An Apple Pie From Scratch, Part VIa: Climate: Global Forcings

NASA/JPL-Caltech/R. Hurt, T. Pyle (IPAC)

Climate is, simply put, not easy to predict. It is a chaotic system influenced by many positive and negative feedbacks, and given a single set of boundary conditions (astronomical context, tectonic activity, etc.) there are often multiple possible stable climates. Within the last half-billion years, Earth’s average temperature has varied from about 5 °C cooler to 15 ° C hotter than today (the current average is about 15 °C), and as such the poles have been variously covered in continent-sized ice caps, temperate woodlands, and seasonal rainforests.

All this happened without any major changes in solar insolation, orbital properties, tectonic activity, or atmospheric content. Minor shifts in the Earth’s orbital and rotational motion, topography, and greenhouse gas production have caused drastic changes in surface conditions. So if we want to predict the climate of a planet similar but not identical to Earth, then we need to get into specifics—though we also need to accept a little uncertainty as well.

But it’s worth saying here that getting into specifics necessarily means we lose broad applicability. This has been a trend with this series from the start, but just to be clear I’ll be going into this post assuming that we’re working with planets that resemble Earth in all the following ways (save for minor variations in specific sections): Broadly similar size; Nitrogen-dominated atmosphere of similar total pressure; CO₂ as the controlling greenhouse gas, itself controlled by the carbon-silicate cycle; Presence within the habitable zone; And mix of above-water landmasses and water seas.

We’ll also be assuming for now that these planets have days far shorter than their years, so that heating is similar all around the equator; tidal-locked planets and other synchronous rotators deserve their own analysis, which I’ll do in a later post.

• Mechanisms of Climate
• Forcings and Climate States
• Astronomical Forcings
• Star Type
• Insolation
• Orbital Period (Year Length)
• Rotational Period (Day Length)
• Obliquity
• Eccentricity
• Planet Size
• Geological Forcings
• Volcanic Outgassing/Weathering
• Land/Water Ratio
• Continental Drift
• Ocean Salinity
• Atmospheric Forcings
• Pressure
• Oxygen
• Biological Forcings 
• Plant-Like Vegetation
• Fossil Fuel Use
• Sea Level
• Tectonics
• Climate State
• Lakes and Groundwater
• Milankovitch Cycles
• Thermal Expansion
• Mantle Cooling
• An Evolving Planet
• In Summary
• Notes

Back to Part Vb

Mechanisms of Climate

Monthly averaged cloud covered across one year (2014) NASA

Earth’s climate, just like its tectonics, is driven by heat transport between the warm surface below and the cold of outer space above. Really it’s beginning to seem like everything interesting that happens here (perhaps including life itself) is just an elaborate heat rejection mechanism for the planet.

But in this case the source of the heat is not the Earth’s interior, but the sun: Sunlight passes through the atmosphere, mostly without interacting with it, and then is absorbed by the surface; so while exposed to sunlight the Earth’s surface is usually warmer than the air above it. The air itself is warmer than outer space, so there’s a gradient driving heat transport upwards from the surface. But because not all areas of earth are equally warmed by sunlight, heat transport occurs not just between the surface and space but between different areas of the surface as well.

To understand what effect this has, let’s consider a few idealized scenarios. First, imagine that the Earth is a perfect sphere with a smooth, homogenous surface, no obliquity, and no rotation—and yet still somehow receives equal sunlight all along its equator (in essence, imagine a static all-ocean planet that is surrounded by some kind of bright ring that replicates sunlight).

In this scenario, high insolation (power per area of incoming sunlight) at the equator causes it to be warmer than anywhere else, and it heats the air directly above it; this warm air expands and rises, until it reaches an altitude where the atmospheric pressure is low enough that it is no longer buoyant—around 10 to 15 kilometers for Earth. There’s still more warm air pushing up from below, so the air moves laterally away from the equator and towards the poles. As it travels poleward the air radiates heat into space, and by the time it reaches the pole it’s cooled down enough to sink back to the surface.

Now we’ve got a lot of air moving away from the equator—creating a low-pressure zone at low altitude—and a lot of air converging at the poles—creating high-pressure zones. The cold air moves down the pressure gradient back towards the equator, and absorbs heat from the surface as it does such that it’s ready to rise again at the equator.

So the result is one large convection cell in each hemisphere, with warm high-altitude winds carrying heat away from the equator and cool low-altitude winds coming back from the poles. A neat, simple closed loop.

Lyndon State College Atmospheric Sciences

Now let’s modify this scenario and rotate the planet at the same speed as Earth, once every 24 (ish) hours; and in the same direction, to the east.

Activity at the equator still looks much the same: warm air rises from the surface and then is pushed polewards. But as warm air moves away from the equator, it is still moving east at the same velocity of the equator (around 1,700 km/h) while the velocity of the ground below it is decreasing. From the ground’s perspective, this makes the air current appear to curve towards the east as it moves poleward, what we call the Coriolis effect. The effect is strong enough that by the time the warm air reaches 30° latitude, it is moving almost due east.

Unable to move any further poleward, the air radiates some of its heat, sinks, and then flows back to the equator. By now it has lost some of its eastward velocity and matches that of the surface, so it starts to lag behind the increasing velocity of the surface as it moves equatorward, and thus it curves towards the west. So: on an eastward-rotating planet the Coriolis effect causes:

• Poleward air currents to curve east, and 
• Equatorward air currents to curve west.

Apparent deflection of poleward (red) and equatorward (green) winds due to varying surface velocity (shown on right). Paul Webb

For westward-rotating planets it would be the reverse, though I tend to think we should just define east as the direction a planet rotates regardless of whether it’s prograde or retrograde relative to its orbit.

Meanwhile, this leaves a lot of extra heat and a high pressure zone at 30°, and still-cold poles beyond. The warm air there cannot rise against the flow of the sinking air flow from the equator, so instead it moves poleward at low altitude. Once again the Coriolis effect causes it to curve east, such that it can only make it to 60° latitude. By this point it is sufficiently warmer than the surrounding air that it can rise into the upper atmosphere, but as it does so it loses some of its heat to the cooler surrounding air. The now-cooled air flows back equatorward to 30°, and sinks to complete the loop.

And then finally, the warmed air at 60° flows at high altitude to the poles, then sinks and flows back to 60° at low altitude.

So altogether, there are now 3 convection cells in each hemisphere. From equator to pole, we call these the Hadley cell, the Ferrel cell (a.k.a. mid-latitude cell), and the polar cell. Heat still moves from equator to pole, but it does it by jumping across convection cells; first at high altitude to 30°, then low altitude to 60°, then high altitude again to the pole. And in each cell, cool air flows equatorward at the opposite altitude to close the loops of air flow.

But when we talk about climate, what we’re usually concerned with is the conditions on the surface, so what we really care about are those low-altitude winds—the prevailing winds—and the pressure zones associated with them.

Pressure zones and prevailing winds caused by atmospheric circulation. Stevenson 2019

So between 30° and the equator, the Hadley cell creates the easterly trade winds (winds are named for the direction they blow from, not the direction they blow towards, a convention I always found confusing). They blow more directly west closer to the equator, but still always tilted towards it. The Ferrel cell creates westerlies (tilted towards the poles) at mid latitudes, and the polar cell creates (equator-tilted) easterlies at high latitudes.

One thing to remember is that as these winds travel, they carry moisture along with them—so in general high-pressure zones where winds originate are dry and low-pressure zones where winds converge are wet. The wettest area of the planet is the Intertropical convergence zone (ITCZ) at the equator, with fairly wet polar fronts at 60° as well. The driest areas are the poles and the horse latitudes at 30° .

Burschik, Wikimedia

Now let’s modify the scenario again, and add continents: Earth’s will do just fine for the moment.

Different areas of land and sea have different albedos, meaning that they reflect or absorb different amounts of sunlight—higher albedo means more light reflected and less heating for a given level of insolation. To review, here are typical albedos for different surfaces on Earth:

Surface	Open Ocean	Dense Forest	Grassland/ Shrub	Desert	Ice/Snow
Albedo	0.06	0.08-0.15	0.1-0.25	0.3-0.4	0.5-0.8

This is only true for modern Earth, of course; different stars, different minerals, and different lifeforms can all affect these values. But it’s a good baseline to work from.

Clouds also have a high albedo, similar to ice, and wetter areas will have more clouds, raising their albedo. On earth today this only has a small effect on local climates, but on a global scale it raises the total albedo enough to significantly cool the planet.

Earth average surface albedo without (above) and with (below) clouds. Giorgiogp2, Wikimedia

The effect of all this is that with continents present, surface temperature is no longer solely determined by latitude; two areas at the same latitude can receive different amounts of heating due to different albedos. You might expect that the oceans, with the lowest albedo, would have the highest temperatures then, but they have their own internal convection as well (and also tend to have high cloud cover), so often the warmest areas are on land.

All of this alters the exact geometry of the atmospheric convection cells. The ITCZ, for example, no longer lies exactly along the equator but swoops north and south through the hottest regions of the globe, and the polar fronts are similarly affected.

And regarding local climates, the fact that the prevailing winds do not flow directly north and south has important impacts on water supply: Land areas downwind of oceans will be generally wetter, even at the horse latitudes, while the opposite is true of areas upwind of oceans. Topography also has an impact: higher terrain is generally cooler, and high mountains can block the transport of moisture through the process of orographic lift (a subject for the next post).

Finally, we can introduce obliquity to our model; Earth’s axial tilt of 23.5°, which causes the seasons—where maximum insolation shifts north and south, and so temperature and the atmospheric convection cells do as well.

Water has a high specific heat, meaning it takes a relatively large amount of heat to increase the temperature of a given mass of water, and this combined with global ocean circulation gives the oceans a high thermal inertia—they resist temperature changes better than land at the same latitude (wet land areas also have a higher thermal inertia than dry areas, to a lesser extent). This means that within each hemisphere, the land will heat up faster than the oceans in summer and cool down faster in winter. Thus, the temperature contrasts between different land and ocean areas become even more pronounced, and the ITCZ sweeps large distances north and south with the seasons.

Mats Halldin, Wikimedia

Where the ITCZ moves farthest, it causes monsoon patterns where regions of land are wet in one season and dry in another. This can create rainforests in the horse latitudes and deserts at the equator. The polar fronts move across the seasons as well, which affects the severity of winters in temperate areas. 

Forcings and Climate States

Concept of the surface of TRAPPIST-1f. NASA/JPL-Caltech/T. Pyle (IPAC)

We’ll discuss all these fine controls of local climate—ocean currents, topography, monsoons—in more detail in the next post. What I want you to bear in mind for now is this overall model for how the interactions of insolation, atmospheric convection, the Coriolis effect, surface albedo, and obliquity create the climate we have. This is the baseline we will be working off—that of our modern Earth.

What I want to do for the rest of the post is look at global climate forcings: conditions that can cause dramatic increases or decreases in average temperature or precipitation, or dramatic shifts in the distribution of temperature and climate zones across the planet. A stable climate is an equilibrium between the many different forcings acting on a planet. When a forcing becomes stronger or weaker (say, for example, a sudden increase in CO₂ supply to the atmosphere) then the climate begins to change until it achieves a new equilibrium.

For example, one strong forcing on climate is planetary albedo; higher albedo lowers the average temperature. Ice has a much higher albedo (0.5-0.9) than liquid ocean water (0.06) and most other types of land cover (0.08-0.4) so increasing ice coverage lowers the planet’s average temperature. Of course, lower temperature increases ice coverage, which lowers temperature, and so on. The same can happen in reverse; high temperature causes ice to melt, which raises temperature, etc.

This positive feedback effect, called ice-albedo feedback, can amplify small initial temperature shifts into a major climate transition. I’ve mentioned before how it can render seemingly hospitable planets uninhabitable, but it also operates on finer scales as well. In these cases it is eventually stopped by stronger negative feedbacks (typically the carbon-silicate cycle; also discussed in the linked post above) or it reaches limits of effectivity: If the ice caps melt completely, or if they grow to the point of extending into deserts that already had a high albedo, then further temperature changes cause a smaller change in albedo.

There are many such feedbacks between climate forcings. Because of this, the Earth rarely transitions gradually between different climates. The trend in the past has been for the planet to leap between discrete climate states, which it inhabits for long periods before a sudden change in conditions causes it to leap again. In some cases the same set of external forcings—astronomical conditions, initial planet size and atmospheric content—can allow for multiple stable climate states, with the particular history of the planet determining which one is inhabited at the present.

Scotese 2016

Throughout its past Earth has fluctuated between three main climate states (exact terminology varies):

Hothouse states (A.K.A. greenhouse states), with warm seas, no polar icecaps, and few or no glaciers on land. This has been the default for the last half-billion years, perhaps for Earth’s entire history. The poles likely still have seasonal snow, but in the warmest periods they may come to resemble tropical rainforests in summer. High sea levels also mean broad shallow seas around the continents, ideal for marine life. But as a tradeoff, arid regions are probably more hostile, and life has to adapt to high humidity in the tropics (some sources distinguish between regular greenhouse and especially hot super-greenhouse or hothouse states, but I won’t bother here).

Icehouse states (A.K.A. ice ages), with cool seas, ice caps, and extensive glaciers. We currently inhabit an icehouse state which began 2.5 million years ago. This period has itself been divided into two climate substates: Interglacials like today, with fairly small ice caps, and glacials, with large glaciers extending into midlatitudes and sea levels falling by tens to hundreds of meters. Throughout the icehouse period Earth has fluctuated between glacials and interglacials on a roughly 100,000 year cycle, with glacials usually lasting longer, driven in part by variations in Earth’s orbital and rotational motion. Previous icehouses have probably had similar cycles.

Snowball states, with total or near-total coverage of the surface in glaciers and ice sheets. Exactly how total remains a matter of debate; It may be possible for a thin strip of unfrozen ocean to remain near the equator, what some call a slushball state. In either case, the poles are essentially uninhabitable except by deep-sea life, and these states are generally a challenge to the survival of complex life. The last snowball ended 650 million years ago; so in Earth’s history they’re fairly rare events, but this may not be true of all planets.

A fourth state, the moist greenhouse state, will occur in the future when Earth passes the inner boundary of the habitable zone. In this state large amounts of water evaporate into the upper atmosphere, warming the planet above habitable temperatures and gradually escaping into space. When the oceans are lost completely, CO₂ builds up in the atmosphere and warms the planet further, likely sterilizing it completely. Fortunately, such states appear to be very difficult to attain for planets in the habitable zone.

 

We can regard hothouses and snowballs as end states of the range of habitable climates, with icehouses as an intermediate state. Earth appears to be biased towards the hot end of the range, but this may not be true for all habitable planets. Stronger positive feedbacks like ice-albedo feedback make it harder for planets to occupy an intermediate state; a planet with stronger feedback due to, for example, higher obliquity, will cause a planet to jump straight between hothouse and snowball states.

Knowing that the strength of one climate forcing can affect another, accounting for all of them is a very difficult task. Trying to isolate the effect of one forcing is perhaps just as difficult. Nevertheless, let’s go through some of the most impactful forcings affecting Earth and similar exoplanets, and try to get a sense of what effect they each have.

Astronomical Forcings

These are forcings originating from the star system around a planet, and the planets motion through it. These are the strongest forcings, and I’ve discussed many in the past, but only to the point of what does and does not allow for habitability. Now I want to dig in a little deeper to the effects of subtle variations in these forcings within the constraints of habitability.

Three of the forcings we’ll discuss are all closely linked together: Star type, insolation, and orbital period. For a given star type, insolation and orbital period are both determined by semimajor axis, so one cannot be changed without affecting the other; and for different star types, keeping the same insolation requires altering the period, and vice-versa. But we’ll try to consider them independently as much as we can, and focus on the strongest effect of each individual forcing.

Star Type

To review, we went over the various types of potentially habitable stars back in Part II, and determined that within the main sequence, more massive stars are brighter, bluer, and evolve faster, and less massive stars are dimmer, redder, and longer-lived. Though I’m mainly concerned with main-sequence stars, the effect on climate mostly comes down to the spectrum of light produced by the star; so e.g. the climate of a planet orbiting a red giant should be similar to that orbiting a red dwarf and receiving equal insolation.

The most important effect of a star’s spectrum, the particular wavelengths of light it produces, is that a planet’s surfaces will have different albedos under different spectra. In particular, ice absorbs more infrared light than visible light. Were Earth transported into orbit of a red dwarf star with the same insolation and no other changes, it would absorb so much more light and be so much warmer that it may no longer be habitable—for stars with effective temperatures less than about 5,000 K, such a scenario would put Earth inside the inner boundary of the classical habitable zone, which should push it into a runaway greenhouse state (though some models suggest that alterations in cloud behavior may still result in habitable, if sweltering, conditions).

In general, Earthlike planets should have planetary albedos 0.1-0.2 lower around M-type stars than G-type stars. To keep the same climate, an unaltered Earth would have to receive 12% less insolation compared to what it receives now when orbiting a typical M-type star, and 8% more insolation when orbiting an F-type star.

But for planets that are in the HZ of red dwarfs, ice-albedo feedback would be weaker, and so climate transitions can be expected to be smaller and more gradual. Pole-equator temperature and humidity differences will be smaller as well, and ice caps smaller for a given average temperature—though weaker ice-albedo feedback increases the likelihood of partial ice cover in general.

Shields et al. 2019

But it’s not all good news: Because it absorbs less light, water would also be less prone to evaporation and so such a planet would have around 10% less global precipitation at a similar average temperature to Earth. Also, once insolation drops below 65% relative to Earth, hydrohalite ice may begin to form that reflects more strongly in infrared and so causes stronger ice-albedo feedback again (though, as we’ll see in a moment, it’s generally weaker at lower insolation anyway).

Albedo also affects the impact of photochemical hazes, like the tholin haze of Titan or possible hydrocarbon haze of early Earth. F-type stars produce high UV radiation that will act to decompose the haze, while M-type stars produce mostly infrared light that will pass through the haze; so it is only around G- and K-type stars that these hazes can significantly cool a planet.

Finally, the other big factor to consider in relation to star types—in particular when reading the next section on insolation—is that of star lifetime and evolution: If a planet is orbiting an F-type star that evolves so quickly that the planet can only remain in the habitable zone for 5 billion years or less, then by the time complex life has developed on this planet (presuming it takes about as long as on Earth and the planet had to remain in the HZ the entire time) it must be close to the inner edge of the HZ. The same would apply for red giants as well, which evolve on similar timescales.

However, such a planet could have developed with an early methane or hydrogen greenhouse, as discussed before, so this is not necessarily always the case.

Insolation

Light from the sun is, of course, the largest forcing acting on Earth’s climate, and it’s fairly obvious that if we increased insolation and changed nothing else, Earth would warm up.

But we know that, within the limits of the habitable zone, the carbon-silicate cycle will compensate for increases in warming by insolation by lowering CO₂ levels and so reducing greenhouse warming. But this doesn’t mean there’s no effect; At lower CO₂ levels the chemical processes that sequester CO₂ are less efficient, so a higher temperature is required to make sequestration match outgassing.

Thus, an Earth with the carbon-silicate cycle at equilibrium (and enough CO₂ outgassing to avoid snowballing or limit-cycling, which may or may not be at Earth-equivalent levels depending on whom you ask) at the inner edge of the conservative habitable zone (with about 112% Earth’s current insolation) would be around 10 °C hotter, and one at the outer edge (36% current insolation) around 15 °C cooler, close to freezing. Though the habitable insolation range is different for other stars, the range of resulting temperatures should be similar.

Predicted average temperature for given insolation with CO₂ outgassing equivalent to Earth (light grey) or 5 times greater (black); this model suggests the former is too low for habitability in the outer HZ. Kadoya and Tajika 2019

However, exactly where the inner boundary is and what sort of climates should appear there is still a matter of debate. Conventional habitable zone models predict that more than 10% increased insolation would push the planet into an uninhabitable moist greenhouse, but more detailed models predict that shifts in atmospheric circulation and cloud formation could keep the climate stable at up to 21% increased insolation.

But by this point average temperatures would exceed 70 °C, and CO₂ levels would fall so low that no modern plant life could survive. The very humid atmosphere would distribute heat very evenly across the Earth, so not even the poles would remain cool. Any remaining life would either have to dramatically change its biochemistry or retreat to the deep oceans—though even those would be lost to space in under a billion years.

Going the other way, as we move further out in the habitable zone warming from initial absorption of sunlight decreases and greenhouse warming increases, so surface albedo and relative insolation become less important; ice-albedo feedback weakens and equator-pole temperature difference decreases. Once the atmosphere has over 3 bars of CO₂, the feedback is essentially gone.

Kadoya and Tajika 2014

In spite of this, snowballing or limit-cycling become more likely in the outer habitable zone, because the minimum threshold of volcanic CO₂ outgassing required for the carbon-silicate cycle to operate increases. Below 62% of Earth’s current insolation, recovery from snowball also becomes more difficult due to formation of CO₂ ice at the poles. So to keep hospitable conditions, these planets would have to either avoid snowball states entirely or compensate with very high volcanic outgassing rates.

Orbital Period (Year Length)

Year length is intimately tied to star type and insolation, both stronger forcings, so there’s been little analysis of the effects of year length alone. But one obvious effect is that longer years means longer seasons, and so more time for increased insolation near the poles to overcome thermal inertia. On Earth the equator stays significantly warmer than the poles throughout the year, but with years 4 times as long the poles may become warmer in summer and the equator undergoes solstice winters and equinox summers just as on a high-obliquity planet.

Each chart shows insolation (lines) and temperature (shading) at different latitudes (vertical axis) throughout the year (horizontal axis), and different charts represent different combinations of year length (ω, vertical axis) and obliquity (γ, horizontal axis). Guendelman and Kaspi 2019

But really most habitable worlds are likely to have shorter years, and so milder seasons. A shorter year also introduces more “lag time” (relative to the year length) between the solstices and polar seasons; at years 1/4 as long, midsummer comes closer to the fall equinox than the “summer” solstice.

Greater pole-equator temperature differences should also lead to stronger and somewhat wider Hadley cells, with the ITCZ moving farther poleward with the seasons. 

Rotational Period (Day Length)

These are not actually exactly the same; the sidereal day, the time it takes the planet to rotate once around its axis, is shorter than the synodic day, the time it takes the star to complete one apparent circuit through the sky (for a prograde-rotating planet). Where the day is much shorter than the year, the difference is small enough to be negligible. Where it is not, tidal forces are likely to lock the planet into a spin-orbit resonance, including 1:1 tidal-locking where the synodic day is effectively infinite, but the climates of such planets are so dissimilar from fast-rotating planets that we’ll leave discussion of them to another day.

Rotation causes the Coriolis effect, and faster rotation causes a stronger Coriolis effect. This means equatorward-flowing air will curve eastward faster, and so the Hadley cell will extend to lower latitude. For slower rotation, the reverse will happen.

However, the Coriolis effect is not the only factor determining the width of the Hadley and other circulation cells; friction between ground and atmosphere and formation of eddies near the equator have an impact as well. So the Hadley cell widens with longer days, but not as neatly as you might expect from the Coriolis effect alone.

Latitude of the top of the Hadley cell (i.e., location of the horse latitudes) for different rotation rates compared to Earth (e.g., days twice as long would be a rotation rate of 0.5 Ω). Kaspi and Showman 2015

As the Hadley cell widens, its associated pressure zones and prevailing winds must shift as well: The dry horse latitudes will shift north, and the easterly trade winds prevail over a larger area of the planet.

The other cells must shift poleward as well, though because they’re weaker their patterns can be a bit harder to predict. When an Earthlike planet has day lengths equal to 4 Earth days, the polar cells disappear; air rises at the poles and sinks at the poleward-shifted horse latitudes, in spite of a thermal gradient pushing air the other way.

At 16 days the Ferrel cells are also reduced to almost nothing, and a single massive Hadley cell dominates each hemisphere. But at the same time, complex momentum transfer in the upper atmosphere causes superrotation, circulation of air faster than the planet rotates, and so some westerly winds appear at the equator. Superrotation is also responsible for the Y-shaped clouds we can sometimes see on Venus, so we might expect similar cloud formations on these worlds (On Venus these clouds are only visible in UV, but this may be different for a habitable world with less atmospheric haze).

Venus's clouds in UV; the "Y" points antispinwards, as the equatorial clouds circle the planet slower than the polar clouds and so lag behind them. NASA

Conversely, doubling the rotation rate (days half as long as Earth) creates two new circulation cells in each hemisphere, continuing the pattern of alternating easterlies, high-pressure zone, westerlies, low-pressure zone. Further increases in rotation rate create even more convection cells. At less than 1/4 days it becomes difficult to count the exact number of cells due to eddies crossing between them and the relative weakness of each individual cell, and indeed the number may not be fully symmetric across the hemispheres or consistent across seasons; but air is consistently rising at the equator and sinking at the poles, implying an odd number in each hemisphere at all times.

Based on this paper, I’ve tried to measure out the approximate latitudes for the predicted convection cell boundaries in each hemisphere:

Day Length (Earth Days)	Convection Cell Boundaries (° Latitude)
16	3	70
8	0	65
4	0	55
2	0	40	70
1	0	30	60
1/2	0	25	40	55	70
1/4	0	18	21	26	33	41	49	56	64
Pressure	Low	High	Low	High	Low	High	Low	High	Low

Wider Hadley cells—helped along by stronger ocean currents—means more efficient heat transfer, and so a smaller equator-pole temperature difference for longer days. It also tends to lead to more regular and predictable wind patterns, and so fewer large cyclones.

Kaspi and Showman 2015

 

Total average temperature increases with day length up to 4 days by about 5 °C due to decreasing cloud formation at high latitude, then decreases thereafter due to increasing cloud formation during the long days; as much as 10 °C lower than current average temperature at 16 days, and 20 °C lower at 256 days (though in the latter case that’s an average between a long, cold night and equally long, sweltering day).

Black: Clouds and 50-meter-deep global oceans (Earthlike) Blue: Clouds and 1-meter-deep oceans (desert planet) Red: No clouds and 50-meter-deep oceans (probably an unrealistic scenario). Yang et al. 2014

Taking these effects into account, the portion of the Earth’s surface with hospitable average surface temperatures is maximized at 16 days—though this model had zero obliquity, so shorter days may be more optimal for higher-obliquity planets.

Average surface temperatures (top) and portion of the surface with temperatures between 0 and 100 °C at different levels of insolation (bottom) for different day lengths in a zero-obliquity model (the maps above correspond to the data points below). Jansen et al. 2018

Indeed, the effect of rotation rate on precipitation depends stronger on the planet’s obliquity. At very low obliquity, widening of the Hadley cells increases precipitation at mid latitudes (~10-50°) but weakening and then disappearance of the polar cell reduces it at high latitudes; areas poleward of 60° will become mostly arid at 8 days or longer.

Average land precipitation at different latitudes (excluding those with no land on Earth) at different day lengths and levels of insolation, using zero obliquity and Earthlike topography. Jansen et al. 2018

But with Earthlike obliquity, the wider Hadley cells allow for the ITCZ to move farther with the seasons. Past 8 days, it reaches almost to the poles near the solstices, causing heavy rains there and leaving the equator relatively dry. In essence, these planets have one giant planet-spanning convection cell near the solstices.

Average precipitation (mm/day) at different latitudes across the year at different rotation rates, using Earthlike obliquity and no topography (global ocean). Faulk 2017

As a final note, longer days for melting and longer nights for freezing lead to stronger ice-albedo feedback, especially past 10 Earth days. But while the proclivity towards a hothouse climate with no permanent ice caps increases, the proclivity towards a snowball state doesn't, so this may not be a big issue for habitability.

Obliquity

A.K.A axial tilt, the angle between a planet’s equatorial plane and its orbital plane, which is also the highest angle that the sun will appear above the horizon at the poles and the highest latitude that will see the sun directly overhead at summer solstice (for obliquities up to 90°; planets with obliquities above 90° are essentially identical to planets with obliquities equal to (180° - obliquity) in terms of climate, unless the rotational period is a significant portion of the orbital period). Any latitudes greater than (90° - obliquity) will receive daylong sun for at least part of the summer, which can cause them to receive higher average insolation over the day than lower latitudes, even if peak insolation at noon is lower.

cmglee, Wikimedia/NASA

I talked about obliquity and its relationship to seasons back in Part IV, but just to review:

• Obliquity causes Earth’s seasons because the insolation of each hemisphere increases and decreases as it is pointed towards or away from the sun.
• Higher obliquity decreases the difference in average temperature between equator and pole (to ~50°), but increases seasonal temperature variability.
• Past 18° obliquity, the poles receive more insolation than the equator at the summer solstice.
• Past 45°, the sun is higher in the sky at the poles than the equator at summer solstice.
• Past 54°, the poles receive more average insolation throughout the year than the equator (they still has lower insolation in the winter and during equinoxes, but this is offset by the constant summer sun).
• Any nonzero obliquity will cause the poles to experience summer and winter at opposite solstices, while high obliquities cause the equator to experience 2 equinox summers and 2 solstice winters every year.

Increasing obliquity somewhat increases average global temperature due to shifts in cloud cover—by about 9 °C from 30° to 90° in otherwise Earthlike conditions. In the other direction, an Earth with no obliquity would be about 4° C colder than today, with permanent polar icecaps extending to ~ 50° latitude. But where a low-obliquity world has comfortable temperatures, they should be near-permanent, while at high obliquity the whole planet will experience large seasonal temperature swings.

Average temperature across the year at different latitudes for different obliquities (using an idealized global ocean planet). Nowajewski et al. 2018

Past ~50°, our assumption in the model scenarios at the start heat flows from equator to poles is no longer valid. Exactly what happens to atmospheric circulation depends on insolation and rotation rate:

At high insolation or low rotation rate (in both cases the transition is near Earth-equivalent values), the major circulation cells are still present, but they switch direction; hot air rises at the pole and cool air sinks at the equator, and so prevailing winds will switch direction as well. These flipped wind patterns will cause moisture to converge at the poles in summer, which could make for intense seasonal storms. However, summer temperatures over the continents could become so high that rain cannot occur even at high humidity, so a typical summer includes an extended drought followed by torrential downpours in fall. Elsewhere, precipitation is distributed fairly evenly over the summer hemisphere (the boundaries of the convection cells aren’t as impactful as at low obliquity, and move far over the year), but the equator remains dry.

Average precipitation in different months with 85° obliquity. Williams and Pollard 2003

At low insolation or high rotation rate, eddy patterns near the equator and friction between the ground and atmosphere help to keep the circulation cells going in the same direction as they do at low obliquity, but with far weaker Hadley and polar cells, wider Ferrel cells, with the ITCZ moving much farther over the year; some heat is still transported from pole to equator, but we can assume these worlds have higher equator-pole temperature differences and are more prone to glaciation.

Of course, varying rotation rate affects the number and strength of circulation cells as well, but we won’t work through all the possible variants here; though one intriguing possibility at particularly low rotation rates is that there may be a single convection cell from one pole to the other that switches direction across seasons.

In either case, winds are fairly weak in winter and temperatures fairly similar across most of the winter hemisphere because a large part of it receives the same amount of insolation; zero. Nevertheless, oceans at any latitude can remain ice-free throughout the year; average sea level temperature at the poles for an Earthlike planet with 90° obliquity and good ocean circulation varies between 12 and 42 °C over the year. But continents are more variable, and deep interiors at high latitudes can reach -30 and 100 °C under similar conditions. Such extreme temperatures—along with the aforementioned seasonal rain patterns—could be a major challenge for any life on such a planet.

Average surface temperature in different months with 85° obliquity. Williams and Pollard 2003

Intriguingly, it is possible for planets with over 54° obliquity and fairly weak equator-pole ocean circulation to develop an equatorial ice belt without permanent ice caps at the poles; it just requires that the entire planet freeze over first, and then the poles thaw. An equatorial ice belt with seasonal polar ice cover is also possible above 30°.

Meanwhile, Earth’s current condition—permanent polar ice caps with an ice-free equator—is only possible up to 35°, and only from an initially fully thawed state; for low-obliquity planets, recovery from a snowball necessarily results in total loss of the ice caps.

Chart of conditions under which it is possible to attain certain climate states starting with global ice cover (left) or no ice (right); cryoplanet (global ice), near cryoplanet (equatorial ice belt and seasonal polar ice caps), uncapped cryoplanet (ice belt and ice-free poles), capped aquaplanet (ice caps and ice-free equator, like Earth), near aquaplanet (seasonal ice caps), and aquaplanet (no ice). Carbon-silicate cycle isn't accounted for, so think of "relative stellar irradiance" not as literal insolation but more an abstract representation of relative warming or cooling, with the widths of climate states representing how stable they are at different obliquities.

Note, by the way, that it appears possible for a planet at 30-35° to have polar ice caps and an equatorial ice belt at different points in its history, though this is the obliquity range where ice-albedo feedback is strongest and so partial ice cover is least likely at any given point in time.

But of course, that can happen to any planet if the obliquity varies over time. Earth’s obliquity oscillates between 22.1° and 24.5° and back over a 41,000 year period, one of the so-called Milankovitch cycles. Small though this variation is, the shift in polar insolation it causes seems to be a major factor in the glacial-interglacial cycle. Any planet in a system with other planets influencing it is likely to undergo such cycling (you can use orbe to get a sense of it for a given system if you like). In certain systems—multiple star systems especially—these cycles might be much shorter, as little as 1,000 years, causing constant rapid shifts in climate.

The amplitude of these variations can vary as well: From its current value of 25°, Mars’s obliquity has likely varied from to over 60° and under 10° over its history.

One last thing to remember: obliquity is not a free value. Initial obliquity could be more-or-less anything thanks to large impacts late in planet formation, but tidal forces from a star or moon will tend to reduce it over time. The forces from the star are stronger in the habitable zone for lower-mass stars, to the point that it’s near impossible for a habitable planet to retain any high obliquity for billions of years around a red dwarf.

However, the gravitational influence of other planets could help maintain some obliquity in these cases, and encounters with other stars could alter the inclinations of these planets—in effect, changing their obliquity.

Armstrong et al. 2014

Eccentricity

Variation of the distance between star and planet due to an elliptical orbit. This is an interesting case of a forcing that is almost absent on Earth but may be significant for other planets. For a given semimajor axis, higher eccentricity slightly increases average insolation, and so leads to increased average temperatures—around 10-20 °C from 0 to 0.5 eccentricity. But with increasing eccentricity comes an increasingly short and intense periapsis summer and a long apoapsis winter. These higher temperatures are accompanied by higher average precipitation, with an annual cycle of wetter low latitudes near periapsis and wetter high latitudes near apoapsis.

Temperature across the year with varying eccentricity, Earthlike obliquity, and a solstice at periastron. Dressing et al. 2010

In general higher eccentricity seems to cause lower equator-pole temperature contrasts and weaker ice-albedo feedback. But exactly how it affects temperature and circulation patterns depends on how it interacts with obliquity. If periapsis lines up with a solstice, it will strengthen seasons in one hemisphere and weaken them in the other; if it lines up with an equinox, hemispheric seasons will be generally milder but with spring/fall transitions varying across hemisphere and low latitudes will experience an annual seasonal cycle.

At low eccentricity, whichever hemisphere has its summer solstice closer to periapsis should be warmer on average and have smaller ice cover. But at very high eccentricity and fairly low insolation, the winters in that hemisphere are so long that an ice cap forms that cannot be fully thawed in the intense but brief summer, and so this becomes the colder hemisphere.

Earth’s eccentricity undergoes its own Milankovitch cycles, somewhat more erratic than the obliquity cycles but with a roughly 100,000-year period. The orientation of Earth’s rotational axis and the position of periapsis in Earth’s orbit oscillate as well, combining to create a roughly 23,000-year period between times when periapsis coincides with northern summer solstice. At such times the ITCZ moves further north in summer, bringing more moisture into what are currently dry deserts and causing periodic “greenings” of the Sahara. The southern hemisphere likely has broader arid zones at the same time, with the situation reversed when periapsis coincides with northern winter solstice.

Altogether this is a secondary control on the glacial-interglacial cycle, but in a world with higher eccentricity it could be the primary control.

Planet size

Though not an element of a planet’s motion, I’m including this as an “external” astronomical factor, given that it’s set at the end of formation and cannot be changed afterwards without effectively sterilizing the planet. We’ll assume all our habitable planets have roughly the same composition as Earth, and so mass, radius, and surface gravity are all closely linked.

Unsurprisingly, increasing radius causes a greater equator-pole temperature difference, because the heat has farther to travel—though the actual gradient of temperature change over distance decreases, thanks in part to how greater gravity affects air flow. A planet with higher density than Earth but the same radius would have a lower equator-pole temperature difference instead.

Equator-pole temperature difference (Blue, left bar) and gradient over distance (Red, right bar, equivalent to °C / 1,000 km) for planets with constant, Earthlike density of 5.52 g/cm³. Kaspi and Showman 2015

This lower gradient causes weaker winds, and though the Hadley and Ferrel cells grow in absolute width, they extend to lower latitudes—by around 5° and 10°, respectively, for a doubled radius.

Greater gravity also reduces the content of water vapor in the air, reducing the greenhouse effect and therefore lowering surface temperature by a few °C. This will probably reduce overall precipitation as well.

Geological Forcings

Those related to tectonic activity, continental drift, or other geological processes on the surface. These can all change over a planet’s lifetime regardless of the wider astronomical context.

Volcanic Outgassing

This is perhaps the most direct way to alter average global temperatures. Other forcings will be restrained by the carbon-silicate cycle, at least to some extent, but altering the rate at which CO­₂ and other greenhouse gasses are emitted by volcanoes directly alters the equilibrium point of the carbon-silicate cycle. So I’ll take this opportunity to talk not only about how this rate can change and what effects that will have, but also about how shifting global temperatures in general (which can still be caused by other factors) affects other climate factors.

I’ve described the carbon-silicate cycle before, but in short it’s the cycling of carbon by outgassing from volcanoes as CO₂, weathering of surface minerals with CO₂ and water to form dissolved bicarbonate, deposition as carbonate minerals, subduction into the mantle, and then decomposition back into CO₂ dissolved in magma that rises to the surface again. Because increased CO₂ increases temperature and increased temperature increases weathering that draws CO₂ out of the atmosphere, a stabilizing negative feedback results: when other factors alter the global temperature, CO₂ levels will tend to increase or decrease until the rate of weathering once again matches the rate of volcanic outgassing. But if the rate of outgassing changes, then the temperature at which this equilibrium is reached changes.

Thankfully, it shouldn’t change by much; small changes in temperature cause large changes in the weathering rate. So long as rates of volcanism don’t rise to something like Io or fall to something like Mars, we should stay comfortably within habitable temperatures. But the subtle effects are worth some consideration.

For one thing, plate tectonics is, while smoother than most of the alternatives, not a perfectly smooth ride. Events like the closure or opening of oceans or other shifts in the global geometry of tectonic plates will cause (relatively) rapid shifts in plate motion, accompanied by increased volcanism, and so increased temperatures by a few °C. But once this episode of increased tectonic activity ends, it’s likely to have left behind large areas of exposed, young rock that can continue weathering for tens of millions of years. Just as increased volcanic outgassing raises the temperature at which the carbon-silicate cycle balances, increased potential for weathering lowers it.

Such a pattern pretty well explains climate shifts over the last 66 million years, the Cenozoic Era. Early in the Cenozoic, intense subduction zone volcanism across the entire western coast of the Americas and around the edges of the Tethys sea caused high temperatures, peaking in the Eocene around 50 million years ago at around 30 °C. Since then, most subduction in North America has ceased (for now) and the Tethys has closed, ending much of the volcanism and leaving large mountain ranges (the Rockies, the Himalayas, etc.) to weather down (there have been some new island arcs forming with active volcanism, but these seem to have a smaller impact on climate). Thus, temperatures have dropped, to 15 °C today and as low as 9 °C in the depths of a glacial episode.

Average temperatures over the last 66 million years. Hansen et al. 2013

And of course, under different tectonic regimes the rates of volcanism could be much more variable, leading to larger and more rapid shifts in climate. Though it causes its upsets, faster plate motion tends to decrease climate variability overall. An episodic or sluggish-lid mode with long periods of slow motion and less vigorous volcanism may allow for wild swings in climate beyond what we’ve seen in the last ~800 million years of modern plate tectonics.

On any planet, the rates of volcanic activity should generally decline over time. What effect this has on climate depends on how this decreasing forcing from volcanic outgassing compares with the increasing forcing from the brightening star. On Earth, the two trends appear to be balanced, but it’s hard to say with poor long-term temperature records. Presuming this is the case, and presuming that most habitable worlds will orbit smaller, slower-evolving stars, the general tendency for Earthlike worlds may be to cool as they age. The general increase in continental area—weatherable land—over time may help as well (presuming that trend actually holds for older planets).

So what effects will increasing or decreasing the global average temperature have on other climate factors? First off, increased evaporation and other shifts in atmospheric circulation tend to make dry areas drier and wet areas—particularly the tropics—wetter. The distribution of these areas will change as well; the Hadley cell widens by about 1° latitude for every 4 °C of increased average temperature, pushing the horse latitudes poleward.

But as the ice caps melt, polar albedo decreases and so the equator-pole temperature difference goes down as well. Mid Cretaceous hothouse seas were around 5 °C hotter at the tropics, but as much as 20 °C hotter near the poles. This causes the trend of a widening Hadley cell to reverse: from a peak width of 35° latitude at 21 °C (perhaps cooler in a world with lower insolation or redder starlight) it shrinks down to 20° latitude, pulling the horse latitudes towards the equator and concentrating tropical rain into a thin equatorial belt. This shrinking likely happens much more rapidly than the initial widening of the Hadley cell.

Hasegawa et al. 2012

Going to the other extreme, a snowball planet would have fairly low thermal inertia across its surface, and so the convection cells would move far poleward with the seasons. The equator would spend more time near the dry edges of the Hadley cells than near the wet ITCZ—not that it would matter much to surface conditions, given that the whole planet would be dry and there’d be little in the way of open surface sediment on which vegetation could grow.

If, as some models predict, there could be a thin belt of open ocean around the equator, then the low albedo of this area would keep it far warmer than the icebound areas, and so the ITCZ should remain close to the equator year-round even at fairly high obliquity. So the open area would have some precipitation year-round, but it still may not be much overall. The coldest climate with open water remaining—what researchers have termed a Jormungand state (because the water belt resembles a world-encircling snake)—has ice reaching to 10° latitude (averaged across the seasons), global average temperatures below -10 °C, poles plunging below -80 °C, and Hadley cells only 20° latitude wide with tropical rains less than half those today.

Average temperatures at different latitudes for hothouse (red), Jormungand/"slushball" (blue) and snowball (black) states, with Earthlike obliquity. Abbot et al. 2011

Land/Water Ratio 

The total area of the planet’s surfaces with land, compared to the amount with oceans, without necessarily accounting for the depths of the oceans or ice cover. This ratio will be determined in part by the age and tectonic history of the world, and in particular I’ve established before that a world with more land than ocean is unlikely to have functioning plate tectonics. But we’ll ignore the deeper implications of that here, and just consider the impact of altering land area on its own.

Generally speaking, we tend to refer to worlds with a higher land/water ratio as “dry”, and in a broad sense this is true, but it can be a bit misleading regarding the critical factor of how much land area with frequent precipitation—i.e., fertile ground—there is. A supercontinent will have less fertile ground than multiple continents that have more total area but are individually smaller. In the extreme cases, an ocean world with a small continent and a desert world with a small sea could have similar areas of fertile ground.

But regarding broad impacts on climate, there are a few clear points: Land has a higher albedo than open water, so all else being equal a world with more land will be cooler. Earth with all land removed (and no alteration in CO₂) would be a few °C warmer than today. There’s also a lower contrast in albedo between ice or snow and land, as opposed to ice and ocean, so ice-albedo feedback is weaker with more land area. A planet in the outer habitable zone with low volcanic activity may avoid limit cycling for billions of years longer with 50% land area as opposed to 10% land area.

More area for weathering will cool the world as well, though weathering also requires water so this may be less true at very high land areas. Except for planets that have very high rates of CO₂ outgassing, the difference in weathering between 50% land area and 1% is unlikely to account for more than a 10 °C difference—and even for high-CO₂ planets, the effect is only significant below ~10% land area.

Likely average temperature for different land areas and initial carbon contents. Foley 2015

As land area increases, ocean circulation will become more restricted and eventually blocked, cooling the poles and warming the equator. Less water also means less thermal inertia, so there will be major temperature swings with the seasons. Overall, then, more land increases temperature contrasts.

In the most extreme cases, of dry planets near the inner edge of the habitable zone (which can be much further in for such worlds), water must be restricted to strips around the poles. Still, the seas here can extend pretty far equatorward—to 70° latitude or more—even while average temperature at the equator surpasses 100 °C. Precipitation will be concentrated towards the poles, but still even lower than these regions receive on Earth.

Continental Drift

Even for a given amount of land, the position of that land can have a profound effect on climate. Though continents are generally cooler than oceans anywhere on the globe, when near the poles they have a better chance of developing large glaciers, cooling themselves and the whole planet. The drift of the continents towards the poles throughout the Cenozoic is another possible factor in the Earth’s gradual cooling trend.

Average temperatures (°C) at different latitudes for different land distributions. Source

The large interiors of supercontinents receive little precipitation from the distant coasts, and are also far from the temperature-moderating effects of the oceans. Central Pangea likely surpassed 50 °C in summer. Winter is similarly cold, but that summer heat can help prevent glaciation even when the supercontinent is over a pole—though supercontinents also tend to have high internal mountains, which can help kickstart the glaciation process.

The position of continents also affects the pathways available for ocean circulation. A broad east-west strip of continent at mid-latitudes can block ocean transport of heat poleward, leaving the poles several °C cooler than they would be otherwise. But open ocean can have a cooling effect as well in the right circumstances: A ring of uninterrupted water circles Earth at 60° south latitude, near the average position of the southern polar front, resulting in a circumpolar current circling the world that, to some extent, isolates Antarctica from warmer waters near the equator. How much is still debated—it probably wasn’t necessary for glaciation, but contributed to it. Overall, shifts in ocean circulation seem to be a secondary factor in global climate.

How the specific geometry of land and ocean can cause warming or cooling at high latitudes is a subject we’ll dig into more in the next post.

And, of course, the position and movement of continents is linked to rates of volcanic activity and mountain formation, and through these outgassing and drawdown of CO₂.  Young mountains in the hot, wet tropics especially seem to increase weathering and so decrease global temperatures. Breakup of supercontinents may bring high rains into what was once dry interior terrain, increasing weathering and cooling the climate—though perhaps only briefly before plate speed and volcanic outgassing picks up.

Ocean Salinity

The salt content of water affects its freezing temperature and density. Ocean salinity has been gradually declining over the last half-billion years, and probably could vary widely between different planets.

High salinity lowers freezing temperature, which inhibits ice formation, and less ice means a lower albedo for the whole planet. It also alters ocean circulation; rather than flowing poleward along the surface, high-salinity water sinks at mid-latitudes and then rises back up at the poles, warming them more efficiently and so reducing the equator-pole temperature difference.

As a result, Earth with twice the ocean salinity may have no sea ice at all, and global average temperatures 6 °C higher.

Freezing and thawing of ice can alter the salinity of the oceans—salt is left out when the ice forms and so salt is concentrated in the shrinking oceans as ice forms, or vice-versa when it melts. Over long periods this is probably compensated for with salt deposition on the sea floor and influx from the continents, but in the short term rapid changes in ice volume can alter salinity and lead to rapid shifts in ocean currents. As Earth transitioned out of its most recent glacial, the influx of fresh water into the Atlantic may have weakened ocean circulation and caused the Younger Dryas event, a cold snap lasting around 1,000 years , before warming resumed.

If this happens consistently—warming causing weaker currents that cool high latitudes, cooling causing stronger currents that heat high latitudes—it might form a negative feedback that helps moderate climate shifts, but it’s not clear exactly how well it would work with different salinities.

Atmospheric Forcings 

Factors caused by the composition and properties of the atmosphere, independent of the controls of other forcing. So though atmospheric CO₂ levels are a main factor in how the atmosphere affects climate, we’ve already discussed the controls and effects and we don’t need to go over them again here.

Pressure 

Exactly how surface pressure affects temperature depends on the composition of the atmosphere, particularly greenhouse gasses. If pressure is increased with Earth’s mix of gasses—including CO₂—held constant, then average surface temperature increases to a peak of 22 °C to 4 bar, then decreases thereafter due to increasing albedo until it reaches 0 °C at 34 bar. However, if we presume that CO₂ levels will vary to counteract these effects (probably partially but not totally true) then there will be a more gradual increase up to about 30 °C at 100 bar. In either case the equator-pole temperature decreases with greater pressure. So in general, thicker atmospheres hold more heat near the surface and distribute it more evenly. 

Average temperatures at different latitudes with different atmospheric pressures and no change in greenhouse heating. Kaspi and Showman 2015

Increasing pressure also narrows the Hadley cell, and can even prompt the formation of more circulation cells; a 10-bar atmosphere has two more circulation cells in its atmosphere, just like on an Earth rotating twice as fast. Wind speed generally decreases, but given how much density of air increases, the power of the wind may not drop much—meaning wind-driven erosion and potential of using wind for energy may not change much (none of the research seems to model this specifically, so I’m not sure exactly how it would change).

Each chart shows insolation (lines) and temperature (shading) at different latitudes (vertical axis) throughout the year (horizontal axis), and different charts represent different combinations of rotation rate (Ω, vertical axis) and surface pressure (P_s, horizontal axis). Guendelman and Kaspi 2019

As I’ve mentioned, the strength of ice-albedo feedback decreases with a greater greenhouse effect, and it appears this may generally happen with increased pressure, one way or another. Albedo of the atmosphere itself also increases, so this may reduce the impact of surface albedo. 

Oxygen

Increasing Oxygen alone increases the molecular weight of the atmosphere, which tends to increase scattering of light and so cools the surface and reduces evaporation. As with increased pressure, it probably reduces ice-albedo feedback.

Average global temperature and precipitation for different levels of oxygen (percentages on chart) and CO₂ (colors, values in legend). Poulsen et al. 2015

Biological Forcings

To wrap it up, a couple forcings that result from the activity of life on the surface—at least, the near-term results of that activity. Life has had a profound effect on the atmosphere and geology of Earth over its history, but we’ve already discussed those forcings.

Plant-Like Vegetation 

Plant life on Earth has a variety of subtle climate impacts—some affecting immediate climate and some playing out over billions of years—but there’s three I want to highlight here (naturally, I can only guarantee this is true for planet life as it appears on Earth, or something very similar, but I wouldn't be surprised if these were common to most forms of complex plant-like flora).

First off, plants put down roots that break up soil and rock, exposing more material for weathering. The first appearance of large rooting plants in the Devonian (and then seeding plants not long after) may have caused a plunge in temperatures, leading to a mass extinction event. Since plant life dominated the continents there has been less untouched soil for them to overturn except in fresh volcanic terrain, so their effect is less dramatic. But it’s probably fair to say that Earth today is cooler than it would be without rooting plants, and that major shifts in plant cover—due to, say, shifts in precipitation patterns—could cause warming or cooling on Earthlike worlds. Though, plants on Earth have a lower albedo than desert, which could partially offset this effect (presuming the dominant surface material of a lifeless Earth would be similar to modern deserts, which is tricky to determine but probably close enough).

Second, plants transport water from the soil into the air by evapotranspiration: they evaporate water from above-ground surfaces on leaves and stems to create negative pressure that pulls water up from their roots. More water in the air means more precipitation, which will cause more plant growth, and so on in a positive feedback until limited by the total water content of the system. The effect is most obvious in arid regions, or arid planets: in what might be a desert without life, plants can kickstart a water cycle with regular rain.

Modelling for Earth suggests that the presence of plant life may have doubled precipitation over land, and particularly increased it in arid regions. Increased cloud cover also cools the planet by around 1 °C (in addition to the effect of increased weathering).

Major climate zones for Earth with (left) and without (right) widespread land vegetation. Kleidon et al. 1999

Some plants also release aerosols (in particular monoterpenes) that enhance cloud formation and so increase local precipitation. This may lead to a negative feedback related to forest growth at high latitudes:

When global temperatures increase, the extent of boreal forests at high latitudes increase, and these forests release more aerosols that increase formation of high-albedo clouds and so cool the planet.

When global temperatures decrease, boreal forests retreat, fewer aerosols are released, and cloud formation decreases and the planet warms.

This may help to counteract ice-albedo feedback, as will the general increase in cloud cover.

Fossil Fuel Use

This one is, you know, us. The causes and effects of anthropogenic climate change are worth their own discussion at another time, but in short under 2 centuries of widespread fossil fuel use has caused around 1 °C of warming, and we could easily add another 2 °C or more by the end of the century. It’s not much compared to most of the forcings we’ve been discussing, but it doesn’t take much to disrupt global agriculture. In a sense, we’re replicating the effects of an extremely violent and sudden outburst of tectonic activity.

Intriguingly, this may not be a hazard for all habitable worlds. Climate change on Earth has been caused by an increase of a couple hundred parts per million of CO₂—a lot compared to pre-industrial levels of ~280 ppm, but nothing compared to the multiple bars of CO₂ that may exist on worlds in the outer habitable zone. So perhaps the inhabitants of such a world could merrily burn through their reserves of fossil fuels without worrying about the consequences (aside from the economic ones when supplies run short).

Sea Level

Though sea level is an element of climate, the mechanisms driving sea level change are distinct enough that the subject deserves its own discussion. There are various geological processes that can cause local sea level change, but for now I’ll only discuss eustatic or global sea level.

Most sea level forcings are cyclical (sea level rises and falls and rises again), though how regular they are varies. We can broadly categorize them based on the timescales over which they tend to operate and the magnitude of changes they tend to cause. Rapid (<100,000 years) shifts tend to be caused by changes in water volume, while more gradual (>100,000 years) shifts tend to be caused by changes in basin volume. You can think of it as the difference between pouring water in and out of a bowl and changing the size of the bowl.

Typical speed and magnitude of major sea level forcings. Miller et al. 2005

Tectonics

The longest-period cycles of sea level change are linked to the supercontinent cycle, which you may recall I described in Part Va; When supercontinents break up, new ocean basins form with young, hot crust that “floats” high on the mantle, while older, less buoyant crust is subducted away, all of which reduces total basin volume. After breakup, the crust in the new oceans ages and sinks, production of new crust and subduction of old crust slows, and so basin volume increases. Once a new supercontinent begins to assemble, continent-continent collisions reduce the area of land and so increase the volume of the ocean basins, and aging mountains and coastlines cause decreased deposition of sediment into the oceans.

Within the last few hundred million years, one of the periods of lowest sea level—close to today’s levels—appears to have occurred around 250 million years ago, just before Pangea’s breakup, while high sea levels—up to 250 meters higher than today—occurred around 100 million years ago, during breakup, causing shallow seas to flood much of the continents. Sea levels have since dropped again as the breakup has slowed.

Modeled depths of the ocean basins (relative to modern sea level) since the mid-Cretaceous sea level high. Muller et al. 2008

Climate State

As Earth has shifted between icehouse and greenhouse states, this has caused increases and decreases in the amount of water trapped in glaciers on the continents rather than in the oceans (sea ice doesn’t cause sea level fall, but can help confine and insulate glaciers on nearby land). The climate states are linked to volcanic activity, which are linked to the supercontinent cycle, so distinguishing between these effects is a bit tricky (glaciations do not always coincide with supercontinent formation, because position and topography of the supercontinent matters as well, but it generally has for the last half-billion years over which we have good sea level records). But based on recent sea level shifts we can pretty confidently say it amounts to over 100 m of sea level change.

Sea level over the last ~550 million years, compared to temperature (climate forcing) and ocean crust production (tectonic forcing). Source

Lakes and Groundwater

Gradual tectonic movement and small shifts in climate will cause lakes to form or fill in, and more or less water to be stored in groundwater. This can cause variations of 10s of meters over millions of years; little enough to be overwhelmed by Milankovitch cycles during icehouse states, but during hothouse states it’ll be the main source of short-term variability.

Milankovitch Cycles

Once ice caps form during an icehouse state, the glacial-interglacial cycle of warming and cooling will cause a linked cycle of glacial advance and retreat, and so sea level fall and rise. During the last glacial maximum, sea level fell to 120 m below its present level. These cycles last 10s to 100s of thousands of years on Earth, but could conceivably be quicker for another world with rapid eccentricity or obliquity changes due to the influence of a nearby gas giant—though in those cases the magnitude of sea level change may be reduced, as glaciers have less time to advance or retreat.

Thermal Expansion

During rapid shifts in temperature—due to volcanic activity, impact events, or, uh, “biological activity”—some small sea level change can occur even before much ice has had a chance to freeze or thaw due to changes of water’s density with temperature. This is unlikely to amount to more than a few meters, but that can be enough to impact coastal ecosystems and communities.

Mantle Cooling

In addition to these cycles of sea level change, there may be a much more gradual trend of sea level change as the mantle cools—too gradual to be observed with current data, but suggested by some models. The hot, young mantle outgasses most of its water to the surface, peaking with oceans about twice as voluminous as today when the planet was 1.5 billion years old. Then as the mantle cools and subduction begins water is drawn back into the mantle. Thus the early Earth may have been completely inundated save for a few volcanic islands, and Earth in another few billion years time may have oceans with 1/4 the volume of the current ones (or it would were the oceans not totally boiled away by the brightening sun long before then).

Projected masses of surface water (compared to Earth's current oceans) for planets of different masses as they age. Schaefer and Sasselov 2015

However, this is only the case so long as plate tectonics lasts; if reduced oceans cause plate tectonics to break down, the remnant oceans may last for far longer. The volume of the oceans is also heavily dependent on the initial water content of the planet, and planet mass; More massive planets with proportionally similar water contents will have a higher peak of surface water (even relative to their greater surface area) but dry out quicker.

An Evolving Planet

What does all of this mean for our example world, Teacup Ae? It’s generally pretty similar to Earth, but we can tally up all the possible effects of the minor differences:

• Average Temperature: Increased by a redder star, longer day, greater eccentricity, and greater atmospheric pressure, decreased by lower insolation and lower obliquity. Size, volcanic outgassing, land area, latitudinal distribution of the continents, ocean salinity, oxygen content, and flora cover are pretty similar. Overall we can justify it as being fairly close to Earth—I intend to give Ae a similar icehouse climate.
• Equator-Pole Temperature Difference: Increased by lower obliquity, decreased by a redder star, lower insolation, shorter year, longer day, greater eccentricity, and greater pressure. Though lower obliquity is the only forcing increasing the difference, it is a pretty powerful one, so again we might end up with a situation similar to Earth but we’ll tweak with the factors a bit in the next post.
• Hadley Cell: Widened by a longer day, shrunk by a shorter year and increased pressure. Again probably fairly similar to Earth—which is getting a bit repetitive, but building a loose Earth-analogue was the idea from the beginning. A redder star, lower insolation, shorter year, and greater pressure all seem to indicate somewhat weaker winds, though.
• Rainfall: Decreased by a redder star and lower insolation. A longer day with low obliquity may offset this for low latitudes, and greater eccentricity should cause some asymmetry between the hemispheres.
• Ice-Albedo Feedback Sensitivity: Increased by a longer day, decreased by a redder star, lower insolation, lower obliquity, greater eccentricity, and greater pressure. Generally it seems that Ae may have a more stable climate with partial ice cover than Earth, and so may spend more of its time with partial ice cover.

We’ll dig into the specifics of Ae’s “modern” climate in the next post, and tweak some of the forcings where necessary. But before then, we can look at the planet’s tectonic history that we constructed in Part Va and use it as an example of how a planet’s climate might change as its continents drift, collide, and break apart at different stages of the supercontinent cycle. I’ll take a few representative snapshots of the planet’s history, and put together a basic picture of the climate and major climate belts in each; I’ll explain more about how to determine climate zones in the next post, and at any rate these are more sketches than definitive maps.

(A quick key to these maps: Orogenies are black while active then fade to grey as they age, white is permanent ice cover, yellow is desert, dark green is dense forest, light blue is inland seas on the continents, and the grey lines are lines of latitude at 30° increments—i.e. where the boundaries of the convection cells should usually be.)

800 mya: Cuvieric

Orogenies start black when active then fade to grey as they age, the horizontal lines are lines of latitude at 30° increments—i.e. where the boundaries of the convection cells should usually be.

At this point we don’t have much pre-existing topography to work with, but we can guess that there would likely be substantial mountain ranges near the core of the initial supercontinent. Deserts (shown in yellow) will prevail in the interiors of the vast continents, especially in areas surrounded by the young coastal ranges blocking winds from the seas.

We’re still in the early stages of breakup, so volcanic outgassing hasn’t picked up yet but much of the former interior is exposed to weathering. Still, there is some new volcanism, and not much land near the poles, so we’ll call this a cool but ice-free hothouse, at least by the midpoint of this period. For reasons I’ll dig into another time, I’ll say that there was a snowball period immediately preceding the Cuvieric, during the tenure of the last supercontinent.

700 mya: Anningic

Well into the breakup process now, widespread volcanism is pushing Ae well into the hothouse zone, and high sea level causes shallow seas (shown in light blue) to push deep into the continent interiors (I’ve had to pretty much just guess at the topography). Even the large continent over the north pole is ice-free save perhaps for some winter snow and mountain glaciers, and can reach pretty pleasant temperatures in summer—though there is as yet no land vegetation to take advantage of this.

At lower latitudes, a thinner Hadley cell brings deserts closer to the equator, though with many small landforms there is generally less desert area—though, again, without vegetation the continents are pretty barren throughout.

550 mya: Owenian

As continents collide and the seas age, the world cools and sea levels drop. The collisions are happening near the poles, so we might expect some glaciers to form in the mountains, perhaps even a few cold snaps with proper ice formation, but this also means that their formation doesn’t increase weathering much so the planet remains mostly hothouse for now.

It’s in this period that I’ll say land plants begin appearing, first as low moss-like forms near rivers and coasts (dense vegetation shown in dark green). As they diversify and spread, they may help cool the climate further in the late Owenian, triggering a proper ice age.

400 mya: Huxleyic

Closure of many subduction zones and movement into or formation of mountain ranges in the tropics has drawn CO₂ levels down and pushed Ae into a proper icehouse state, with large icecaps (shown in white) spreading from the poles—especially in the southern hemisphere, helped along by the high mountains and vast interior of the large continent assembling there. Where the ice has not reached, new deserts are spreading.

Still, life survives and proliferates, and by now vast forests spread across the areas that are still warm and wet. Take note of where these first major forests appear; they’ll become the coal-producing regions of the future.

250 mya: Marshian

New subduction and shifting of the continents away from the poles has helped pull Ae out of the ice age, though it remains relatively cool and large glaciers probably still remain in the southern mountains.

But as the supercontinent grows, a vast desert spreads across much of its interior, and will only grow more as assembly continues. The center is almost completely dry and lifeless, and sweltering in summer. Still, substantial forests remain flourishing on the coasts.

150 mya: Copian

The supercontinent breaks up, accompanied by a period of intense volcanism (an LIP) that probably causes a brief rise to extreme temperatures, though they fall again as moisture enters the formerly dry interior and weathering picks up. As the breakup continues, temperatures and sea levels will rise again

The formerly indomitable desert is split in half and forests spring up on the coasts of the young Bischoff Ocean, helped along by wider Hadley cells. The splitting of the supercontinent and meeting of species from formerly separated coasts will no doubt spark a period of diversification for Ae’s life.

50 mya: Andrewsian

Continuing breakup and volcanism drives temperatures and sea levels up, and shallow seas once again flood the continents. Marine life flourishes in these seas, and the deposited carbon will eventually become the oil that can be extracted by future civilization.

The very hot conditions pull the Hadley cells inwards, and a fortuitous arrangement of continents means there’s relatively little desert area in this period. Instead, lush forests straddle the equator and spread deep into the polar regions.

0 mya: Ostromian

Just as Earth has cooled through the Cenozoic, Ae cools through the late Andrewsian and Ostromian, for similar reasons: Collisions in Hutton/Lyell and Wegener close subduction zones and create weatherable mountain ranges in the tropics, and several continents move towards the poles. The situation shown here is during an interglacial; during a glacial episode, the glaciers will extend over much of Hutton, Steno, and Agassiz.

As this is the “modern” condition of Ae, we’ll use my sketch as a starting point to flesh out a more detailed picture of the climate on this world in the next post. But for now, here’s a quick look at the climate history I’ve assembled—as with the climate sketches, this is all more suggestive than definitive, and probably misses some brief climate excursions:

In Summary

• The climate is driven by atmospheric convection of heat from the equator (usually) to the poles.
• Earth’s rotation and the Coriolis effect causes each hemisphere to be split into 3 convection cells (Hadley, Ferrel, and polar), with alternating equatorward easterly winds and poleward westerlies at low altitude. 
• These winds create rain belts near the equator and 60° latitude, and dry belts near 30° latitude and the poles. 
• Variations in surface albedo and thermal inertia and seasonal shifts in insolation cause the boundaries between these cells to move over the year. 
• Earth has passed through 3 distinct climate states: Hothouse with no ice caps, icehouse with partial ice cover and glacial-interglacial cycles, and snowball with total ice cover. 
• Greater sensitivity to ice-albedo feedback causes the climate to transition more rapidly between hothouse and snowball states, with reduced stability in the icehouse state. 

Regarding the many forcings, here’s a chart of the main climate impacts of each (Green = Increases; Red = Decreases; Yellow = Mixed impacts; Blue = No known major impact independent of other factors; Darker/more saturated color = Stronger effect; forcings are increased from minimum to maximum habitable values, not just relative to Earth):

Forcing	Average Temp.	Equator-Pole Temp. Difference	Hadley Cell	Rainfall Patterns	Ice-Albedo Feedback Sensitivity
Redder Star	Increases	Decreases	Weakens	Global Decrease	Decreases
Increased Insolation	Increases	Increases	Strengthens	Global Increase	Increases
Longer Year		Increases	Widens and Strengthens
Longer Day	Increases to 4 days, decreases thereafter	Decreases	Widens	Shifts to or away from poles depending on obliquity	Increases
Increased Obliquity	Increases	Decreases to 50°, increases and inverts thereafter	Weakens, Possibly inverts direction	Possible seasonal inversion (heavy polar storms)	Increases to 35°, Decreases thereafter.
Increased Eccentricity	Increases	Decreases		Can widen or shrink tropic rains in each hemisphere	Decreases
Increased Size	Decreases	Increases	Shrinks (in latitude) and weakens.	Global Decrease	Decreases
Increased Outgassing/Decreased Weathering	Increases	Decreases	Widens until ice caps melt, shrinks thereafter	Increased contrast (wetter tropics, drier deserts)
Increased Land Area	Decreases	Increases		Global Decrease	Decreases
More Polar /Less Equatorial Land Area	Decreases	Increases (especially with polar ice caps)
Increased Ocean Salinity	Increases	Decreases			Possibly counteracts
Increased Atmo. Pressure	Increase to at least 4 bar, possible decrease thereafter	Decreases	Shrinks and weakens		Decreases
Increased Oxygen	Decreases			Global Decrease
Increased Plant-Like Flora	Decreases			Increase over land, particularly arid regions	Decreases, Counteracts with aerosol feedbacks

• Sea level change is caused by various cycles of water volume and basin volume shifts, of varying period and amplitude: 
• Supercontinent Cycle and Climate State: 100s of meters, 100s of millions of years.
• Lakes and Groundwater: 10s of meters, millions of years.
• Milankovitch Cycles: 10s to 100s of meters, 10s to 100s of thousands of years.
• Temperature shifts: 1s of meters, 10s to thousands of years.
• Aside from these cycles, sea level is likely to peak when planets are ~1.5 billion years old and gradually decline thereafter.

Notes

You’ll sometimes hear people describe the Coriolis effect in terms of a “Coriolis force”. Like centrifugal force, this is a fictitious force that only appears in rotating, non-inertial reference frames; but within that context it’s perfectly valid to describe as a force.

There are multiple possible etymologies for "horse latitudes" from the Age of Sail, all a bit bizarre:

• A ship with too little wind to sail but that could still make good progress by currents was said to be "horsed", and this often happened at the horse latitudes.
• Spanish ships transporting horses to the colonies would often be trapped at these latitudes with no wind and run short of water; unable to sustain them, the crew would throw the horses overboard.
• The last one is particularly convoluted: Horses were regarded as a symbol of hard work, and paying for such work in advance was considered a good way to guarantee it got done; withholding payment was analogized to expecting work from a dead horse (this may also be where "beating a dead horse" comes from). Now, sailors were often paid their first month's wages in advance while in port, where they quickly spent it all; once underway, they would then have to work the first month without wages to pay off the debt, and so they would say their "horse"—symbolizing their motivation to work—was dead (this doesn't quite seem to line up with the above definition of "dead horse", but whatever). Once the first month passed—which, for ships outbound from Europe, was often near the horse latitudes—wages resumed, and the sailors would hold a ceremony parading a straw horse around the ship before throwing it overboard.

I’m glad Hoffman et al. (2017) got 8 sources to support their statement “The Snowball Atmosphere is cold.”

I just want to say that Hasegawa et al. (2012) is a really nice example of good geology relying on a variety of types of evidence and clearly presenting the hypothesis, data, and link between the two.

Buy me a cup of tea (on Patreon)

Part VIb