookbook: http://web.archive.org/web/20130619132254/http://jc.tech-galaxy.com/bricka/climate_cookbook.html

Software: You can use any software you like to do this but using multiple layers is almost a necessity. I would recommend Photoshop or Gimp. Personally, I used Photoshop CS3.
I have worked on a script to make step 7 easier. It’s not impossible to do without the script but it’s much faster/easier since it’s automated. I haven't tested it much yet.

Map projection:
Examples where made with the Winkle triple projection, it’s the same projection used by National Geographic. It’s useful because it minimizes all the types of distortions. But you can use other projections too. The euqirectangular is useful if you want to convert it to other projections later. Winkle triple cannot be converted into another projection, use it only if you’re sure you won’t need to convert it later. Otherwise, you’re better to use equirectangular.

Summary of the tutorial:

1. Base elevation map, one level for each 1000m of altitude ideally
2. Ocean currents map: one map will do, currents do not change over the year.
3. Atmospheric pressure systems: these air masses have a big impact on weather; you need one map representing January and one for July.
4. Dominant winds: you need one map representing January and one for July.
5. Temperature map: you still need to make 2 maps
6. Rain pattern: again you need two different version of the map.
7. Climate map: magic!

Step 1 basic elevation map

The first step is to make the elevation map. I made one level of elevation for each 1000m above sea level. I added several others layers of elevation under 1000m but they are mostly aesthetical. It can also be a good idea to include the continental shelf of the oceans because it can have an impact on the oceanic circulation if you have shallow waters.

And by the way, it will also help a lot if you have latitude lines since we will refer to them all the time.


Great, I made a typo in the title.

Step 2 oceanic circulation:

Here are two reference maps for oceanic circulation:

A simple one: http://www.google.com/imgres?imgurl=http://science.kennesaw.edu/~jdirnber/oceanography/LecuturesOceanogr/LecCurrents/0908B.jpg&imgrefurl=http://science.kennesaw.edu/~jdirnber/oceanography/LecuturesOceanogr/LecCurrents/LecCurrents.html&h=618&w=1000&tbnid=130jbciJH-gicM:&docid=wlG_uioTGw9N9M&ei=LfLcVbW5EMvp-QHptKDoDQ&tbm=isch&ved=0CB0QMygAMABqFQoTCPWDzfKpxccCFct0PgodaRoI3Q
And another, more detailed: https://upload.wikimedia.org/wikipedia/commons/6/67/Ocean_currents_1943_%28borderless%293.png

We will do the surface currents only. Deep currents are also important but their dynamics is not yet fully understood and their impact on the climates of specific areas is not clear either. One thing worth mentioning about deep currents is if you don’t have north/south ocean flow, it means that the oceans and surroundings will be more affected by the temperature of their latitudes. Equatorial waters will be hotter while polar ones will be colder since there is no heat exchange between cold and hot waters.

Color signification:

• Current parallel to the latitude lines are black, temperature difference between water and land is minimal, resulting in a marginal heat exchange
• Current flowing poleward are red, the water is hotter than the land, especially in winter.
• Current flowing equatorward are blue. The water is colder than the land, especially in summer. (It makes the temperature colder compared to other areas at the same latitude)

Mapping the currents:

1. Close to the equator there are usually two currents flowing westwards. Between the two of them is a counter current flowing eastward. You can ignore this current because it’s much weaker the other two and it’s also a more seasonal current depending on the monsoon winds. Draw these currents in black.

2. As these two currents meet a large landmass (or shallow waters) they get diverged away from the equator, flowing toward the poles. They will stick to the eastern coast of that continent approximately until 40º to 45º of latitude. They are hot because they bring hot water into cooler places: draw them in red.

3. At about 45º, the Westerlies (West -> East winds) are strong enough to create an eastward current. However, this is no sudden right angle turn, it actually starts much sooner. The warm current you drew earlier is now cooler and it’s now moving probably at more or less stable latitude. The exchange of heat is decreased: draw it in black.

4. When this current meets continental shores, it spreads north and south following the coasts. The currents moving back to equator will take reasonably cold water: draw them blue. The part that flows north takes relatively warm water to the poles: draw in red.

5.Poles : All you need to do if you have polar oceans/seas, is to close the loops. Unless you close every loop, you are saying water piles up somewhere, that’s wrong. Cold current flowing back to mid latitudes tends to stick to the coast until they encounter the hot currents mentioned at point 2. They fill the void left by the north-eastward movement of this hot current and then turn abruptly to the east. In the current map, see the Labrador Current east of New England.

Near the poles, the currents are flowing from the east to the west. This only happens at very high latitudes (over 70 degrees approximately) , where the polar easterlies are the dominant winds.

Circulation around the North Pole: http://www.google.com/imgres?imgurl=http://www.grida.no/prog/polar/bsc/pict/arccur.jpg&imgrefurl=http://www.grida.no/prog/polar/bsc/fig5.htm&h=540&w=509&tbnid=b1AGtmqNM8sxoM:&docid=2dKTOAbfx1CKZM&ei=qffcVcPICsex-wG6jp6IBg&tbm=isch&ved=0CB0QMygBMAFqFQoTCIPBhJCvxccCFcfYPgodOocHYQ

I made all the current with their respective colors but it's clear that the equatoward currents on the coasts of South Asia are not really cold.


Step 3 : Atmospheric pressure systems

Reminder: Pressure systems depend on temperature. Hot air means rising air and low pressure at sea level, while cold temperature make the dry air sink and pushes it away. We don’t have the temperature maps yet but we can figure how where the hot and cold spots are. The hottest place is near the equator, where it is the hottest. It’s the Intertropical convergence zone, ITCZ for short. The poles are the coldest. Air rises in the ITZC and fall back at the Poles.

The position of the pressure systems changes over the course of the year, with the ITCZ roughly located where the planet receive the most energy from the star. It should be close to the tropic n the summer but the position is influenced by the actual temperatures. Land heat up more than water and large landmasses will pull it to them. This mechanism is the main engine of the atmospheric circulation.

Look here for a more detailed explanation of atmospheric circulation: https://en.wikipedia.org/wiki/Atmospheric_circulation

The ITCZ, a low pressure system

The position of the ITCZ is close to 10-15° normally since it’s roughly the center (hottest) point of the solar radiation. During the summer, the movement of the ITCZ is strong in Asia but limited elsewhere. In order to have an impact, continents need to be large, hot, with significant landmasses around the tropics.

The position of the ITCZ is not clearly defined. The Intertropical latitudes are always hot and therefore, there is a consistent low pressure system. Since the water is cooler than the land at these latitude, it’s not as low but it’s still making the air converge smoothly.

The relation between the size of the pressure system and the size of the continent is not linear but exponential. If a continent is twice as big, it will have a low system 4 times larger (my theory).

A. High pressure center


In the oceans, they tend to be located on the eastern side, next to the continents because it’s where the cold currents are flowing and cold mean higher pressure here. The pressure lessens going west as the ocean is less affected by the cold. In winter, the high pressure systems tend to make a more or less continuous band around the planet. In summer, the high pressure system breaks apart as the continents are affected by low pressure systems due to hotter temperatures.

Localization over the oceans (large oceans don’t have larger zones)
Cold season: 30° in a more or less continuous line
Hot season 35° separated, mostly on the eastern side of the oceans


Cold season: high pressure systems develop over the continents.
Hot season: hot temperatures prevent the formation of high pressure systems.

The larger and the colder the continent, the larger the system will be.

The inland systems must be poleward of the high pressure systems that are over the ocean (unless they are very large like in Asia). Areas between the tropics are not cold enough to become high pressure although; it does not mean that the pressure is low.

The Mediterranean prevent (lessen) the formation of a high pressure system in winter. The temperature cools off less rapidly than inland locations and thus contributes at preventing the formation of what could be a high pressure system.

Furthermore, during winter, another high pressure system develop over the poles, over 70°
It is much weaker in summer.

B. Low pressure systems:

Cold season: No low pressure overland.
Hot season: large landmasses become hot and this makes the pressure drop. It can cover most of the continent if it’s large enough. Moderate landmasses have small to moderate low pressure systems unless most of the land is concentrated between the tropics.

There are other low pressure systems in the higher latitudes. They are at their strongest in winter and fade away in the summer. In winter, they are centered a couple of degrees south of the polar circle (50-60). They move north in summer following the movement of the sun. They move by about 5 to 10°. They simply disappear in the north because it’s mostly land.





10-07-2015, 10:46 AM
Right on, Azelor! There are valuable additions to the method and wording so far. I'm certain this will turn out (finally) to be a self standing complete tutorial.

... And there's a good chance it will become a sort of Geoff's Cookbook 2.0 over the internet ;)

10-07-2015, 04:23 PM
Step 4, winds:

Use the pictures to figure out how the winds are blowing. Figure A
All the figures are from the North Pole except E.
Or use the main map at the bottom at the page.
I'm having an easier time to explain it with images than with words.


Dominant winds

• Near the equator, the dominant winds are usually blowing to the west (Trade Winds)
• In mid latitudes, it’s blowing to the east (Westerlies)
• And close to the poles they are blowing to the west again (Polar Easterlies)

Wind usually flow from the highest pressure to the lowest. The larger the difference in pressure between two areas, the stronger the winds will be. Inside a large high or low pressure zone, the winds can be very weak. Figure B

Mid latitudes winds (most common): Starting with your low pressure bands at high latitudes, the Westerlies should blow from west to east where it’s blue, where the pressure is lower, where there is mostly water. They do tend to converge but not as much as with the ITCZ where the air look like it’s sucked out to space by a vacuum cleaner. The Westerlies should extend to much of the area till the high pressure systems.

Low pressure have 2 types:

• Those that really converge like the ITCZ in Asia, see figure F
• And those of the mid and high latitudes, the North Pacific (round , isolated) and the North Atlantic (continuous band of low pressure) Figure G

The high pressure systems at mid latitudes (also called subtropical highs):

• Draw the poleward winds first. They have a curved shaped because they quickly change direction when encountering the Westerlies. Figure C
• Winds blowing from the equatorward side of the system tend to blow toward the equator, or if any, toward the closest low pressure center. Winds are converging near the equator; they tend to blow to the west but not always. Figure D

Polar highs: The high pressure systems on the poles brings dry and extremely cold temperatures. Figure E (South Pole)

Overland highs are spinning according to the schema. See Figure A, or look at Eastern Asia in winter.

January winds


July winds


So far, is it easy to follow?

10-12-2015, 06:08 AM
Excellent! I guess it's time I get my height-map ready...again. Will let you know when I'm ready to start trying these steps.


Hi Azelor.

I've started running your revised tutorial. I'm quite confused at "Step 3 : Atmospheric pressure systems":

For example, at:

"B. Low pressure systems:

Cold season: No low pressure overland."

Bu then, in the January map you do paint low pressure (blue) on land.

By the way, I didn't catch anywhere that altitude has much of an influence, so I wonder if these old preassure maps you made long ago (http://www.cartographersguild.com/showthread.php?t=27111&page=26&p=280569&viewfull=1#post280569) are still valid?