Gigantica -- A Study in Maximum Habitable Planet Size

**MrBragg** · 06-19-2022, 09:42 AM

I don't have any advice on how to help things, but I do want to point out that your Koppen map doesn't seem to necessarily correspond to your temperature maps. Your northern summer map has >10 C temperatures extending well north of 30 N (up to ~45 N it seems in some places), though the Koppen map shows almost everything north of 30 degrees as ET/EF, which can't be the case if a month has an average >10 C (>0 C for EF!). Similar conditions exist in the south, if anything it's more prominent there with nearly all the land being >0 C in the winter. You may want to look into why this is as, if the temperature maps are correct, you'll have way more land with less hostile climates than you do now.

[Edit]: Also wanted to say thanks for posting all of this information. It interesting and really helpful to see how much the resolution affects the result that you're getting.

[Edit 2]: It also looks like your latitude scales are messed up in your most recent outputs as they start and end at 60 N/S even though the displayed maps seem to go from 90 N/S based on your previous posts, so something funny might be going on with whatever scripts you're using.

**worldbuilding pasta** · 06-19-2022, 08:23 PM

Yeah there's definitely something up with the koppen and temperature maps there. Are you sure these are from the same file? Are you pointing the koppenpasta script to a specific .nc file from the end of the model run or to a folder containing the entire run?

Also, I forget I hadn't uploaded the latest version of the eps_avg script I was using, that should be fixed now; and I forget to mention it needs nco to run properly (both the nco package in yast and the python package you can get with "pip install nco").

Also also, I think the wind map you're using there might be from the top of the atmosphere; some of the climate data is recorded for every level of the atmosphere in the model, and Panoply will open to the first level at the top of the atmosphere by default.

**Peter Toth** · 06-25-2022, 01:40 AM

Thank you MrBragg and worldbuilding pasta, for diagnosing the problem plaguing my climate zones!

I've done two things to remedy the situation: first, I've re-run the koppenpasta algorithm using ONLY the very last (MOST.0009.nc) file; second, I've used my own algorithm to generate an alternate Koppen map for the purpose of cross-checking. Amazingly, not only did I end up with a more congenial range of climates this time around, but the two approaches turned out almost identical, even though I wasn't able to generate arid climates in my own algorithm. The problem: I couldn't convert the precipitation units using the eps_avg.py file, even after I had downloaded the latest version on GitHub. Here's the error message that resulted every time I ran eps_avg.py:

peter@localhost:~> python3 eps_avg.py
File "eps_avg.py", line 8
<!DOCTYPE html>
^
SyntaxError: invalid syntax

Without further ado, here is the revised Koppen map from the koppenpasta algorithm:

Click image for larger version.

Name: Koppen.png
Views: 59
Size: 81.7 KB
ID: 133779

And here is my own algorithm, which dovetails nicely with the one above, save for the lack of B climates:

Click image for larger version.

Name: Climate.png
Views: 28
Size: 77.2 KB
ID: 133780

Green: A (Tropical)
Yellow: C
Cyan: D
Magenta: E (Polar)

I'm wondering if someone could tell me: What is the difference between the MOST and the MOST.SNAP files (from the snapshots subfolder)? I assume choosing a later file (higher number) is better?

Yes, MrBragg, it looks like I made an error sizing the legend for my latitude lines; this resulted in everything poleward of 60 degrees being cut off from the map. Thank you for the note.

Now that I have the climates I've been desiring, I'm moving on to the next step: generating detailed topography.

Please let me know what you think.

Peter

**worldbuilding pasta** · 06-25-2022, 08:05 AM

The MOST files contain data across the whole year, averaged into months; i.e. the ts data is the average temperature of each month. The MOST.SNAP files are snapshots, showing the data at specific points of time, usually once every 15 days. So the snapshots aren't appropriate for making climate maps like this, but might be interesting if you want to see what's going on at a specific time of day. And yes, later is better; Earth-like climates usually balance out within 10-20 years, but I often run models for at least 50 to be sure (but I can afford to do that as I have a spare laptop that basically just runs models all day).

I can't replicate your error with eps_avg and that line definitely shouldn't be in the script, did you paste in some text?

**Peter Toth** · 06-27-2022, 03:17 AM

Hello Guild,

I've taken a temporary hiatus from the climate modelling to focus on some topography. I mentioned earlier that I was intending to use real DEM data to construct my geological features, but after cutting and pasting mountains from the real world, I felt somewhat "unauthentic." So I laboured for several hours on Wilbur and Photoshop to try to develop a brand new process, incorporating the many lessons I've learned from previous conworlds and maps. I was able to generate the following map for the first continent, most of which is covered in desert, save for an oceanic climate on its north coast, and a humid subtropical on its northeast coast. With this in mind, I've adjusted the rivers and tried to also mirror those climates in the topography. Hope you like it. Please let me know if this looks accurate. Thanks!

Peter

Click image for larger version.

Name: Reta Continent.png
Views: 49
Size: 2.19 MB
ID: 133788

Click image for larger version.

Name: Reta Climate.png
Views: 15
Size: 19.8 KB
ID: 133789

**Naima** · 06-27-2022, 04:30 AM

Originally Posted by Peter Toth

Hello Guild,

Thanks for the feedback, Naima.

Because I'm personally not satisfied with the results of my Wilbur approach, I've decided to use morne's method of juxtaposing actual DEM data to create a more realistic topographic map. It may take longer, but I'm aiming a bit higher this time around.

But first, I've run into some problems. Calculating my climates using worldbuilding pasta's script, I've discovered that much of my world will be covered in cold icecap climates and thus remain uninhabitable. Here are two maps comparing my initial parms to Earthlike parms. I'm definitely in favour of the Earthlike parms, but a 24 hour day is much too long to generate an adequate magnetic field for atmospheric protection.

Initial Parms:

Attachment 133714

Earth Parms:

Attachment 133715

Is there anything I can personally do to the parameters to make more surface area of my planet livable? Must I speed up the rotation or increase atmospheric density? I didn't want to increase the insolation as I already have average temperatures of 40 C in some regions. What other parameters could I change to ensure a more congenial spread of climates?

Thank you for any suggestions!

Peter

Personally I wouln't put too much trust and blind faith in algorithms for climate modeling , unless you are using more advanced ones with deep machine learning AI aid, I use only partially if none at all and I go mostly by common sense and study of the currents/winds/orography.

**Peter Toth** · 06-30-2022, 08:07 PM

Thanks Naima,

Yes, I'm personally never completely convinced about any complex calculation unless I can reproduce and cross check it with a separate algorithm. That's why I generally use both Clima-Sim and ExoPlaSim together (plus a couple spreadsheets), even though both have their limitations, most importantly a lack of ocean current modelling among others. Thus, to account for the influence of ocean currents, I'm intending to simulate the effect manually by painting a 10-15 C temperature change in the appropriate areas using worldbuildingpasta's blog as a guide. I figure that Azelor's tutorial is fantastic for Earthlike planets (as is Artifexian's), but as soon as the parameters stray far from Earthly standards, (such as the Polar cell disappearing with a 4-day rotation, for example), I find it difficult to trust my own calculations, or rather my ability to follow complex guidelines without error. Unlike a human brain, the GCM software always considers every programmed variable and follows every step in an algorithm, although sometimes the equations themselves are based on approximations and assumptions. Perhaps in 20 years, conworlding will be done completely by software, which will create every aspect of your world with scientific accuracy and high-precision resolution. I dread that day, for I believe such a feat will exise all the fun from conworlding!

Peter

**Peter Toth** · 07-01-2022, 12:27 AM

Update: I've tried to make the first continent even more realistic, using another experimental method (that I forget the exact algorithm for, unfortunately.) Here it is:

Click image for larger version.

Name: Reta Done 1.png
Views: 50
Size: 1.78 MB
ID: 133825

Is this any better?

**Turambar** · 07-01-2022, 01:10 AM

Originally Posted by Peter Toth

Update: I've tried to make the first continent even more realistic, using another experimental method (that I forget the exact algorithm for, unfortunately.) Here it is:

Is this any better?

Out of curiosity, was the algorithm used to generate the heightmap? or just to determine more realistic placement of topography?

**Peter Toth** · 07-01-2022, 01:26 AM

Hi Turambar,

The algorithm was supposed to take everything into consideration, including the heightmap and the realistic placement of topography. I kept switching from Photoshop to Wilbur over a dozen times, making layer masks for each new layer and juxtaposing them together at the end. Sometimes I experimented with juxtaposing two identical images but enlarging one and setting the blending mode to screen, for the purpose of eliminating that "Wilburified" look and to create some natural lakes.

The Wilbur process relies on several applications of Erode (Precipitation) followed by incise flow; this pattern is repeated several times with various entries for the incise flow variables. In the end, I've run through the entire image to adjust coasts and contour lines.

Finally, I've overlain a bitmap of the Wilbur image and set the blending mode to vivid light.

Does this image look any better than the last? Am I getting close to realistic topo?

Peter

Thread: Gigantica -- A Study in Maximum Habitable Planet Size

Thread Tools

Display

Posting Permissions