Theia revisited - Climate check (WIP)

[nwisth]

Another update. I shouldn't have, but the unfinished climates are bugging me to no end. So, I installed GIMP and downloaded Charerg's 6-step script, hoping that would help. I got the script running - and even hacked off a pixel in the top left corner, as instructed. The script started strong, creating a few more layers, but then I didn't really see anything happening. Task Manager told me GIMP kept using between 4 and 20% of my Xeon processors, and about 10GB of RAM - but after about forty minutes it gave up, leaving me with a white map with a couple of purple patches over the driest, hottest places, and grey over the absolutely coldest regions.

Obviously, I'm doing something wrong, since none of the three script alternatives work. :/

I think it has to be the colors. I've copied and pasted example maps from the Köppen tutorial thread into Photoshop and used the color picker to create my swatches, but now that I check I see that it hasn't been precise enough. Well, I haven't given up! Luckily, Charerg posted RGB values for all zones in his script post, so I'll add them manually, repaint everything using PENCIL, and try again.

-Niels

[Azélor]

Personally, I haven't tried that version of the script yet.
I assume that missing colours would simply be ignored. It seems that it is not the case.
It was done by Azure wings, hopefully he knows why it doesn't work.

Always paint with the PENCIL tool

I think I mentioned it very clearly in the tutorial.
I guess it's a no-brainer for anyone well versed to graphic software but it's not for those new to map making.
The paint tool produces a lot of anti aliasing. It look smooth but the result is that only the pixels close to the middle have the right colour. The rest is a gradient, useful to make transitions.
The pen uses sharp colour with no transition /aliasing. It's usually ugly so we avoid using it. But in this case it's what we need.

[Charerg]

I took your source maps and processed them through the 6-step GIMP script (after manual modification to remove artifacts and correct the RGB values of the colours).

First off, some notes:
- Your temperature maps are completely missing the "Warm" (18-22 °C) category, effectively skipping this completely. I compensated for this by converting "Mild" into "Warm" and similarly ramping all lower categories up a level.
- Your precipitation maps, on the other hand, have 7 categories instead of 6. I merged the two highest categories into a single one.

Here are the source maps after the aforementioned changes:

 

JanTemp.png

JulTemp.png

JanPrec.png

JulPrec.png

And the resulting climate map (I added the sea colour and the land border manually, those are not created by the script):

Theia Generated Climates.png

Basically take a look at those source maps and see how they look like: they include ONLY the relevant colours (with ocean as transparent). I suppose I should have emphasized that your source maps should not include any other colours except the ones listed, otherwise the results may be wonky.


Now, regarding the climates themselves, as you see the results are a bit strange. What strikes me as odd is the extreme temperature gradient: some areas are extremely hot but transition very rapidly into extremely cold regions. I haven't followed the thread closely, so I'm not sure if there exists a reasoning for this, but at first glance this strikes me as inaccurate in comparison to Earth. Particularly striking is the almost complete absence of continental climates, with deserts and steppes going all the way to the tundra (though this may be due to the modifications I made to the temperatures to add the missing category, not sure).

[nwisth]

This is awesome - thank you very much, Charerg! I agree that the result is a bit strange - omitting the Warm category was definitely not intentional, and I also need a thorough rework of the precipitation patterns to take into consideration the high polar pressure and strong circular winds Theia's large ice caps would result in.

With your example maps I am sure I will be able to re-create working files and get the script working myself.

-Niels

[Azélor]

There is a blue climate in the north around the tundra, I don't recognize the colour. Is it Dw something?

Yes, some transitions are clunky.
From Cfa to Dfb to tundra in a short distance.
It is the right order but missing transition climates. The more plausible transition should be : Cfa, Cfb, Dfb, Dfc, tundra
Maybe this script is outdated?

There are places where C climates are surrounded by lowland D climates despite being at higher altitude and normally colder.
Other places where hot climates are surrounded by much colder climates but I assume these are mistakes or forgotten areas.

[nwisth]

I'm sure the script is fine - it is I who have made a ton of noobish mistakes. But, seeing the climate results makes it a lot easier to adjust things and figure out how to make it right!

So far, Theia seems a lot hotter than I envisioned. I'm getting closer - I've gotten mediterranean climates in a few of the places I hoped, but the deserts are a bit too large, and I'd like a bit more boreal forests and subarctic climes around the northern part of my active continent (the third one from the left, in the northern hemisphere). I've made the northeast coast too hot, over playing the warm water effect, and the eastern area that's desert now ought to be steppes of some sort. I guess I need to make it a wee bit wetter and colder.

It is super exciting to see results, though, and I'm really thrilled to be getting all this help from the resident climate experts.

-Niels

[Azélor]

Hotter? You have 2 huge ice sheets and extensive tundra extending well into mid latitudes. That is not hot compared to Earth at least.

One thing I notice is the relatively sharp gradient between the equator and the poles. I assume the poles to be very cold because of the large ice sheets but at the same time the equator temperatures looks similar to our ours, which makes a rapid transition from torrid to frigid climates. Did you say you have a low axial tilt or a large one?

[nwisth]

I have an Earth-standard axial tilt, but I'm trying to figure out how much colder the ring would make the winter hemisphere. The effect is close to zero near the equator, but ramps up fast as you go north - so I'm going for a sharp gradient. This is especially noticeable due to the small size of the planet - it only has a circumference of 28400 km, which means the distance from the 45th parallel to the equator is just 3550 km - the distance from Oslo to... well... somewhere in Sahara.

I've placed the ring shadow during winter solstice and the scales on this height map, to show it:

Theia-Equirectangular-ringshadow-and-scales.jpg

I need to find a better technique on how to effectively distribute the temperature scales. Maybe using some sort of grayscale layer, which can be affected by both elevation and other manual touches... How do you guys do it?

I'm thinking the ice sheets would melt quite a bit in summer, and re-freeze in winter - but then again, the ring isn't totally opaque either. It would filter out somewhere around 30-40% of the sunlight, I think (all I can say without spoiling too much is that it's not made of rocks).

-Niels

[Charerg]

There is a blue climate in the north around the tundra, I don't recognize the colour. Is it Dw something?

I think it's Dwb.

It is the right order but missing transition climates. The more plausible transition should be : Cfa, Cfb, Dfb, Dfc, tundra
Maybe this script is outdated?

The script should be up to date (it's the latest of my 6-step scripts uploaded in the main thread of the tutorial), in this case the source maps are probably somewhat off (especially since I modified them, ramping up the temps to account for the missing category in the originals).

There are places where C climates are surrounded by lowland D climates despite being at higher altitude and normally colder.
Other places where hot climates are surrounded by much colder climates but I assume these are mistakes or forgotten areas.

The January temp map has the western coast colder than the interior (which is inaccurate, you're correct that the coast should have milder temperatures), hence the C climates turn up in weird places in the interior (keeping in mind I ramped up most temps to one category warmer to account for the "warm" category that was missing in the originals).

@nwisth
Btw, in summer the coast should be milder and the interior hotter. Right now you have the hottest temperatures basically along the coastline, which is inaccurate since the sea is a very strong moderating influence.

Edit:
And about generating the temperature maps, yes, a grayscale base layer with an elevation adjustment layer on top works (and using a gradient to convert the grayscale into the desired discrete temp categories), assuming you're familiar enough with your image modifying software to use those features.

Alternatively just manually painting the temp categories works too. However, both methods require a degree of understanding about the basics of how temperature is affected by the various factors like oceanic currents and continentality. Taking a good look at the sample maps of Earth helps (see this post for the 1970-2000 temp maps, for example).

Edit2:
A slight aside about the ice sheets: note that ice itself reflects ~90% of sunlight hitting it back to space (it has a high albedo), so vast glaciers would have an immense effect on the climate all year round. It's also worth noting that the size of glaciers does not notably change on an annual basis: vast glaciers take millenia to form, and a long time to melt as well. Even with the present global warming going on, it is estimated the 2-3 km thick glacial cover of Greenland could take a thousand years or more to melt entirely.

[srm038]

By counting hexes and using historical European population figures from 1000-1500, I found the population density my various countries could support, and have made an enormous excel sheet which calculates everything from average distance between villages to happiness scores. But that's probably not something you want to hear about on a cartography forum.

I certainly am interested in this bit.