The following is a collection of individual posts recovered from the NewMars forum thread on a Minimally Terraformed Martian Atmosphere:

I'm making this (admittedly massive) post in an attempt to correct some stuff I've seen on these forums and offer some new information about the requirements for a breathable atmosphere. The overall goal is to present an upper limit to tolerable CO2 levels (jump to the end if you just want to know), but I'd like to start with two general points:

Saying that humans can't live with more than [insert safety data sheet value] mbar of CO2 because of tocixity is like saying people can't live in the Andes because of lack of O2.
Occupational Safety and Health Administration guidelines dictate that the Permissible Exposure Limit for CO2 is 0.5%, and the level Immediately Dangerous to Life and Health is 5% (both assume a total pressure of 1 atm, i.e. sea level). At these levels shortness of breath, headache, and dizziness are reported. Similar symptoms are associated with altitude sickness, which begins to occur when someone from sea level ventures up to altitudes of 1500-3500 m. Contrast this with the fact that the city of Potosi in Bolivia is located at 4000 m and has a population of over 130,000 according to the INE 2001 census. The point is that the acclimation capabilities of humans (and other organisms) should not be underestimated.
High Altitude Illness, Ivan Schatz, M.D., Western University of Health Sciences
Ansul Incorporated Carbon Dioxide Material Safety Sheet

Breathing 5% CO2 on Earth is not the same as breathing 5% CO2 on Mars.
This is a common misconception I've noticed on these forums, and it really needs to be cleared up. The truth is that respiration in organisms depends only on partial pressure of gases. Going by just percentages will get you in trouble. Let's do an example. Say that Bob has acclimated to breathing 5% CO2 on Earth at sea level. What this really means is that he has acclimated to breathing (5% CO2)*(1 atm total) = 50 mbar CO2. Now let's put him in a terraformed Mars atmosphere of 250 mbar. The amount of CO2 Bob can handle on Mars is the same as on Earth: 50 mbar CO2. But when you convert to percentages, that's (50 mbar CO2)/(250 mbar total) = 20% CO2. I hope this simple example gets the point across that taking straight gas percentages can be inaccurate when considering the breathability of a terraformed Martian atmospshere.

So, how much CO2 can we take? Here are the key effects to consider:

#1) CO2 replaces O2 on hemoglobin, thus depriving you of oxygen
#2) CO2 causes acidosis, i.e. a blood pH imbalance
#3) CO2 causes an inert gas narcosis, similar to nitrogen in scuba diving


The first effect, inhibition of O2-hemoglobin bonding, turns out to be negligible:

Only a small fraction of the CO2 in your blood is transported by hemoglobin.
The vast majority (~90%) of the CO2 in blood is transported by bicarbonates.
Source: Clinical Anesthesiology, M. J. Murray, G. E. Morgan, and M. S. Mikhail, New York: McGraw-Hill (2002)

Blood CO2 levels have only a small effect on O2 binding to hemoglobin.
To illustrate: Like elevated CO2 levels, raising body temperature negatively impacts O2-hemoglobin binding. Normal CO2 arterial pressure is 40 mmHg. Increasing this over seven times to 300 mmHg (400 mbar) has as much effect on O2-hemoglobin binding as raising core body temperature to 102.6°F, which is only a moderate fever. Fevers aren't fun, but no one dies from respiratory failure due to inadequate O2-hemoglobin bonding.
For more information, play with this applet:


The second effect, acidosis, is the major cause of the toxicity typically associated with CO2. But what seems not to be so widely known is that it can be overcome through acclimation over a few days, much like altitude sickness. As I've noted in a previous post, this is because the carbonic acid formed in the blood by the CO2 is neutralized by the kidneys retaining bicarbonates ions. Just to reiterate, humans have been adapted without significant complications to 4% CO2, Rhesus monkeys to 6% CO2, and sheep to 12% CO2.

Only in the experiment with sheep was some sort of limit of acclimation reached. The study was reported on in at least two papers from the early 70s by Hoover et al., who detailed the nutritional and physiological responses. Negative effects began around 8% CO2, but were tolerable up till 12% CO2. Eventually serious degradation occured at 16% CO2, with the sheep becoming lethargic, beginning to drool, and ceasing to eat. Subsequently, the experiment was stopped. But the question is, what caused these symptoms? There does not seem to be any obvious reason why the bicarbonate buffer of the body would fail at that level.


The answer is that the third effect, inert gas narcosis, is coming into play. It turns out that every gas has some pressure at which breathing it will cause symptoms similar to being drunk. At even higher pressures it will act as an anesthetic. These effects are thought to be due to the gas diffusing across cell membranes and interfering with neural signals. Thus, the narcotic potency of a gas would be expected to be related to its solubility in lipids and oils. This is confirmed by experiment. The graph below depicts the points at which a number of gases cause inability to feel pain in a standardized test:

Image not available

So chloroform is a good anesthetic because it is extremely oil soluble and will knock you out at only trace pressures, while helium is almost impossible to "overdose" on in terms of narcotic effects and is thus used in deep diving where you have to breathe high pressure gases. Nitrogen isn't quite as good as helium, and so its narcotic effects can be significant when diving at relatively shallow depths. For nitrogen a biological response, not significantly impairing, can occur as shallow as 60 ft of sea water, or 1.5 atm of pressure, which for breathing air would be 1.2 atm N2. The safety guideline limit is 100 ft of sea water, or about 2.4 atm N2. Deeper than this and serious impairment of judgement similar to drunkeness can occur.

Now look at where carbon dioxide is on the graph. It's much more narcotic than one would expect based on the lipid solubility theory. This increased potency is due to acidosis, the second biological effect of CO2 on our list. But we know that the effects of acidosis can be compensated for by the kidneys. This means that, given time to acclimate, the narcotic potency of CO2 will decrease, and it will move up on the graph to join the company of nitrous oxide (laughing gas) and xenon. Thus, the theory of inert gas narcosis provides us with the real limit to breathable CO2 levels.

So what exactly is this limit? According to lipid solubility theory CO2 should be 20 times more potent than N2. This suggests that biological response will begin at about (1.4 atm N2)/20 = 70 mbar CO2, or 7% CO2 at sea level, and the safety limit will be (2.4 atm N2)/20 = 120 mbar CO2, or 12% CO2 at sea level. I was quite happy when I saw this, because it agrees well with the study on sheep that I mentioned before.

Now I want to note that there's significant "wiggle room" in these calculations. For example, it is difficult to evaluate the narcotic contribution of O2 at high pressure when looking at diving, but some people think that it's similar to N2. If we accept that and redo the calculations, we find a biological reaction threshold of about 90 mbar CO2 (9% CO2) and a safe limit of 150 mbar (15% CO2). These are a bit different from the numbers we got assuming just N2 was narcotic, but they're still consistent with the sheep experiment. Also, while it's true that sheep and humans have similar inert gas narcosis reactions, it turns out that humans are 8% more "durable", i.e. the sheep study pressures have to be bumped up 8% to get a correct pressure/response correlation for humans. Finally, and perhaps more intriguing, it seems that there may actually be the possibility of acclimation to inert gas narcosis. Some frequent divers are capable of operating safely at depths down to 180-220 ft, which, depending on whether you try to include O2 narcosis or not, is equivalent to between 220-340 mbar CO2, or 22-34% CO2 at sea level. This is somewhat speculative though.

Ultimately, we can state with confidence a conservative lower estimate of 120 mbar for the limit of safe CO2 partial pressure. This raises the possibility that CO2 might actually be used as a significant buffer gas in terraformation.

RickSmith: So to be clear, you are suggesting that a mixture of gases like:

CO2: ....... 120 mbar.
O2: ......... 200 mbar.
N2 .............. 5 mbar. (Plants can now fix nitrogen.)
Ar ............... 1 mbar.

could be breathed in both greenhouses and the main residential areas?

Sure, Rick, that'd work just fine. In fact, at those levels of ambient O2 and CO2 your blood O2 pressure would be almost exactly the same as at sealevel on Earth, ~100 mm Hg. Let's walk through the math, converting from "mbar" to "mm Hg" to make comparison with medical data easier:

Ptot = 200 + 120 + 5 + 1 = 326 mbar = 245 mm Hg
FO2 = 200 / 326 = 0.61
FCO2 = 120 / 326 = 0.37

Now that we have the total atmosphere and fractional components, we can find the inspired pressures by accounting for water vapor in the lungs (this is in mm Hg now):

PIO2 = 0.61*(245 - 47) = 121 mm Hg
PICO2 = 0.37*(245 - 47) = 73 mm Hg

Ok, here's where a little research was required. Turns out there's a linear relationship between PIO2, the inspired O2 pressure, and PaO2, the arterial O2 pressure. Using data from (Holstrom 1971), which can be found half-way down the page here, converting to inspired values as we did above, and fitting to the points PIO2 > 64 mm Hg, i.e. altitude < 20,000 ft (because the body begins to reach its limit there and the linear relationship fails), we obtain:

(Eq 1)
PaO2 = 0.84 * ( PIO2 ) - 23 mm Hg

This great little equation works down to PIO2 = 64 mm Hg with <10% error. Caution! It only works for normal CO2 levels.

By culling data from many of the references in my last post, I obtained a very similar equation for adult rats. The PaO2 values it produces agree well with the human equation. In fact, the difference is <10% for values of PIO2 < 100 mm Hg, roughly equivalent to breathing conditions found at altitudes above 10,000 ft on Earth. Since we're interested in minimal terraformation (i.e. low O2 levels), that suggests that rat respiration is a very good human analog for our purposes.

So, what about high CO2 levels? Well, CO2 stimulates respiration in animals, which means you transport more O2 from the air to your blood. Since rats seem to be a good model, I again culled data for adult rats from that list of references and found the relationship between PICO2 and PaO2 for low O2 levels to be:

(Eq 2)
PaO2 = 0.35 * ( PICO2 ) + C

where 'C' is the PaO2 one would obtain for normal CO2 levels, i.e. (Eq 1). So, substituting that in we obtain:

(Eq 3)
PaO2 = 0.84 * ( PIO2 ) + 0.35 * ( PICO2 ) - 23 mm Hg

This is the culmination of a lot of research! It allows you take inspired O2 and CO2 levels in units of "mm Hg" and determine what one's blood O2 pressure would be.

OK, now that we have the tools, back to Rick's example. Plugging PIO2 = 121 mm Hg and PICO2 = 73 mm Hg into (Eq 3) obtain PaO2 = 104 mm Hg, which is practically the same as the 103 mm Hg recorded for sea-level on Earth by (Holstrom 1971). I think that's some serious serendipity on Rick's part!


Holstrom, F. M. G. Hypoxia. In: Aerospace Medicine edited by H. W. Randel. Baltimore: Williams & Wilkins Co. (2nd Ed.) 1971, pp. 56-85.

In order to colonize a terraformed planet it is not enough for adult organisms to function, they must also be able to produce healthy offspring. This fact, combined with the knowledge that newborn animals generally have lower tolerance thresholds than their mature counterparts, has been bugging me on the issue of atmospheric breathability for a while now.

Luckily, the use of therapeutically elevated CO2 on infants for lung protection is currently of interest and quite a few studies have focused on collecting data from neonatal rats and mice, and even human infants. Here are summaries of some of those studies:

Investigating the prenatal and neonatal effects of CO2, (Dean et al. 2001) exposed pregnant mother rats for 1 week pre and post delivery. They found that 7% CO2 slightly slowed the early growth of newborns, but that they caught up with control about 2 weeks after birth. Those exposed to 10% CO2, however, experienced more severe growth retardation and did not seem to recover; much of the litter was sacrificed at birth.

Studying neonatal rats, (Rezzonico & Mortola 1989) found no significant difference in weight between control and subjects exposed to 7% CO2 during the first week of life.

(Kantores et al. 2006) examined the combined effects of hypoxia and hypercapnia on neonatal rats, finding that elevated CO2 levels helped relieve many of the effects of breathing low O2. The most extreme gas mixture used, 10.5% O2, 10% CO2, caused excessive mortality. However, mixtures of 13% O2 and between 5.5-10% CO2 allowed newborns to maintain normal (i.e. 21% O2) blood O2 pressure and red blood cell levels.

In an extremely interesting study by (Li et al. 2006) the gene expression in neonatal mice exposed to 8% CO2 and 12% CO2 for 2 weeks was compared with control to see if there were any differences. The 8% CO2 mice experienced upregulation of 365 genes and the downregulation of 342 out of a total genome of 24,878. These alterations acted to improve lung function and offered increased protection to extreme oxygen levels. In contrast 12% CO2 cause a significant expression change in only 2 genes. The authors point out that at high enough levels CO2 acts as a depressant on biological function, and invoke this as an explaination.

Another fascinating experiment was done by (Gu et al. 2007) on mice beginning at 2 days old which sought to examine the effects of elevated CO2 on neuron activity in the hippocampus. They found that after 2 weeks of exposure, the mice on 8% CO2 has nearly identical neuron operation with control, while neuron excitability in mice on 12% CO2 was significantly inhibited.

In study of premature human infants by (Mariani et al. 1999) it was demonstrated that newborns on assisted respiration required significantly less time on the respirator when given a hypercapnic air mixture than when not (2.5 days vs 9.5 days, on average). Blood CO2 pressure was raised by ~10 mm Hg, so I'd guess they used ~5% CO2.

These results are fairly consistent with the previous estimates of CO2 toxicity presented in this thread, and indicate that for optimal newborn growth a biological threshold of 50-64 mm Hg inspired CO2 (7-9% CO2 at 1 atm) should be followed rather than the adult tolerance threshold of 86 mm Hg inspired CO2 (12% CO2 @ 1 atm).

In terms of O2 requirements, (Xu & LaManna 2006) identify PaO2 ~ 50 mm Hg as the limit of chronic mild hypoxia, i.e. the point below which degradation of biological functions begins over long periods. They also suggest PaO2 ~ 35 mm Hg as the limit of consciousness for long term exposure. These values agree with previous estimates in this thread as well as this one.

So using the theory of "normal adult function threshold" as the safe limit for infants, should we take PaO2 = 50 mm Hg as an acceptable limit? Well, the city of Potosi says "Yes!" Having over 130,000 people, and existing since 1546 A.D. at an elevation 4000 m, adults and infants alike in Potosi inspire O2 at PIO2 ~ 90 mm Hg, which corresponds to a PaO2 ~ 50 mm Hg. While (de Meer et al. 1995) do find a moderate reduction of children's stature at such an altitude, in part attributible to the mild hypoxia, severe growth retardation is not observed. If you want to read more about the adaptation of animals, including humans, to high altitudes (Monge and Leon-Velarde 1991) is a wealth of information.

My gut says that the respiration stimulation by CO2 should not be relied on when thinking of how much O2 is in the blood. I think the effect of CO2 should be used as relief for hypoxia, not a tool that allows us to drive the atmospheric O2 lower. The fact that all the rat newborns on 10.5% O2, 10% CO2 in (Kantores et al. 2006) died supports this. Though one might have expected sufficent PaO2 from (Eq 3) in my previous post, the actual PaO2 was much lower, suggesting some limit had been reached and the linear relationships involved failed. I suspect the hypoxia was too much and the effect of CO2 couldn't cope.

My suggestion is that an inspired O2 of PIO2 = 90 mm Hg (13% O2 @ 1 atm) be used as the minimum limit for oxygen.

A simple atmosphere of just O2 & CO2 at PIO2 = 90 mm Hg and PICO2 = 57 mm Hg (equivalent to breathing 13% O2, 8% CO2 @ 1 atm), would look like this:

O2:.......160 mbar
CO2:.....100 mbar

You can add any inert gas (N2, H2O, Ar, etc.) within reason and this will still be breathable. In fact, its breathability will be slightly enhanced by adding other gases due to the increase in total pressure.

From (Eq 3) in my previous post, one would expect to have a blood O2 pressure circa 68 mm Hg, which is comparable to the oxygenation experienced at ~9000 ft on earth. Substantial inhabited portions of Colorado are above this elevation.


Developmental changes of the response of in vivo ventilation to acute and chronic hypercapnia in rats
JB Dean, PB Douglas, JA Filosa, AJ Garcia, RW Putnam and CE Stunden, Respiratory Research 2001, 2(Suppl 1):P13

Respiratory adaptation to chronic hypercapnia in newborn rats
R Rezzonico & JP Mortola, J. Appl. Physiol. 67(1): 311-315, 1989

Therapeutic hypercapnia prevents chronic hypoxia-induced pulmonary hypertension in the newborn rat
C Kantores, PJ McNamara, L Teixeira, D Engelberts, P Murthy, BP Kavanagh, and RP Jankov, Am J Physiol Lung Cell Mol Physiol 291: L912–L922, 2006.

Effect of carbon dioxide on neonatal mouse lung: a genomic approach
G Li, D Zhou, AG Vicencio, J Ryu, J Xue, A Kanaan, O Gavrialov, and GG Haddad, J Appl Physiol 101: 1556–1564, 2006.

Chronic High-Inspired CO2 Decreases Excitability of Mouse Hippocampal Neurons
XQ Gu, A Kanaan, H Yao, GG Haddad, J Neurophysiol 97: 1833–1838, 2007.

Randomized trial of permissive hypercapnia in preterm infants.
G Mariani, J Cifuentes, and W A Carlo, Pediatrics. 1999 Nov; 104 (5 Pt 1):1082-8

Chronic hypoxia and the cerebral circulation
K Xu & J C LaManna, J Appl Physiol 100: 725-730, 2006

Physical adaptation of children to life at high altitude
K. de Meer, HSA Heymans, and WG Zijlstra, Euro J of Ped, V 154; 4, 1995

Physiological Adaptation to High Altitude: Oxygen Transport in Mammals and Birds
C Monge & F Leon-Velarde, Physiological Reviews, V 71; 4, October 1991

SpaceNut: Plants cannot survive in a partial pressure CO2 level greater than 0.2 kPa.

That's a lot lower than the sources I've seen:

(Tikhomirov et al. 2007) grew radishes, beets, carrots, and cabbage at 0.7-0.9 kPa CO2 and found significant enhancement in edible biomass (~20%) over the previously assumed optimal range of 0.15-0.3 kPa CO2.

(Chagvardieff et al. 1997) grew wheat, tomatoes, potatoes, and peas at 0.37 kPa CO2 from seedlings to harvest. They did find that wheat suffered at these levels, but response was extremely variable between species: one variety suffered a 50% reduction in edible mass while another suffered only a 9% drop. The main thrust of the paper, however, was the tremendous benefits for all the other vegetation grown: tomatoes increased edible output by 46-59%, potatoes a modest 6-16%, and peas an incredible 245%. A further experiment raising potatoes in 2 kPa CO2 found the only effect to be that the tubers grew much faster.

(Wheeler et al. 1994) found that while 1 kPa CO2 was not optimal for the growth of soybeans and potatoes, neither was it injurious: the plants were in all respects either comparable to or far more productive than those grown in normal conditions.

(Wheeler et al. 1997) successfully grew tomatoes at 1 kPa CO2 and found that there was no significant effect on mineral and nutritional composition of the fruit.

(Grotenhuis et al. 1997) studied wheat and (Bugbee et al. 1994) studied wheat and rice at up to 1 kPa CO2. Like many other studies, they both found that while not optimal, such CO2 levels were not lethal or even seriously detrimental to the plants. It should be noted that a drop in seed productivity associated with CO2 interference with ethylene was observed. This is a major problem with high CO2 among plants. CO2 stimulates ethylene production at a few kPa, but then begins inhibiting it above ~5 kPa. While plants can still survive at these levels, ethylene is an important plant hormone, and modification of reproductivity is experienced.

I would also like to reference the series of papers by Kidd in the first half of the 20th century. He remarks upon many studies that found pea growth to be stimulated by a few percent CO2, and then inhibited by higher levels. The point at which inhibition drew growth below normal levels was consistently 7% CO2 (7 kPa, 70 mbar). Kidd himself also did extensive research into the effects of very high CO2 on germination and respiration in a wide variety of plants. He found that sensitivity in germination varied significantly between species, with white mustard being totally intolerable of >18% CO2, while peas were still capable of germinating at 100% CO2, albeit at greatly reduced rates.

But what takes the cake in my experience is the study of (Pfanz et al. 2007) on timothy grass around mofettes, i.e. natural CO2 springs. The researchers found plants growing with air in the soil being 26% CO2; that's 26 kPa or 260 mbar. Described in the paper as living "a life on the verge of death" at 800 times normal CO2 levels as well as permanent hypoxia and even transient anoxia in the soil, these plants were severely stunted. They also had nitrogen, phosphorous, zinc, and sulfur nutrition absorption deficiencies. But they were alive, persisting, and reproducing in a natural microecosystem. In a similar study with reeds, Pfanz et al. found that photosynthetic inhibition occured at 20-98% CO2 (20-98 kPa) in both mofette and control plants, but that the mofette acclimated plants were far more robust. They were even able to maintain 20% normal photosynthetic electron flow for several hours at 98% CO2; by comparison the control plants were totally shut-down.

Effect of Increased CO2 Concentrations on Gas Exchange and Productivity of Cultivated Vegetables Contributing to the Phototrophic Component of Biological Regeneration Life-Support Systems (Tikhomirov et al. 2007)
Effects of Modified Atmosphere on Crop Productivity and Mineral Content (Chagvardieff et al. 1997)
Growth of Soybean and Potato at High CO2 Partial Pressures (Wheeler et al. 1994)
Effect of Elevated Carbon Dioxide on Nutritional Quality of Tomato (Wheeler et al. 1997)
Photosynthetic performance of timothy grass is affected by elevated CO2 in post-volcanic mofette areas (Pfanz et al. 2007)
Physiological Reactions Of Reed Growing Under Co2 Extremes Within a Co2 Emitting Mofette Field (Pfanz et al.)