Last edited by Ottehcnor; 05-12-2022 at 10:13 AM. Reason: grammar correction
Agree with Morne, this is definitely some of the best Wilbur output I've ever seen! It's *almost* enough to get me to make the switch (but not quite, mostly because I'm on a Mac). Would love to see a tutorial for this process at some point.
Well, this has advanced a lot! Congratz on finishing the first continent (which looks amazing, btw).
Very solid work overall, and looking forward to seeing more!
Because topography can get boring, and because climate and topography can feed back on each other, I’ve taken a bit of a detour into temperature modeling lately. Months ago I followed the Azelor / Charerg / AzureWings tutorial here and I wanted to compare the results of that to the Clima-sim method covered by Nikolai Lofving Hersfeldt which generates temperature isotherms on the basis of a very low resolution input geography.
For all of these, no attempt was made to correct for orographic temperature changes and no modification was made to the Clima-sim output to account for the influence of currents (which it ostensibly already includes). Apart from the input geography, all other Clima-sim parameters were kept at the default Earth values.
July Tutorial (left) and Clima-sim (right)
map6_july_temp_small.png climasim_july.png
The July maps have pretty remarkable agreement between the two methods and there really are only (to my eye, anyway) very minor differences. This is really neat and makes me happier than it probably should.
January Tutorial (left) and Clima-sim (right)
map6_jan_temp_small.png climasim_jan.png
For the January maps, the qualitative agreement is still pretty good, though the differences are much more pronounced. The Clima-sim maps are a bit warmer at the equator, which likely doesn’t impact all that much since it’s always going to be hot, and are also appreciably warmer at the poles. Particularly in the northern hemisphere, the 0 C isotherm is quite far north, which pushes the “C” climates far to the north and eliminates the Dfc climates in the northernmost latitudes entirely.
I’m not 100% sure what to make of this January discrepancy. Perhaps the lack of land at very high latitudes allows for much better heat transport by the oceans and so moderates the high latitude temperatures a lot. There could also just be a systematic temperature anomaly in winter that makes all of the Clima-sim output warmer than it should be. I'm going to keep messing around with this, but wanted to post these results here in case they're interesting for anyone.
Thanks for posting this! I have always wondered about the practical differences between the Azelor tutorial and the Nikolai's much-more-involved version. The latter honestly struck me as a great deal of additional work and processing power for what seems like a less-exact result. I was also skeptical of his results that seemed too warm at the poles for my liking (the final result for Teacup Ae had no south polar tundra despite having a large continent at something like 85º south). My initial guess is that the Clima-sim methods understate the cooling effects of having a great deal of land directly next to the poles, which is likely to result in extensive ice-caps and thus cooler temperatures even in summer than what an agnostic model predicts. Clima-sim even allows for inputting ice caps into the land cover map, but it seems this doesn't matter much.
Of course, the Clima-sim algorithm is much more advanced than the Cartographer's guild model that much more strictly uses Earth as a base, and I'm not a climatologist and haven't fiddled around with Clima-sim at all. Perhaps there's something obvious I'm missing here, but to my eyes the results produced manually using the Azelor/Charerg/AzureWings method seem more plausible to me.
Yep, that's the exact same boat I'm in here. The results from the Azelor/Charerg/AzureWings tutorial feel more right, but I also know nothing about climate forcings and so am entirely unqualified to be making these judgements. That the July results match up so well while January is so different is confusing, so I'm playing around with changing parameters somewhat systematically to see if I can't figure out why this is. If I stumble onto anything interesting I'll definitely let you know
The warm south pole in that model is due to Teacup Ae's eccentricity, which causes stronger seasons in the southern hemisphere; the significantly warmer summers (and generally flat terrain of the southern continent) evidently prevent permanent glaciers from forming, and without that ice-albedo feedback the south pole stays remarkably temperate. Later runs with a formal climate model have largely confirmed this result (though that also helped highlight some of the shortfalls to my approach to determining precipitation).
There may be something similar going on here: If there isn't sufficient landmass at the south pole to maintain a large glacier, the overall albedo of the poles will be far lower and so they will be warmer compared to Earth's (you can play around with the ice cover in clima-sim to see if you can get different results, though; there's a certain inertia to global climates wherein the presence of a reflective ice cap can help keep itself cool in a situation where an initially ice-free pole might not form an ice cap).
First of all I wanted to say thanks for putting together your tutorials; they're super informative and I always enjoy reading your material
This is what I'm currently playing around with. If I give the north pole partial ice coverage it dramatically lowers the northernmost temperatures and pushes the 0 C isotherm somewhat farther south than without the ice. I don't really have an intuition for 1) whether it's appropriate to add this type of sea ice cover since there is no land up there (as you say, feedback is potentially quite important), and 2) whether the ice-free version accurately reflects the temperature distribution or whether that is "too mixed" and thus too warm in the winter. As a baseline initial test, I'd be curious to see what ExoPlaSim predicts for a water world so I could compare that to a Cilma-sim water world output, I just need to find the time to run that test...
1, Clima-Sim should determine sea ice cover on its own.
2, Not too sure what you mean by "too mixed". The general approach I'd advise if you're aiming to have ice caps is to begin by over-estimating the extent of ice, running the model, and then remove ice from areas that get over 0 C in summer (be sure you're looking at temperatures at the surface, not sea level), and iterate until you find an equilibrium.
Regarding the test you ran with Antarctica, there should indeed be a significant difference there--and you'd be surprised how warm polar regions can get without ice thanks to constant sunlight in summer. This might be overstating it a bit though; Clima-sim doesn't seem to do a good job of allowing heat to flow from the poles to lower latitudes when the former gets warmer than the latter, which is why it essentially breaks at higher axial tilts.
re: Peter, I took a look at morne's world and that high-latitude desert is a reasonable outcome of a large mountain range casting a large rainshadow. ExoPlaSim has its weaknesses (most notably the low resolution, static glaciers, and lack of deep ocean currents) but it's in a whole other league compared to Clima-sim in terms of the complexity and accuracy of the underlying simulation and its flexibility in handling a range of climate parameters.
That it was warmer made sense, but the magnitude seemed off. As a followup here I removed the ice from Greenland and saw qualitatively similar but somewhat less crazy warming of Greenland / Siberia / northern Canada. My guess is that's due to the moderating effect of water near the pole, though removing half the land from an ice-free Antarctica still gives temperatures nearing 30 C in the summer.
Ice-free Greenland
lastmap.png
As a point of comparison, I'd actually be really curious to know what ExoPlaSim predicts for the temperature of your southern continent in the summer to see how that compares to the >40 C predicted by Clima-sim. Even with its systematic issues, I'd be much more inclined to trust the output from ExoPlaSim given how much more explicitly it treats things.
Last edited by MrBragg; 05-08-2022 at 10:23 PM.