Fictional Exoplanet creation

Linguofreak

Well-known member
In the total energy balance yes, but not per wavelength. What you say is true for line emission, but that's not the mechanism we're dealing with here - we have line absorption but broad blackbody emission.
My understanding is that it is per wavelength: insofar as a body failsto be a perfect absorber it fails to be a perfect blackbody radiator, and at the same wavelengths. I was first made aware of this in a discussion on the vulnerability of spacecraft radiators in combat: you can't protect them from laser fire by making them reflective or transparent, because they'll fail to emit in the affected wavelengths to the exact same degree as they'll fail to absorb. I remember that the discussion specifically covered wavelength specific effects on blackbody radiation.

After doing a bit of research,

gives the equation

$\alpha_\lambda = \epsilon_\lambda$

which, given the lambda subscript, and if I read the article correctly, I believe bears out that absorbtivity and emissivity are equal per wavelength.

Thorsten

Active member
My understanding is that it is per wavelength

That is in explicit contradiction to your basic idea of a photosphere having a given temperature though.

Let's clear up the terms so that we see what law applies where.

The relevant issue is - how do we excite states in the molecules? And how do we de-excite them?

In the case of aurorae for instance, the solar wind coming in excites molecules. Since they're 100 km and more up, the atmosphere is very rarefied, and so they can stay in excited states for tens or even hundreds of seconds, and their only way to de-excite is line emission - they have to emit a photon to go down a specific energy level. Thus, Northern Lights do not have a blackbody spectrum, you see colors because the air molecules radiate at specific wavelength only, and the upper atmosphere doesn't get a temperature from the process, because it is highly off-equilibrium.

In the case of Greenhouse, the atmosphere is much denser. So rather than remaining undisturbed by any collisions, excited molecules bump into others all the time, and so the vast majority of them ends up doing thermal de-excitation - the energy ends up not being emitted as photons at specific wavelength but as kinetic energy of other molecules. And these, while bumping into each other, emit photons now and then - in a broad distribution of blakbody radiation (pretty much like the solar photosphere in fact). This process is in thermal equilibrium and hence we an assign a temperature to the atmosphere.

However, this isn't how we feed energy into the atmosphere, because the atmosphere isn't 'black' or 'opaque' throughout - so it doesn't receive energy like a black body would do. It receives energy by doing wavelength specific absorption only (in the limit that the whole atmosphere is opaque, you can reason using thermal equilibrium and just do a simple diffusion equation).

The situation is that the way we feed into the atmosphere is a non-equilibrium process, we have to do the line physics, but once the molecules get excited they thermalize the received energy quickly, and the process by which the energy is radiated away is in thermal equilibrium.

This difference between off-equilibrium and in-equilibrium process is not accounted for in your reasoning. Kirchhoff applies to a blackbody receiving and emitting thermal energy in equilibrium - which however isn't what happens, so the law doesn't apply.

Likewise, the interaction cross section for absorption is high at wavelengths where a state excitation is possible - and that is also where emission is high - but that applies to emission and absorption in off-equilibrium - which also isn't what is happening here, so that reasoning doesn't apply either.

Does that give you a better idea?

Thorsten

Active member
Up to this point, things were fairly calculable. Now the Fi-part of SciFi starts - I draw a collection of landmasses underneath the climate lines. The distribution should be Earth-like (more water than land) since the Albedo was similar. From there then follows a measure of geology, and ultimately the different ecological niches, but fist, as promised, let's talk about weather.

My arguments here will be more qualitative as I seriously do not have the resources to run a full weather code for a fictional world.

The basic pattern is that air underneath the subsolar point heats and rises. This sucks air from the ice-covered dark hemisphere inward while the warm air reaches the upper atmosphere and migrates outward.

The massive inflow of air at ground level leads to a spinning vortex due to angular momentum conservation (that's a common pattern observed in objects as diverse as bathtubs and black holes), aka we get a super-cyclone. If the air at the subsolar point is moist, the water condenses, releasing latent heat and leading to copious cloud formation and rain - which provides substantial energy to fuel the process - the mechanism is really the same as for tropical storms on Earth.

Now, as seen above, the whole process is self-quenching, because the cyclone clouds have a high albedo and cool the subsolar point - which means that the highest energy is suddenly next to the subsolar point where there are no clouds and Mime can warm the sea undisturbed.

So the cyclone will generally go to where it can suck most energy from the sea, and hence 'wobble' around the subsolar point. In addition, because the orbit is eccentric, it will weaken at apoapsis simply because the available energy is less - so a winter pattern will look more like this

Likewise, the process of cold air flowing in at ground level won't equalize with the high atmosphere outward flow easily, because the upper atmosphere has 1/5 of the density, so it's not easy to transport the same mass that is coming in. There will instead be boundary instabilities at the division line between warm and cold air, with fingers of cold air reaching inward while pockets of cold air being pushed outward - leading to the developments of frontal weather - cold fronts with massive thunderstorms and clear weather in their wake, warm fronts with plenty of snow on the ice shields - here's a coldfront coming in.

Now, the consequence of a massive coldfront is that there's a broad stretch of planet which has a few days of exceptionally fine weather - and if the coldfront reaches up to the subsolar point, there's a massive heating of the ocean and hence a juicy path filled with energy to fuel the cyclone - so it may react and use this energy - and thus detetch from the subsolar point and migrate outward - creating a far-reaching warmfront over the ice (and sucking cold air inward in other places.

Finally, just like a hurricane can spawn secondary storms at its fringes, it would be fairly normal that the super-cyclone generates secondary cyclones which migrate outward (they're not much restricted by Coriolis forces, as we've already established these are weak). Again these would move warm and cold air when they reach the boundary between ice and thawed land.

So that, in a nutshell, is what I believe the typical weather patterns on this world would be and why.

The consequence is a high variability of weather - when a coldfront moves in, temperatures can easily drop by 20-30 K within a few hours. Winds can reach hurricane-force quickly when a cyclone makes landfall. Precipitation is copious nearly everywhere, especially at the windward side of mountains. All of this shapes the terrain which will have deeply-cut water erosion created canyons. Every lifeform needs to cope with the potential of high windspeeds and sudden freezing.

I'm currently in a discussion with a fellow biologist tossing around ideas what plausible adaptions to these conditions could be for plants and animals, and (with the understanding that this gets more into the Fi-realm) I will eventually present some of that as well.

Linguofreak

Well-known member
Likewise, the process of cold air flowing in at ground level won't equalize with the high atmosphere outward flow easily, because the upper atmosphere has 1/5 of the density, so it's not easy to transport the same mass that is coming in. There will instead be boundary instabilities at the division line between warm and cold air, with fingers of cold air reaching inward while pockets of cold air being pushed outward - leading to the developments of frontal weather - cold fronts with massive thunderstorms and clear weather in their wake, warm fronts with plenty of snow on the ice shields - here's a coldfront coming in.

<image snipped>

Now, the consequence of a massive coldfront is that there's a broad stretch of planet which has a few days of exceptionally fine weather - and if the coldfront reaches up to the subsolar point, there's a massive heating of the ocean and hence a juicy path filled with energy to fuel the cyclone - so it may react and use this energy - and thus detetch from the subsolar point and migrate outward - creating a far-reaching warmfront over the ice (and sucking cold air inward in other places.

View attachment 26764

Finally, just like a hurricane can spawn secondary storms at its fringes, it would be fairly normal that the super-cyclone generates secondary cyclones which migrate outward (they're not much restricted by Coriolis forces, as we've already established these are weak). Again these would move warm and cold air when they reach the boundary between ice and thawed land.

Something I'd heard said would likely be a typcial weather pattern for planets like this, but didn't quite understand the reason for, is a cross-shaped pattern with cold air from the night side coming in from two opposite arms of the cross, and warm air coming from the subsolar point returning to the night side on the other two. I'd gotten the impression that it was supposed to be a coriolis effect of some sort, but when I tried drawing out what coriolis effects would do to the winds, it didn't look like it would produce that kind of pattern (nevermind your point about coriolis effects not being strong enough).

But looking at what you have there, I think I may have misunderstood the argument that was being made. It may have been more along the lines of those boundary instabilities pulling the cyclone off center, creating a situation like that illustrated in the image I left in the quote. With the spawning of secondary cyclones, you might then get a situation where two cyclones got drawn off in different directions by different cold fronts, and ended up creating two copies of the illustrated situation.

Thorsten

Active member
I don't know the source you're referring to, but I would assume that the general pattern must be a mixture of

• inflow to the subsolar point at ground level, outflow in the upper atmosphere and
• inflow and outflow by 'fingers' of air driven by the boundary instability

It is clear that heat exchange must take place in some form, but the precise balance would depend on factors like whether the subsolar point is over land and water, the general moisture level on the planet (massive storm systems are much easier to maintain over water with the high latent heat release when it rains off), the scale altitude of the atmosphere vs. the presence and location of a stratosphere, that in turn depends on the gas mixture as well as the stellar radiation spectrum,... So I would guess that under different circumstances, on a much drier planet perhaps, the permanent presence of a cyclone is not really obvious and in/outflow patterns could dominate the dynamics most of the time.

(I haven't tried to calculate this though, my weather reasoning is heavily based on Earth as atmosphere conditions on Stormhold are sufficiently similar)

Thorsten

Active member
Just a quick note that I have now published the code I've been using for this thread and am working on a tutorial series how it can be used. See the project page:

The code is GPLv2+, so enjoy.

Linguofreak

Well-known member
Any chance of a 3D solver? Eccentric orbits with orbital resonances offer interesting worldbuilding scenarios, and inclination can be a critical element in the stability of such arrangements (see Neptune/Pluto).

Also, it would be nice to have a some kind of calendaring tools, to answer questions like "if primitive astronomers take observations from this lat/lon on body A at sunrise/sunset, on what solar days will they observe the angle between bodies B and C to be at a maximum/minimum?". Most n-body solvers aren't built with calendaring in mind, and most calendar-building tools for worldbuilding don't include n-body solvers, so they're no use for planets librating around the L4/L5 point of a gas giant or brown dwarf .

Climate-wise, it would be nice to have provisions for determining the effect of eclipses on temperature (a moon in a close orbit around a gas giant will tend to experience daily eclipses of significant length on the giant-facing side, and if the inclination of its orbit to the ecliptic is significant, the length of the eclipses will vary with the seasons, up to the point of some or all latitudes not seeing eclipses at the height of summer/winter). Handling of climate influence from reflected light, and reradiated infrared (which can be significant for close-in moons) would also be nice.

Regarding the comment on the "setting up a star" page "it is not quote clear to which degree e.g. a white dwarf, which is the remnant of a stellar explosion, would even have something resembling a planetary system", I have two nitpicks; First, a white dwarf does not result from a stellar explosion (the mass loss in that case is gradual), so it can generally be expected to have whatever planets it did not swallow in its red giant phase (though orbits may be altered due to the mass loss). Secondly, for neutron stars, it was expected that any planets would be destroyed by the supernova in which the neutron star was born, but in fact the first confirmed extrasolar planets were found orbiting a pulsar (the arrangement may be fairly rare, but pulsar planets are much easier to detect than other planets as the regularity and short repetition time of pulsar pulses allows variations in the timing due to orbiting objects to be detected easily). That said, neither type of star is likely to offer a hospitable environment on any planets it does have.

Thorsten

Active member
Any chance of a 3D solver?

Yeah, it's a logical extension (and in fact I have one without GR corrections already). It'll probably come, I'll need to do some work on a user-friendly multi-body init first.

Also, it would be nice to have a some kind of calendaring tools, to answer questions like "if primitive astronomers take observations from this lat/lon on body A at sunrise/sunset, on what solar days will they observe the angle between bodies B and C to be at a maximum/minimum?".

That's a good idea, yes.
Climate-wise, it would be nice to have provisions for determining the effect of eclipses on temperature (a moon in a close orbit around a gas giant will tend to experience daily eclipses of significant length on the giant-facing side, and if the inclination of its orbit to the ecliptic is significant, the length of the eclipses will vary with the seasons, up to the point of some or all latitudes not seeing eclipses at the height of summer/winter). Handling of climate influence from reflected light, and reradiated infrared (which can be significant for close-in moons) would also be nice.

I started the code before I discovered gas giant moons as interesting targets for analysis, so it is kind of biased towards planets at the moment... But yeah, that too is a very good idea.

First, a white dwarf does not result from a stellar explosion (the mass loss in that case is gradual)
Well... semantics. It's the core of a star that has shed its outer layers, it definitely is more gradual and slow than a supernova, but... likely pretty violent anyway (I wouldn't want to be around).

So I kind of agree halfway - but only that
Secondly, for neutron stars, it was expected that any planets would be destroyed by the supernova in which the neutron star was born, but in fact the first confirmed extrasolar planets were found orbiting a pulsar (the arrangement may be fairly rare, but pulsar planets are much easier to detect than other planets as the regularity and short repetition time of pulsar pulses allows variations in the timing due to orbiting objects to be detected easily).
Interesting, I didn't know that. So... I'll add a comment then.

Linguofreak

Well-known member
Well... semantics. It's the core of a star that has shed its outer layers, it definitely is more gradual and slow than a supernova, but... likely pretty violent anyway (I wouldn't want to be around).

Well, there's "hostile environment for life" violent, and there's "disrupt a planet" violent. No source that I've ever seen has said that the mass loss on the (primary or asymptotic) red giant branch gets to the latter level of violence. All planets not physically engulfed by the star are taken for granted to remain present when the star reaches the white dwarf phase in every source I'm aware of.

Linguofreak

Well-known member
Oh, just thought of this: Is it possible at present to specify binary stars? An interesting test case would be a circumbinary planet at Delta Trianguli (binary separation of ~0.1AU, spectral types G0V and something between G9V and K4V, period of 10 days).

Thorsten

Active member
Well, there's "hostile environment for life" violent, and there's "disrupt a planet" violent.

I don't think we disagree in what happens - merely semantics.

Is a process where stuff flies outward from the center an explosion? Certainly the formation of a White Dwarf is a slow process - but then again, if you'd watch a real-time movie of a supernova, it wouldn't fit our intutition of an explosion as a 'sudden bang' either - even on the timescale of minutes, nothing much could be seen, it'd be actually quite boring in real time.

Is Earth 'destroyed' when the biosphere is gone? Or when atmosphere and oceans are gone? Or when the surface features are gone? Or only when most of Earth's mass is removed?

Obviously the code forsees that you might want to simulate a piece of slag orbiting a White Dwarf (or a Black Hole for that matter).

Oh, just thought of this: Is it possible at present to specify binary stars?

Yes, in fact I've been designing it with Brian Aldiss 'Helliconia' fresh in mind.

The issue is a bit making it user-friendly - you can easily simulate the short year (Helliconia's orbit around Batalix) with decent time resolution, but waiting for the longyear (Batalix' orbit around Freyr) takes a while.

Also, binary stars have the other configuration you mentioned - a planet orbiting the barycenter of two close-by stars - which needs specified differently in the UI.

So the code is there, the UI is not quite.

and inclination can be a critical element in the stability of such arrangements (see Neptune/Pluto).

I would be reluctant to see stability questions as applications because you really need to know what you're doing when running such an analysis. I have written the first version of the code because I wanted to understand why Jupiter both shields the inner planet against impacts and why it prevents the orbit of another planet in its vicinity - and the code gives the qualitatively correct answers, aka I could see how test planets in the critical region would statistically be gone after waiting some period - but is was not easy to tease this out, as you need to be sure the physical instability is driving what you see rather than the numerical one.

Linguofreak

Well-known member
I don't think we disagree in what happens - merely semantics.

Is a process where stuff flies outward from the center an explosion? Certainly the formation of a White Dwarf is a slow process - but then again, if you'd watch a real-time movie of a supernova, it wouldn't fit our intutition of an explosion as a 'sudden bang' either - even on the timescale of minutes, nothing much could be seen, it'd be actually quite boring in real time.

Not in the optical, anyways. If you're watching a neutrino scintillator the ~10s surrounding the core collapse would be impressive, and several hours later, the literature suggests that the X-ray shock breakout burst could ramp up on the order of seconds, and certainly less than an hour. The optical brightening would be slow and boring, though.

Is Earth 'destroyed' when the biosphere is gone? Or when atmosphere and oceans are gone? Or when the surface features are gone? Or only when most of Earth's mass is removed?

Well, for the purposes of an n-body simulation, it would be the removal of most of the mass, or the complete disruption of the body, which is why I'm not sure the "it is not quote clear to which degree e.g. a white dwarf, which is the remnant of a stellar explosion, would even have something resembling a planetary system" line really fits.
I would be reluctant to see stability questions as applications because you really need to know what you're doing when running such an analysis. I have written the first version of the code because I wanted to understand why Jupiter both shields the inner planet against impacts and why it prevents the orbit of another planet in its vicinity - and the code gives the qualitatively correct answers, aka I could see how test planets in the critical region would statistically be gone after waiting some period - but is was not easy to tease this out, as you need to be sure the physical instability is driving what you see rather than the numerical one.

I wasn't talking about investigating stability, I was saying that there are configurations whose known stability characteristics are negatively impacted if you squish them down to a plane, thus the desire for a 3D solver.

Thorsten

Active member
which is why I'm not sure the "it is not quote clear to which degree e.g. a white dwarf, which is the remnant of a stellar explosion, would even have something resembling a planetary system" line really fits.

Re-reading this, you're actually right - it is misleading. I'll change it, thanks for being persistent

Linguofreak

Well-known member
Here's something that might be good to include eventually: A clean-sheet implementation of the "Accrete" algorithm.

Now, if you're into writing code for sci-fi worldbuilding enough to be putting together a project like this, you might ask "sources for the accrete algorithm are already available on the net. Why reinvent the wheel?"

The answer is in this post: In short, the most popular implementations floating around out there have murky copyright status (the common ancestor is technically all-rights-reserved, but that was probably not intended by the copyright holder), and one that I thought was promising to use as a base for my own work turned out to flat-out have portions under incompatible licenses.

Fortunately, the 1969 paper describing the algorithm at the root of it all is available online:

So the possibility exists to create a clean-sheet implementation that will then have a known-good copyright status.

Of course, the original algorithm only generates Sol-type systems, when we now know empirically that a more diverse set of possibilities shows up in nature, but there are accrete-based system generators out there that manage to create fairly systems that are fairly believable in light of modern knowledge, so we at least know that the algorithm can be modified in such a way, even though we want to avoid using any existing implementation in order to create a clean code base without licensing issues.

Thorsten

Active member
Now, if you're into writing code for sci-fi worldbuilding enough to be putting together a project like this

Actually, my main motivation is the hands-on experience of astrophysics - I just happen to like doing science

So yeah - one could include all sorts of things, but then again, I've made it available GPL, so there could also be all sorts of contributors other than myself.

***

Here's an appetizer for what I'll upload 'soon' - orbit of a planet around a binary pair orbiting each other, sadly way too close - it's pretty cool to see it wobbling around before being ejected from the system.

Thorsten

Active member
Okay, I managed to get the update out - this can now also do planets orbiting binary stars. Here's an example from 'Helliconia', a novel series by Brian Aldiss.

Helliconia orbits sun-like Batalix which in turn orbits Freyr on an elliptical orbit with an eccentricity of 0.5 - at periapsis Freyr keeps the planet tropically hot, at apoapsis only Batalix keeps it from freezing totally.

Here's a simulated temperature distribution (without atmosphere transport) from summer with Batalix and Freyr in opposition - one nicely sees to 'heating peaks' moving across the planet in longitude while the axis tilt towards Freyr keeps the northern hemisphere hotter.

For code and tutorials see my page.

Linguofreak

Well-known member
Huh. A couple questions I thought I'd asked, but apparently haven't, looking back at the thread:

Is it currently possible to specify mean anomaly for bodies?

Is it currently possible to specify moons?

Is it currently possible to plot the time evolution of a parameter (such as temperature at a given location)? Most of the examples seem to involve evolving the simulation to a given time and then plotting, for example, the instantaneous value of temperature at a range of latitudes and longitudes. It would be useful go be able to run the simulation a while to allow things to equilibriate, then plot the time evolution of temperature at a given lat/lon over a day, or a year, or a multi-year seasonal cycle.

Thorsten

Active member
Is it currently possible to specify mean anomaly for bodies?

Not upon init, but you can run the orbital solver to get to your point of interest and then start the thermal solver on top.

Is it currently possible to specify moons?

No. I know they're interesting for habitability by e.g. stabilizing the rotation axis, but that's a long-term stability thing and outside of the scope of the program - so a moon wouldn't do much relevant stuff otherwise.

The solver can do it of course, but there's no interface to expose that.

Is it currently possible to plot the time evolution of a parameter (such as temperature at a given location)?
Again, it's not exposed explicitly. I also think for the more interesting examples with atmosphere it is misleading, because the simulation does climate, the day-to-day change of temperature however is weather (which can be very different from mean values).

For the less interesting examples which are symmetric in longitude, you can use longitude as time axis and read off daily slices from the surface plots.

So in a sense it's there, but the global average picture is more meaningful, a time series for a given location often suggests an accuracy which imo isn't there.

Linguofreak

Well-known member
Not upon init, but you can run the orbital solver to get to your point of interest and then start the thermal solver on top.

I was thinking for a Trojan arrangement, where the tertiary body needs to lead/trail the secondary by approximately 60 degrees.

No. I know they're interesting for habitability by e.g. stabilizing the rotation axis, but that's a long-term stability thing and outside of the scope of the program - so a moon wouldn't do much relevant stuff otherwise.

The solver can do it of course, but there's no interface to expose that.

I'm more thinking in terms of a gas giant moon.

Thorsten

Active member
I was thinking for a Trojan arrangement, where the tertiary body needs to lead/trail the secondary by approximately 60 degrees.

Those would be most conveniently specified as 'at L4 (or L5)' and placed using a Lagrange point finder (surprise, I have one of those as well...)

I've never heard that you can have planets developing there though... it seems kinda hard to accumulate enough mass, usually it's just debris that's been captured

I'm more thinking in terms of a gas giant moon.

(In fact, you can cheat a bit and specify a 'star' with the temperature and luminosity of a gas giant, let that orbit your companion star and you get the results with the existing code and interface)

I guess for me the point is - of course there's all sorts of weird situations where orbital dynamics is interesting. But most of these cases would be rare and exotic. Whereas what is a major influence and not rare at all is... atmosphere. So I'm actually more interested in spending my time with improving the atmosphere simulation and see whether one can tease some gross weather patterns out for instance (like the (non)-existence of a jet stream) than coding an interface for arbitrarily weird situations (also with extending the solver to 3d - the issue isn't so much the solver, 3d is easy there, the issue is interface and analysis tools, they become much more complicated and messy without providing much more insight - most systems would be near-planar problems..)

Replies
6
Views
614
Replies
4
Views
272
Replies
3
Views
287
Replies
1
Views
166
Replies
19
Views
2K