You Are Not Smelling Rain

What you're smelling is drought. What NASA is looking for on Mars is the same thing.

I spent a week in Kauai last month, and it rained almost every day. Not the way it rains in Los Angeles — the halfhearted winter apology that everyone pretends is weather — but real tropical rain, heavy and immediate, the kind that falls from clouds that were never far away to begin with.

It didn’t smell like anything.

Not nothing exactly. But not what I grew up with in Arizona — the smell that arrives before the storm front does and stops you mid-sentence. That smell has a name, petrichor, that feels like it should be much older than 1964. In Kauai, the rain was just water falling. I noticed its absence the way you notice silence after a sound you didn’t realize you were used to.

I spent the rest of the trip thinking about what was missing, and the answer turned out to be the desert, and of course, minerals.

What Bear and Thomas Named

In 1964, two Australian researchers — Isabel Joy Bear and Richard Thomas — published a paper in Nature¹ describing the distinctive smell that arises when rain falls on dry earth after a long dry spell. They named it petrichor, from the Greek petra (stone) and ichor (the fluid that runs in the veins of gods). The word is sixty years old. The feeling it names is not.

Bear and Thomas were precise about something that has largely been forgotten in six decades of popular writing about their discovery: the smell doesn’t come from rain. It comes from how rain affects a specific mineral.

The mineral is montmorillonite.

A transmission electron micrograph of montmorillonite at high magnification ² — the stacked, sheet-like interlayer architecture that creates the enormous surface area. A single gram of this mineral, unfolded, covers most of a tennis court. The smell of rain begins here. Source options: DOI: 10.1088/2053-1591/ad62c2

Montmorillonite is a smectite clay — its atomic structure is built in repeating sandwiches: two silica sheets bonded to one aluminum sheet, stacked layer upon layer at a scale far smaller than anything visible. That “2:1” architecture is what gives it an internal surface area of roughly 700 to 800 square meters per gram. The number is harder to picture than to state. A single gram of this clay, if you could unfold all its interlayer surfaces, would cover most of a tennis court.

During dry periods, those surfaces act as an enormous chemical sponge, adsorbing — sticking compounds to their surfaces, the way dust clings to a screen, rather than soaking them in — volatile organic compounds: plant terpenes (the aromatic molecules responsible for the smells of pine, sage, and citrus), fatty acids, and the waxy residues of everything that grew and dried in the surrounding landscape. The clay holds them through weak molecular attraction — van der Waals forces, the same physics that let a gecko climb glass — and through the slight electrical charge that clay surfaces carry, which grips certain molecules the way a magnet grips iron filings.

Bear and Thomas tested multiple rock and soil types. Montmorillonite produced the strongest petrichor by a significant margin. Catching the compounds is only part of it. The mineral surface acts as a catalyst — it speeds up chemical reactions without being consumed by them — oxidizing odorless fatty acids into smaller aromatic molecules called aldehydes and ketones, the classes of compounds responsible for smells like vanilla, fresh bread, and cut grass. These are the waxy, stone-like notes in what you smell.

When humidity rises just before rainfall, water molecules compete for the same mineral-binding sites. Water wins. The organic compounds are displaced and lifted into the air as fine aerosol particles. What you smell is the montmorillonite letting go — releasing what it had been holding since the last storm. Not rain arriving. Something older.

The second component is geosmin.

Geosmin

Geosmin is a sesquiterpene — a naturally occurring aromatic compound in the same chemical family as the scents of sandalwood, ginger, and cedarwood, built from 12 carbon atoms — with the molecular formula C₁₂H₂₂O. The human nose detects it at concentrations as low as five parts per trillion. Most analytical chemistry equipment can’t find what your nose handles without trying.

Streptomyces are filamentous bacteria — they grow in long branching threads rather than as individual cells — and they live in soil in extraordinary abundance. During wet periods, they actively colonize the microscale pore spaces between clay particles, where montmorillonite’s water retention capacity exists — its ability to hold water and release nutrients slowly, acting as a slow-release fertilizer — and concentrate the resources the bacteria need. When conditions dry, they form desiccation-resistant spores — dormant, protective capsules that can survive complete drying for months or years — not dead exactly, but suspended on the mineral surface, waiting for something to change. This is a seasonal system. Months, occasionally a few years. The geosmin they produce adsorbs onto the surrounding clay surfaces in the meantime, held by the same mineral-binding mechanism that holds the plant oils, until rain arrives and the clay releases both at once — bacteria exhaling for the first time in months, the mineral letting go of what it stored.

The smell was never for you. It's a 450-million-year-old call-and-response between a bacterium and a soil arthropod, aerosolized by rain, intercepted by a human nose that evolved sensitivity to it for reasons we still don't fully understand.³ The bacteria broadcast a signal to springtails — arthropods to eat dead biomass — you receive it; it stops you mid-sentence.

This is what petrichor is, in full: a mineral releasing two things simultaneously — the chemistry it trapped from the landscape and the chemistry produced by the organisms it sheltered. The smell isn’t separable from the clay, nor from the bacteria the clay houses. The mineral isn’t just a container here.

This is also why Kauai doesn’t smell the same — there are two reasons, not one.

Kauai’s soils are not montmorillonite. They are what you get when basalt — volcanic rock — weathers completely over millions of years in a wet tropical climate: gibbsite (an aluminum mineral), goethite (an iron mineral — what gives Hawaiian soil its deep red-orange color), and halloysite, a different type of clay altogether with far less surface area and adsorption capacity than smectite. These minerals simply can’t trap and hold plant chemistry the way montmorillonite does. And those soils are already saturated — the mineral trap was never set because in a rainforest receiving over a meter of rainfall each year, the montmorillonite that would write the record isn’t there, and the drought that would charge it never happens. The Streptomyces never cycle through dormancy. Two absences, same result. What you call the smell of rain is, more precisely, the smell of drought ending in the right kind of rock — and if there was no drought, and the rock is wrong, there is nothing to smell.

Inside the Rock

The story gets stranger in the Atacama.

In the hyperarid core of the Atacama Desert in northern Chile — one of the driest places on Earth, where rainfall in some areas occurs once every twenty to fifty years — there are no plants. Very little soil in any biological sense. The mechanisms Bear and Thomas described don’t apply: no plant-derived terpenes accumulating on mineral surfaces, no seasonal Streptomyces dormancy cycle to speak of — not really an ecosystem in any form the word usually implies.

And yet.

My own research has included work on halophilic microorganisms — salt-loving organisms that don’t merely tolerate high salt concentrations but require them — in Atacama salts. The halite nodules that cover large sections of the desert floor are not dead rock. Break one open and the interior is green — a biofilm of cyanobacteria (bacteria that photosynthesize, using sunlight to make energy the way plants do), dominated by a genus called Halothece, that has found a way to survive inside a salt crystal.

It has been alive, inside a crystal, without rain, possibly longer than anything else in this landscape. 4,5

The reason is mineralogical. Halite — NaCl, table salt — is hygroscopic in a specific and consequential way: it undergoes deliquescence at 75% relative humidity. Hygroscopic means it readily absorbs water from the air; deliquescence means it takes this further — when the air reaches that humidity threshold, halite doesn’t just get damp. It actively pulls water vapor out of the atmosphere until it has dissolved itself into liquid brine within its internal pore network. The crystal creates its own water supply from air that appears bone-dry. Simultaneously, halite’s translucency — its partial transparency, like frosted glass — reduces incoming UV radiation to levels where photosynthesis is possible without the DNA damage that kills exposed cells, while the few millimeters of salt above the community insulate it from surface temperature extremes — sometimes above 70°C — that kill everything outside. Gypsum creates similar conditions through different mineralogical properties. Quartz does neither — it doesn’t deliquesce and can’t maintain internal humidity. Hypolithic communities shelter under quartz pebbles, using the mineral as a sunshade but remaining on the surface beneath it rather than inside it. The endolithic lifestyle — living within the mineral itself — depends on the specific chemistry of evaporite minerals: minerals formed when ancient bodies of water evaporated and left their dissolved salts behind, the way a ring of mineral crust forms when a glass of water dries on a wooden table, but over geological time and at massive scale.

The salt isn’t keeping life out. It’s what life is built on, and the specific mineralogy is the reason that’s possible at all.

Actinobacteria — the large bacterial phylum, or major biological division, that contains Streptomyces and from which many of our antibiotics were originally discovered — are confirmed heterotrophic members of these endolithic communities. Heterotrophic means they get energy by consuming organic matter, as opposed to making their own through photosynthesis. They live alongside the photosynthetic Halothece, eating what the cyanobacteria produce. What has not yet been directly measured is whether these organisms produce geosmin when the Atacama receives rainfall. A 2025 review in FEMS Microbiology Reviews on microbial volatile organic compounds — the airborne chemicals that microorganisms release as metabolic byproducts — in extreme environments identifies exactly this as an open research gap, noting that production of these compounds by endolithic communities is “likely, as MVOC-mediated interactions have been observed in related species.” The measurement hasn’t been done. The organisms are inside the rock, and the rain, when it comes, comes every few decades.

But the timescale of their survival is probably not decades — the mechanism that extends it is a property of the mineral itself.

As water evaporates from a brine, the solution concentrates toward saturation — and then beyond it, into supersaturation, holding more dissolved salt than it normally could at equilibrium. The evaporation rate slows dramatically at this point — the brine resists its own disappearance. Bacteria live on and within supersaturated brines for extended periods during this process, far longer than a simple drying timeline would suggest. Then, as halite begins to crystallize from the supersaturated solution, something happens that has no equivalent in any other mineral environment: the growing crystal traps microscopic pockets of the original brine as fluid inclusions — sealed droplets of ancient liquid, locked inside the crystal structure as it formed. The organism that was living in that brine is now inside the mineral — not sheltering against its surface, not in a pore between grains, but encased within the crystal structure itself, in a sealed pocket of the original liquid, cut off from everything outside when the crystal finished forming around it.

Fluid inclusions in halite have yielded viable microorganisms at timescales of hundreds of thousands of years. These organisms aren’t dormant in any seasonal sense. Preserved is closer — metabolizing at rates near zero inside a pocket of ancient brine, chemistry sealed in when the crystal grew.

^{Above: First, note the scale bar. The aqua bar is 2 microns wide, a little bit larger than the bacterial cell. Next, see all those moving dots; those are ancient bacteria trapped in this crystal aquarium for the last 1,000 years. Video by me.}

In the temperate version of petrichor, the mineral controls the chemistry: it traps the VOCs, catalyzes their transformation into aromatic compounds, holds the geosmin — and releases all of it at the moment of wetting. Remove the montmorillonite and you remove the record. In the Atacama version, the mineral control extends beyond surface chemistry. The halite makes its own water supply, filters radiation to survivable levels, buffers temperature swings that would otherwise kill everything — and then, as it crystallizes, takes the organism entirely inside itself. Without those specific properties — hygroscopic, translucent, thermally insulating, crystallizing from brine in a way that traps inclusions — nothing survives long-term. No organism waiting. No geosmin.

The mineral isn’t context for the smell — it’s the reason anything was alive to produce it, possibly across timescales that make a drought seem brief.

The Rover Problem

The Atacama Desert is NASA’s primary terrestrial analog for Mars. It is studied specifically as a model for where life might persist on another planet — inside minerals, not in soil or liquid water. Halite and gypsum — the same minerals sheltering endolithic communities in Chile — are found in abundance on Mars, particularly in regions like Meridiani Planum and Gale Crater that once hosted evaporitic environments.

The airborne chemicals that microorganisms release as metabolic byproducts are among the biosignature candidates astrobiologists prioritize. A biosignature is anything a non-living system couldn’t produce on its own — a chemical fingerprint that points specifically to biology. Geosmin and related compounds carry exactly this kind of fingerprint: they’re detectable at extremely low concentrations with the right instrumentation, and they’re very difficult to replicate through purely non-biological chemistry.

The smell you encountered on a sidewalk in the first minutes of a storm — that specific, emotionally loaded signal your nose has been processing since before you had language for it — is chemically related to what a Mars rover is designed to detect.

Bear and Thomas named petrichor in 1964 and understood that the mineral was doing something. What they couldn’t have anticipated is that at the extreme limit of the story, the mineral isn’t acting on chemistry from outside. It’s where life went when everything else became impossible — and the smell is what escapes when conditions briefly allow.

---

What You Were Smelling

You were not smelling rain.

You were smelling the end of a waiting period you never knew was happening. In the Arizona version: clay minerals releasing stored plant chemistry, Streptomyces exhaling after months of dormancy. The mineral held the record of the drought. The rain opened it.

In the Atacama version — the extreme limit of the same process — you might be smelling an organism that was sealed inside a halite fluid inclusion long before you were born. Released for the first time since the crystal formed around it.

In Kauai: nothing, because there was no record to open.

The thing I cannot resolve is this. The most emotionally reliable smell in human experience — the one that stops people mid-sentence and arrives before the storm, that has been doing something to humans in every culture that has experienced rain after a dry spell — is a biosignature. Chemical evidence of life persisting through conditions it was not supposed to survive.

We have been moved by it before we understood it. That part hasn’t changed.

The rover is looking for the same thing, because we know those are signs of life.

Aaron Celestian is the Curator of Mineral Sciences at the Natural History Museum of Los Angeles County, a former and long-time Affiliate Research Scientist at NASA's Jet Propulsion Laboratory, and recently elected as a National Fellow of the Explorers Club. He has published research on halophilic microorganisms in Atacama salts, biosignature preservation in evaporites, and the challenges of detecting life within a mineral on another planet. He is also an adjunct professor at USC and West Los Angeles College. He writes Pocketful of Χtals because mineralogy is stranger and more alive than most people have been told.

---

References

1. Bear, I.J. & Thomas, R.G. (1964). Nature of argillaceous odour. Nature, 201, 993–995.

2. Kumara, P.A. Amila & Deng, Xiaoxuan & Cooper, Paul & Cathro, Peter & Dias, George & Gould, Maree & Ratnayake, Jithendra. (2024). Montmorillonite in dentistry: a review of advances in research and potential clinical applications. Materials Research Express. 11. 10.1088/2053-1591/ad62c2.

3. Becher, P.G., Verschut, V., Bibb, M.J., Bush, M.J., Molnár, B.P., Barane, E., Al-Bassam, M.M., Chandra, G., Song, L., Challis, G.L., Buttner, M.J., & Flärdh, K. (2020). Developmentally regulated volatiles geosmin and 2-methylisoborneol attract a soil arthropod to Streptomyces bacteria promoting spore dispersal. Nature Microbiology, 5, 821–829.

   https://www.nature.com/articles/s41564-020-0697-x

4. https://eos.org/articles/desert-microbes-mine-for-water

5. NASA image repository.