You could grow food on Mars, but only indoors and only after cleaning the soil. Martian soil carries about 0.4 to 0.6% perchlorate, a salt that blocks the thyroid and stops seeds germinating, so it must be washed or eaten away by bacteria first. Sunlight is 43% of Earth's, the air is 95% carbon dioxide at under 1% of our pressure, and by my estimate each person would need roughly 50 to 60 m² of potatoes under a pressurised roof to live on them.
- 0.4 to 0.6% perchlorate by mass was measured in soil at NASA's Phoenix landing site in 2008.
- 586 W/m²: sunlight at Mars's distance, against 1,361 W/m² at Earth's.
- About 60 m² of potatoes per person on filtered Martian sunlight, by my estimate; about 19 m² under ideal lamps.
- Over 90% of perchlorate removed from a Mars simulant in 72 days by a microbial community, in a 2026 study.
- About 58 kg of oxygen, roughly 70 days of breathing, locked in the perchlorate of one person's farm soil.
This tier-2 what-if is the sequel to cooking dinner on Mars. A Mars mission lasts two to three years with no resupply, and NASA's own risk record says no packaged food system yet keeps for the five years needed. Fresh food grown on site is the obvious answer. The question is how hard it is.
Can plants grow in Martian soil?
Plants can germinate and grow in Mars soil simulants, but real Martian soil adds a problem simulants usually leave out: perchlorate. The first large controlled test, in the Netherlands in 2014, grew 14 species in simulants for 50 days.
The team sowed wild plants and crops in NASA-made Mars and Moon regolith simulants and in a nutrient-poor Dutch river soil, with no added nutrients, for 50 days. Plants grew far better on the Mars simulant than on the Moon simulant, and slightly better than on the river soil. Tomato, wheat and cress did well; cress and field mustard set seed. The limit, which the authors flag: the simulant contained no perchlorate, and its water-holding and physical properties may not match real Martian soil.
Read the paper (free)The perchlorate discovery came from NASA's Phoenix lander in 2008, near Mars's north pole. Its wet-chemistry lab mixed scoops of soil with water and analysed what dissolved.

Phoenix's Wet Chemistry Laboratory dissolved soil from the Vastitas Borealis plains in water. The solutions held about 10 mM of salts, with 0.4 to 0.6% perchlorate by mass leached from each sample. Magnesium and sodium dominated the cations, the pH was a mildly alkaline 7.7 ± 0.5, and surface soil matched soil from the ice table about 5 cm down. The limit: one landing site at high latitude; perchlorate levels elsewhere on Mars may differ.
Read the abstractHalf a percent sounds small. It is not. Perchlorate stops seeds germinating at similar levels in simulant tests, and for humans the issue is the thyroid. Here is why simply rinsing it out by dilution fails.
Worked solution: can you just wash the perchlorate out?
- Perchlorate in a kilogram of soil at 0.5%:
\[ m = 1000\ \text{g} \times 0.005 = 5\ \text{g} = 5 \times 10^{6}\ \mu\text{g} \]
- Water needed to dilute it to the EPA's proposed drinking-water goal of 20 µg/L:
\[ V = \frac{5 \times 10^{6}\ \mu\text{g}}{20\ \mu\text{g/L}} = 250{,}000\ \text{L per kg of soil} \]
That is a quarter of a million litres of water for every kilogram of soil, on a planet where water must be mined from ice. So washing only moves the problem into a toxic brine. The perchlorate has to be destroyed, not diluted. Perchlorate salts are very soluble, so washing still helps as a first step, but the wash water then needs treatment.
How do you get rid of perchlorate?
With bacteria. Some microbes "breathe" perchlorate instead of oxygen, splitting it into harmless chloride and oxygen. A 2026 study showed a natural microbial community stripping more than 90% of it from a Mars soil simulant.
The authors spiked a Mars simulant to about 22,500 mg/kg (2.25%) perchlorate and inoculated it with a perchlorate-reducing microbial community from hypersaline sediments in Utah's Pilot Valley Basin. Over 72 days at 22°C, perchlorate fell to about 7.8% of the starting level in both aerobic and anaerobic trays, roughly 1,760 mg/kg. Aerobic trays also gained plant-available phosphorus, and a small carbon and nitrogen pool formed. The limit: final levels were still slightly above the EPA industrial soil screening level, the community's make-up was not disclosed, and no plants had yet been grown in it.
Read the paperWorked solution: the oxygen bonus
The reaction the microbes run is, in net terms,
- Oxygen per gram of perchlorate: 64 g of O₂ per 99.45 g of ClO₄⁻, a ratio of 0.644.
- Soil in one person's farm (from the area calculation below): 60 m² × 0.2 m deep, at an assumed bulk density of 1,500 kg/m³:
\[ 60 \times 0.2 \times 1500 = 18{,}000\ \text{kg}, \qquad 18{,}000 \times 0.005 = 90\ \text{kg ClO}_4^- \]
- Oxygen released:
\[ 90 \times 0.644 \approx 58\ \text{kg O}_2 \]
- In breathing days, at NASA's planning figure of 830 g of oxygen per person per day:
\[ \frac{58{,}000}{830} \approx 70\ \text{days} \]
In practice much of that oxygen would be used by the microbes themselves in aerobic treatment, so treat 70 days as an upper bound. But the poison in Mars soil is also a stored oxygen supply.
In 2016 and early 2017, scientists at the International Potato Center (CIP) in Lima, Peru, tried to grow potatoes in Mars-like conditions. First they used dry, salty soil from the Pampas de La Joya desert in southern Peru as a Martian stand-in; potatoes grew only once Earth soil and fertiliser were mixed in, because the desert soil lacked the structure and nutrients for tubers to form. Then engineers built a sealed CubeSat, a Mars in a box, that set Martian temperature, pressure and oxygen and carbon dioxide levels. In March 2017 CIP reported positive preliminary results, with a salt-tolerant variety already grown in Bangladesh standing out.
Discover: Good news, it looks like we can grow potatoes on Mars (March 2017)Why potatoes for a Mars farm?
Potatoes are a favourite of space-farming research because they turn a large share of the plant into food. Raymond Wheeler's NASA team listed them among crops with a high harvest index, the edible fraction of total biomass, in the 60 to 80% range alongside beet and sweet potato, far above wheat or soybean.
A NASA-funded study of potato production for a lunar base set out the numbers I use below:
- Up to 40 g of tuber dry matter per m² per day in ideal controlled environments, about seven times the average US field yield.
- 3.73 kcal per gram of tuber dry weight.
- 2,800 kcal per person per day as the design requirement.
- Tubers keep well, are rich in carbohydrate with adequate protein, and can be cooked many ways.
Potatoes also have an Indian history of arriving from the Americas, much as chillies did (see how potatoes and tomatoes reached India).

How much farm does one person need on Mars?
About 50 to 60 m² of potatoes per person on filtered Martian sunlight, by my estimate, which matches NASA's rule of thumb of about 50 m² per person for an autonomous colony. Under ideal electric lighting it could shrink to about 19 m².
| Growing factor | Earth | Mars | What it means for a farm |
|---|---|---|---|
| Sunlight (top of atmosphere) | 1,361 W/m² | 586 W/m² | 43% of Earth's light |
| Day length | 24.0 h | 24.66 h | Plants keep a familiar rhythm |
| Surface pressure | 1,013 mb | 6.36 mb | A pressurised greenhouse is required |
| Air | 0.04% CO₂ | 95.1% CO₂ | Plenty of carbon, once pressurised |
| Mean temperature | about 15°C | −59°C | Heated, insulated growing space |
| Soil perchlorate | Trace | 0.4–0.6% (Phoenix) | Treat before planting |
Worked solution: potato area per person on Mars sunlight
Assumptions (mine unless sourced): equatorial site at equinox; light averaged over a full sol; 45% of sunlight energy is photosynthetically active (PAR) and each joule of PAR carries about 4.57 µmol of photons, standard conversions for sunlight; dust and the greenhouse roof cut 30%; radiation use efficiency 0.75 g of biomass per mol of photons (the value NASA's Wheeler uses); harvest index 0.7.
- Average sunlight over a sol. A flat surface at the equator gets, on average, \(1/\pi\) of the noon flux:
\[ \bar{S} = \frac{586}{\pi} \approx 187\ \text{W/m}^2 \]
- Daily light integral (DLI). Over a sol of 88,775 s:
\[ \text{DLI} = 187 \times 0.45 \times 4.57 \times 88{,}775 \times 10^{-6} \approx 34\ \text{mol m}^{-2}\,\text{d}^{-1} \]After a 30% loss to dust and glazing, about 24 mol m⁻² d⁻¹.
- Edible yield.
\[ Y = 24 \times 0.75 \times 0.7 \approx 12.5\ \text{g m}^{-2}\,\text{d}^{-1} \]
- Daily need in tuber dry matter:
\[ \frac{2800\ \text{kcal}}{3.73\ \text{kcal/g}} \approx 751\ \text{g/d} \]
- Area.
\[ A = \frac{751}{12.5} \approx 60\ \text{m}^2 \text{ per person} \]
Sensitivity check. At the best documented radiation use efficiency of 1.0 g/mol, the area falls to about 45 m². At the record 40 g m⁻² d⁻¹ under strong lamps, cool temperatures and high CO₂, it falls to 19 m². In a global dust storm, which can dim the sun for weeks, natural light could fall several-fold and the farm would need lamps or stored food. A crew of four would need a greenhouse roughly the size of a badminton court, about 240 m², before paths and equipment.
See the data as a table
| Area per person (m²) | |
|---|---|
| Ideal lamps, record yield (40 g/m²/d) | 19 |
| Mars sunlight, RUE 1.0 | 45 |
| Mars sunlight, RUE 0.75 (base case) | 60 |
| NASA rule of thumb, autonomous colony | 50 |
Source: This post's calculation using Wheeler et al. (potato productivity, 3.73 kcal/g, 2,800 kcal/d) and Wheeler, NASA KSC (RUE, about 50 m² per person) · Estimates for a potato-only diet; a real diet would mix crops.
Use this to check what you just read. Change the light, the crop's efficiency and the calories needed.
How big is your Mars farm?
What else would a Mars farm grow?
A potato-only diet is a design exercise, not a menu. A real Mars farm would mix fast salad crops for freshness with staples for calories and pulses for protein:
- Salad and herbs (lettuce, cress, coriander, mint): fast, high harvest index, and the flavour crews miss most.
- Chillies: already grown on the ISS in 2021; a few plants transform a pantry of dried food.
- Pulses such as moong and peas: protein, and they fix nitrogen, which Martian soil lacks.
- Millets: short-season, hardy grains with good nutrition per gram (see millets ranked by nutrition).
- Potatoes and sweet potatoes: the calorie base, with high harvest indices.
On Mars the farm is not measured in hectares. It is measured in watts of light and kilograms of cleaned soil.
Growing food on Mars is possible, but not in the open and not in raw soil. The 0.4–0.6% perchlorate must be destroyed (bacteria removed over 90% in 72 days in a 2026 simulant test), because diluting it to safe levels would take about 250,000 litres of water per kilogram of soil. The crops need a pressurised, heated greenhouse, and light sets the size: about 60 m² of potatoes per person on Martian sunlight, or about 19 m² under ideal lamps. The deciding number is the light budget, and the bonus is that the cleaned-out perchlorate yields about 58 kg of oxygen per person-sized farm.
Where my model breaks down: I assumed equatorial sunlight, a fixed 30% loss and a potato-only diet, and I ignored cosmic radiation, low gravity's effect on watering, and the energy to heat and pressurise the greenhouse. Any of these could enlarge the farm. The calculator lets you test them.
What's new
- Microbes strip over 90% of perchlorate from Mars soil simulant in 72 days Gravitational and Space Research, Sep 2026
- US EPA proposes a perchlorate drinking-water rule with a 20 µg/L health goal US EPA, Jan 2026
- NASA lists bioregenerative food among the fixes for its red-rated Mars food risk NASA, Aug 2025
What if next?
- What if humans could digest grass? A Mars crew that could eat potato leaves and stems would shrink the farm.
- What does food taste like in zero gravity? On the six-month trip out, fresh herbs would matter even more.
- What if you lost your sense of smell tomorrow? What flavour is left when the nose is blocked.
Questions people ask
Can you really grow potatoes on Mars like in The Martian?
In principle, yes, but not in raw soil. Real Martian soil contains about 0.5% perchlorate, which the film The Martian skipped. The International Potato Center's 2017 Mars-in-a-box test grew potatoes under Martian air conditions, but only with Earth soil and fertiliser mixed in.
Is Martian soil toxic?
Yes, because of perchlorate, measured at 0.4–0.6% by the Phoenix lander. It blocks the thyroid's uptake of iodine in humans and harms seed germination. It can be removed by washing plus treatment, or by perchlorate-reducing bacteria.
How much space do you need to grow food for one person?
NASA's Raymond Wheeler puts it at about 50 m² per person for a self-sufficient colony. My potato-only estimate on Martian sunlight is about 60 m², shrinking to about 19 m² under ideal electric lighting.
Why not just use Earth soil?
Mass. A 20 cm deep bed of 60 m² holds about 18 tonnes of soil, and every tonne launched to Mars costs a fortune. Using local regolith, treated, is the only scalable route.
Sources
- Hecht MH et al.: Detection of perchlorate and the soluble chemistry of martian soil at the Phoenix lander site, Science 325:64–67 (2009) (0.4–0.6% perchlorate, pH 7.7)
- Wamelink GWW et al.: Can plants grow on Mars and the moon, PLOS ONE (2014) (14 species, 50 days, no added nutrients)
- Kelly L et al.: Microbial pre-conditioning of perchlorate-bearing Martian regolith simulant, Gravitational and Space Research 14(1) (2026) (over 90% removal in 72 days)
- Wheeler RM, Morrow RC, Tibbitts TW, Bula RJ: Scenarios for optimizing potato productivity in a lunar CELSS (40 g/m²/day, 3.73 kcal/g, 2,800 kcal/day)
- Wheeler RM: Development of bioregenerative life support for longer missions, NASA KSC (RUE up to 1.0 g/mol, 830 g O₂/person-day, about 50 m² per person, harvest indices)
- NASA NSSDC: Mars fact sheet (586 W/m², 6.36 mb, 95.1% CO₂, 214 K, 24.66 h day)
- Discover: Good news, it looks like we can grow potatoes on Mars (2017) (CIP experiment, Pampas de La Joya soil)
- US EPA: Perchlorate in drinking water (2026) (proposed 20 µg/L health goal; thyroid effects)
- NASA: Food and Nutrition Risk Record, Revision D (2025) (5-year shelf-life gap, bioregenerative supplementation)
