≡ Menu

An Internal Source for Earth’s Water?

The last time we caught up with Wendy Panero’s work, the Ohio State scientist was investigating, with grad student Cayman Unterborn, a possible way to widen the habitable zone. Slow radioactive decay in elements like potassium, uranium and thorium helps to heat planets from within and is perhaps a factor in plate tectonics. In 2012, Unterborn argued that planets with higher thorium content than the Sun would generate much more heat than the Earth, allowing a habitable zone with liquid water on the surface correspondingly farther out from the star.

You can read about that work and its implications in Widening the Habitable Zone. I was reminded of it because Panero reported at the recent American Geophysical Union meeting on her latest direction, a study involving the formation of the Earth’s water. Recall that analysis of data from the Rosetta probe implicated asteroids rather than comets as the main delivery mechanism for Earth’s oceans (see Rosetta: New Findings on Cometary Water). Panero’s new work indicates a substantial part of Earth’s water was made right here.

Hydrogen atoms trapped inside crystal defects and voids within minerals can bond with the oxygen already plentiful in these substances, which is how rock that appears dry can contain water. The potential for a great deal of water exists in mantle rock, given that the mantle makes up more than 80 percent of the total volume of the planet, according to this Ohio State news release. Working with doctoral student Jeff Pigott, Panero subjected mantle minerals to high pressures and temperatures to create conditions like those deep inside the Earth.

Changes to the crystal structures of these minerals as a result of compression helped the team gauge how much hydrogen the minerals can store. The minerals in play here include bridgmanite (a high pressure form of olivine), which is abundant in the lower mantle, and ringwoodite (another form of olivine). 525 to 800 kilometers below the surface, both minerals exist in a ‘transition zone’ that can hold large amounts of water, which could be carried to the surface by convection of mantle rock, the same process that produces plate tectonics. The researchers’ tests indicated that bridgmanite contained too little hydrogen to be significant for water delivery, but ringwoodite emerged as a candidate for deep-earth water storage.


Image: This plate tectonics diagram from the Byrd Polar and Climate Research Center shows how mantle circulation delivers new rock to the crust via mid-ocean ridges. New research suggests that mantle circulation also delivers water to the oceans.

Panero and Piggott’s computer calculations helped to reveal the geochemical processes that would allow the minerals to rise through the mantle to the surface, the prerequisite for release of water into the oceans. Panero calls the relationship between surface water and plate tectonics “one of the great mysteries in the geosciences.” Her calculations with Pigott indicate that another mineral, garnet, may also play a role in delivering water from ringwoodite back down into the lower mantle, a circulation cycle enabled by plate tectonics that could maintain half as much water below the Earth as is currently flowing in today’s surface oceans.

A cycle like this would keep mantle water replenished even as it fed the oceans — the mantle would never exhaust its water supply. Says Panero:

“If all of the Earth’s water is on the surface, that gives us one interpretation of the water cycle, where we can think of water cycling from oceans into the atmosphere and into the groundwater over millions of years. But if mantle circulation is also part of the water cycle, the total cycle time for our planet’s water has to be billions of years.”

Such a process also relaxes the need for accounting for all of Earth’s water through bombardment from asteroids or comets. The latter surely played a role, but perhaps supplemented water that has been reaching the surface through plate tectonics ever since.


Comments on this entry are closed.

  • Mark Zambelli December 23, 2014, 15:29

    With the likes of the Champagne White-smoker belching out bubbles of liquid CO2 it shouldn’t be too suprising that water is bursting forth also. I’ve been puzzled lately by the bias towards thinking ocean-water had to come from asteroids/comets leaving out the mention of early volcanism supplying it’s share. Yes asteroids rather than comets as a delivery mechanism seems the way to go but I have reservations about how much they contributed, especially when we have direct evidence of the sheer volume of water that ‘rains down’ onto accreting solar systems from their contracting nebulae… I’m unaware of any mechanism that would remove the majority of this water from an entire planetary body.

  • Harold Daughety December 23, 2014, 16:23

    I have often wondered about the solubility of water vapor in molten rock, especially under high pressure. Gaseous H2O is a highly polar substance and should be the most soluble gas in high pressure magma.

  • Michael December 23, 2014, 17:20

    I am also favouring silicate dust and rocks interacting with the early hydrogen atmosphere to generate water. As they come in they would get pretty hot, enough for the hydrogen to bond to the oxygen in the silicates and be released as water.

  • Mike December 23, 2014, 19:17

    Kind of makes you wonder just how much planetary water might still remain on those inner-system M-dwarf planets that have to endure their stars’ lengthy (.5 Gyr to 2 Gyr) overheated adolescence. If we’re considering a billion year or more rock water cycle.

    During the process of planet formation when planetesimals are colliding and forming proto-planets how much of the planetesimals’ water is lost as compared to being retained in the deep sub-surface rock of the building proto-planet?

    How much of Earths’ water was acquired and retained during the initial planet formation and how much contributed by asteroids and comets? These questions directly address the possible aridity of planets in systems without the large outer system planets needed for a late heavy bombardment type of water acquiring process.

  • Alex Tolley December 23, 2014, 20:14

    If there is water in the mantle, then this may also be the source of methane.
    Perhaps this idea is testable on deep gas deposits if we can find them.

    Wouldn’t water in the mantle impact “weathering” and rock types as well?

  • Ashley Baldwin December 23, 2014, 21:16

    I don’t think this is about solubility of water so much as it being an intrinsic part of the chemical structure of the rock. Whether that is there from the formation of the rock at the beginning of the solar system or subsequently following asteroidal/commentary bombardment is lead important than the fact it was always there . Totally independent of planetary surface conditions such as a the “drying” effect of the prolonged exposure to a protostar or young sun’s “space weather” that has been described in the literature recently and diligently reported on thus site . Whatever is thrown at a planet , it can always shield its volatile content to released at a later date via plate tectonic related vulcanism. The key to the kind of secondary atmosphere here on Earth and critical to habitability. Even the fearsome battering handed out by M dwarfs could be circumvented against a day when they settle down to a slowly rotating, non flaring, middle age. TESS just got even more exciting .

  • Mike lorrey December 23, 2014, 23:01

    This also explains why mars and Venus have little and no water respectively. One is too small. The other has too thick a crust. Both lack the large moon necessary to keep a tectonic system running. This explains the large moon rare Earth theory.

  • Michael December 24, 2014, 17:05

    @Mike lorrey December 23, 2014 at 23:01

    ‘This also explains why mars and Venus have little and no water respectively. One is too small. The other has too thick a crust. Both lack the large moon necessary to keep a tectonic system running. This explains the large moon rare Earth theory.’

    If you look at Venus it looks as if it turned itself inside out with massive volcanic resurfacing at least within 500 million years. So any water there would have released, I suspect there was little water to start with in the mantle.

  • Dmitri December 24, 2014, 18:35

    Absolutely fascinating writing. It correlates with many things published in recent week but the most I regard this as possibility presence of water (liquid) not always as confirmation of biological / microbial life – purely geological byproduct.

    Let’s pair this with recent claims & discoveries:
    1) 2012 claims by NASA teams after statistically re-evaluating 1975 Viking 1 & 2 results of soil biological activity to be too complex to be explained non-biological only.

    2) Russian Academy of Science member Leonid Ksanfomaliti (Леонид Ксанфомалити) 2012 claims after re-evalutating Russian Venera probe pictures of Venus surface spotting “symmetrically shaped object with tail” resembling Earth’s scorpion. He has some substantial & verifiable arguments up to verdict from designers of Venera cameras “We don’t appreciate what you claim. Yet we can’t come up with any other explanation what we see”
    https://www.youtube.com/watch?v=xI0EOucWt2A (in Russian, no subs)

    3) NASA’s Mars Science Laboratory announcement on last week AGU on their episodic substantial discovery of methane presence over longer period where they say explicitly this is not confirmation of biological activity but from all the explanations biotic source can not be ruled out.

    4) NASA’s Systems Analysis and Concepts Directorate recent announcement to carry out concept study of manned Venus mission hovering in Venus atmosphere @ 50 km in High Altitude Venus Hovering Vehicle or in layman’s term Zeppelin.

    With the knowledge of 2014 A.D, me including, the Curiosity result hints to “we are a step away for confirmation of microbial life on other planet”.
    There are strong circumstantial evidence not explainable by other than “the signs of life on very first missions to Mars were misinterpreted due to unexpected results not within in our understanding”. Add to that claims of “a symmetrical object with tail on Venus appearing & disappearing between the shots of the Venera probe” gets to the realm “confirmed (soon)”.

    This might mean:
    a) water on Europa doesn’t mean anything even if there is liquid ocean underneath – byproduct of geological process.

    b) Burp of methane on Mars would most probably end up in non-biological explanation like byproduct of geological process.

    c) If comets generally are ruled out as the vessels for liquid water during early bombardment / their effort is not substantial does asteroids bear the test to be the ultimate source of water?

    d) The general process of water as geological byproduct would explain why Titan is as it is – hydrocarbon based – yet heavily resembling Earth geological activity & logic. Arbitrary as it seems this pops in my mind as 1st circumstantial evidence of validity for such process.

    See the video of Dec 2013 AGU presser on Titan findings & direct comparison w Earth’s body of liquids – https://www.youtube.com/watch?v=YmLm3rto64I

    d) NASA’s vision to have manned mission to hover in Venus atmosphere @ 50 km strongly resembles with claims the atmosphere between 50 – 70 km being very Earth like in its meteorological terms. Venus is next to us with atmosphere that resembles Earth’s one yet in extremes. Probability finding a bone on Mars is close to zero yet if comparing such probability with known to us similar atmosphere manned mission to Venus for answers of origin of life – possible confirmation – seems now as priority considering getting to Venus & breaking in high altitude is much easier than in any other planet in Solar system.

    e) As space has shown us it doesn’t operate as the man has foreseen it most probably signs of biological activity would be found on Venus rather than on Europa, Enceladus. I would exclude Titan completely.

    f) If Mike Lorrey can elaborate on “Venus has too thick crust to have water”. Some radar data of Venus Express indicating on recent volcanic activity on Venus not more than in last 2.5 million years.

  • Marshall Eubanks December 25, 2014, 18:27

    I actually regard this as good news for the terraforming of Venus. Venus appears to have little or no surface water*, but it certainly has surface volcanology (see http://www.hou.usra.edu/meetings/lpsc2014/pdf/2556.pdf), which will release water from the mantle. If / when we make Venus into a decent place, we will also have to keep it that way, and a source of deep water will certainly help in the long run.

    * This is based on Deuterium/Hydrogen ratios of Venus and the Earth. Given the recent variations of D/H ratios found in the solar system, not least by Rosetta, I don’t regard this assumption as quite as robust as I did a couple of years ago. We need to go and find out.

  • Harry R Ray December 26, 2014, 10:27

    Dmitri: You forgot to mention the most shocking news of all; a claim (unfortunately JUST a claim for now) made by Dr Nora Noftke, that a rock formation named “Gillespie Lake Member” most likely consists of a group of microbial mats (kind of like stromatolites with their rounded tops sheared off and flattened. Click on http://www.themeridianijournal.com to view the image).

  • Harold Daughety December 26, 2014, 18:46

    I do not believe that hydrogen gas can reduce silicates to produce oxygen and reduced silicon compounds: the reaction goes the other way. The equilibrium physical and chemical properties of supercritical H2O/molten “rock” systems could be such that a substantial portion of the incoming water in the formation of the earth remained dissolved in the molten rock. A small mole fraction of H20, given the low molecular weight, would result in a higher weight fraction. And volcanism is a vivid example of supercritical H2O and CO2 in solution in magma.

  • Dmitri December 27, 2014, 14:26

    Terrific find! Here is a 14 minute podcast by Dr Nora Noffke explaining her claims. Clearly she talks what she knows. Finds like these truly needs time for oneself to let it settle. Hoping her voice would be heard.

    Let this be for future to come for things recall later.

    Like her Norwegian accent.

  • Dmitri December 27, 2014, 14:39

    Link to Dr Nora Noffke article explaining her claim with direct link among others to her detailed 24 pages paper.


  • Michael December 28, 2014, 3:10

    Is it possible that the heat of the drilling is causing a chemical reaction to form methane? So in a way even if life is not present it is a good discovery as there is now more pieces to the complex puzzle of life in place. Missing organics is big hurdle to life’s development.

  • Dmitri December 28, 2014, 14:09

    The episodic burp of methane release on Mars lasted over 2 months steadily increasing every time when measurements were taken Ending up with maximum 11 parts per billion per volume averaging all 4 measurements in 7,02 ppbpv. This is not a mistake. This is in-situ registered unknown phenomenon with unknown planetary origin.

  • Eniac December 30, 2014, 0:26

    Harold Daughety is correct, silicate cannot be reduced with hydrogen. Other metal oxides, however, can, and a large fraction of rock is not silicate.

  • Michael December 30, 2014, 10:38

    @Eniac December 30, 2014 at 0:26

    ‘Harold Daughety is correct, silicate cannot be reduced with hydrogen. Other metal oxides, however, can, and a large fraction of rock is not silicate.’

    Silicates can be reduced by hydrogen (not necessarily complete), if silicate dust or rocks move at high velocity through the hydrogen envelop of the early Earth environment water will be released during the reaction. Elevated temperatures are required and are certainly available during entry into the upper hydrogen dominated atmosphere of the early Earth. The temperatures would also be high enough for reactions with nitrogen and carbon dioxide to form nitrides and carbides again releasing water as well.

  • Dmitri December 30, 2014, 19:27

    I’ll post a brilliant blog written by Emily Lakdawalla in October about AGU’s meeting on Mars geological findings by Opportunity & Curiosity that actually turns out to be unimaginably more complex and not so clear cut as we have taken comparing to geological understanding of Earth’s geology. Mars was never a wet planet. Not as wet as Earth but not dry as Earth. Water on Mars was for very brief period.

    Neatly hooks with this article.


  • Dmitri January 1, 2015, 10:47


    I wrote Earth twice. 2nd should have been Moon.

    “Mars was never a wet planet. Not as wet as Earth but not dry as Moon. Water on Mars was for very brief period.”

    Now there comes another twist regarding Venus specific geophysical features referring on better suitability for life (as we know it).

    Fresh article from Space.com by Dima Bolmatov on his computer simulations show CO2 on Venus has been in supercritical stage more naturally than anticipated resulting in bubbles encrusted with thick layer of liquid. That is coupled with recent find bacteria do fine in supercritical CO2 as in Earth’s liquid environment.

    The beauty with Venus is the fact it still has very active climate like on Earth what in turn severly hinders its robotic exploration capability just like on Earth & probably the robotic explorers won’t have any survivabilty like on sterlie Mars. Sterile here doesn’t mean bacterial life forms are excluded.

    Maybe the space exploration has strived too far into understanding origina of life through geology that still gives us magnificent inside on each planet geological unique past but none of them have shown us do we have life on other planets. Maybe it’s time to shift the paradigm and push them into find the bone on another planet. We could spend forever on puzzling how specific geological feature / process corresponds to theoretical habitability on the planet. Even finding the answer that would be still applicable to microbs.


  • Dmitri January 1, 2015, 12:43

    Addition to Curiosity methane finding it’s time to recall ESA’s Mars Express Planetary Fourier Spectrometer (PFS) result from 2004 where they repoerted finding methane on Mars in quantities of 10 ppbv.

    In 2009 Michael Mumma of NASA Goddard released detailed map of methane spectra plotted on Mars map result in truly astonishing feat – http://www.nasa.gov/images/content/303598main_mjmumma_vid_01_1280_01.jpg

  • Dmitri January 1, 2015, 12:50

    The blog entry of the Mars methane distribution map – http://dawninhampton.blogspot.com/2014/12/like-bad-penny-methane-on-mars.html

    I’m more astonished on realizing Mars has natural process of methane destruction. 30 ppbv is astonishing peak but considering how long it takes naturally to decay the methane detected by Curiosity but how rapidly it went away is a true eye opener.

  • Harold Daughety January 1, 2015, 16:22

    On the source of the earth’s water, there is no reason to look for synthesis. My point is that water has existed as long as oxygen has, and from the earliest accumulation of material to form the planets water was brought in. But is it lost from the heating that turned the proto-earth into a ball of molten rock? For a certain answer to that, the solubility of supercritical H2O in the molten rock must be considered. The equilibrium vapor composition would depend on the thermodynamic properties and the liquid phase composition of the magma. H2O as a supercritical gas would be dissolved in the molten rock: the quantity, as my old physical chemistry prof would say, is a problem for the students, meaning he, and also I, have no idea.

    Metal oxide reduction with hydrogen is known and is surely a source of water. But hot metals react with water to produce hydrogen and metal oxide. The driving force to produce water would be the escape of water from the reaction zone. As to the “partial” reduction of silicates, what is the reaction?

    On hydrocarbon formation in the earth, metal carbides react with water to form hydrocarbons: acetylene from calcium carbide is an example. Other metals generate other hydrocarbons, including methane. Acetylene, methyl acetylene, and a variety of unsaturated hydrocarbons could be formed, but only if the reactants were present and conditions were right. Were they?
    A lot of unknowns and even more speculation.

  • Michael January 1, 2015, 17:18

    @Dmitri January 1, 2015 at 12:50

    I think it is unlikely that Venus had anymore CO2 than it has now as the temperature is just so hot it would have baked out almost completely if not entirely. However when the Sun was a little cooler I would think the surface atmospheric density would have increased giving more weight to erosion processes.

  • Michael January 2, 2015, 16:32

    @Harold Daughety January 1, 2015 at 16:22

    ‘Metal oxide reduction with hydrogen is known and is surely a source of water. But hot metals react with water to produce hydrogen and metal oxide. The driving force to produce water would be the escape of water from the reaction zone. As to the “partial” reduction of silicates, what is the reaction?

    In a simple silicon dioxide partial reduction reaction the hydrogen from the surround environment of the earth interacts with the incoming high velocity dust converting silicon dioxide into silicon monoxide and water.

    SiO2 + H2 –> SiO + H2O (endothermic)

    The incoming silicate dust interacts with the hydrogen high up in the earths atmosphere (or accretion region) with the water and SiO products which are denser falling to the surface of the earth where it is cooler preventing or limiting recombination. My thoughts are that water that we see on earth formed around the same region not by a huge number of comets although they did add some water.

  • Dmitri January 2, 2015, 17:28

    Can someone pinpoint how to get access to presentations made on Dec 16 2014 American Geophysical Union. German Aerospace Center DLR Jörn Helbert did interesting presentation on interpretation of Venus Express VITRIS spectral analyses leading up to understanding oldest surfaces on Venus could be made of granite. Trick is spectral obtained by Venus Express can’t be compared with Earth minerals as the pressure & temperature yield completely unique to Venus result. Quick search didn’t reveal any hint to presented materials.

    1) “The results suggest that the oldest surfaces on Venus could be made of granite, which on Earth forms through tectonic activity. That might mean Venus once had continents. Many geologists think tectonic activity requires the presence of water, which could mean Venus once had oceans and, perhaps, could have hosted life.

    “This kind of data… provides a fantastic way to peek to the surface from orbit, so we can do actual mineralogy from orbit like a real geologist,” says Thomas Widemann at the Paris Observatory in France.

    Widemann is now working with Helbert to build an instrument specifically for orbital composition scans. It could fly on a future Venus mission, such as ESA’s planned EnVision mission or the NASA proposal known as VIRTIS.”

    2) The permanent cloud cover of Venus prohibits observation of the surface with traditional imaging techniques over most of the visible spectral range. Venus’ CO2 atmosphere is transparent exclusively in small spectral windows near 1 µm. The Visible and Infrared Thermal Imaging Spectrometer (VIRTIS) team on the European Space Agency Venus-Express mission have recently used these windows successfully to map the southern hemisphere from orbit. VIRTIS is showing variations in surface brightness, which can be interpreted as variations in surface emissivity. Deriving surface composition from these variations is a challenging task. Comparison with laboratory analogue spectra are complicated by the fact that Venus has an average surface temperature of 730K. Mineral crystal structures and their resultant spectral signatures are notably affected by temperature, therefore any interpretations based on room temperature laboratory spectra database can be misleading.
    In order to support the interpretation of near-infrared data from Venus we have started an extensive measurement campaign at the Planetary Emissivity Laboratory (PEL, Institute of Planetary Research of the German Aerospace Center, Berlin). The PEL facility, which is unique in the world, allows emission measurements covering the 1 to 2 µm wavelength range at sample temperatures of up to 770K. Conciliating the expected emissivity variation between felsic and mafic minerals with Venera and VEGA geochemical data we have started with a set of five analog samples. This set includes basalt, gneiss, granodiorite, anorthosite and hematite, thus covering the range of mineralogies. Preliminary results show significant spectral contrast, thus allowing different samples to be distinguished with only 5 spectral points and validating the use of thermal emissivity for investigating composition.
    This unique new dataset from PEL not only allows interpretation of the Venus Express VIRTIS data but also provide a baseline for considering new instrument designs for future Venus missions.


  • Josh January 3, 2015, 19:45

    Anyone know how water could (theoretically) be extracted from Ringwoodite?

  • Michael January 6, 2015, 11:01

    @Josh January 3, 2015 at 19:45

    ‘Anyone know how water could (theoretically) be extracted from Ringwoodite?’

    Heat it until it melts and reduce it with hydrogen as I believe the contained water is in the form of a hydroxide.

  • Dmitri January 14, 2015, 15:37

    Don Lowe 2010 talk in SETI on Late Heavy Bombardment brings up some very important questions regarding LHB but also hooks with the current topic on origins of Earth terrestrial water. Per Don Lowe et. al LHB not just didn’t ended suddenly, it continued after we assumed being the end of LHB. Worse – when the bombardment frequency is extrapolated per evidence in sedimentary the frequency is not bigger than the current nowadays bombardment. He specifically stresses the geological period we talk about is significantly long one. There are definitely substantial fluctuations w/o clear traces just all the data pointing to continuous bombardment that continues up to this day.

    The objects under consideration are meteorites, heavy, metal rich ones. That in turn raises question if comets didn’t play major role in bringing water to Earth & there is not enough meteorites to have such volume of water where did it come from?

    There’s suspicion of Hungarians class (E-belt, E stands for Extended) meteorites are the building blocks that has now exhausted their balance an part of Earth, Moon, & Mars (craters in) crust.
    Lunar evidence of Late Heavy Bombardment has been interpreted to suggest that large-body impacting declined rapidly after about 3.8 Ga and that by 3.5 Ga the terrestrial bombardment rate was not much greater than the impact rates of today. In 1986 and 1989 Dr. Lowe and colleagues described four major layers of spherical particles in the 3.22-3.55 Ga Barberton greenstone belt (BGB), South Africa, ranging from 3,472 to 3,243 Ma, and interpreted them to represent the products of large terrestrial impacts of bolides 20-50 km in diameter. Since describing and interpreting these early impact layers, they have identified at least three additional thick layers of spherules in the Barberton belt that likely represent deposits of large impacts, and two new layers that display some geological features associated with impacts. Large impact layers have been identified to date in most of the major sedimentary units in the BGB. Intervening sections are composed largely of volcanic rocks where the record of impact events is unlikely to be preserved: it seems likely that other large impacts occurred during this period without leaving a record. These layers suggest that Earth continued to be bombarded by large extraterrestrial objects late into the Archean, at least until 3.2 Ga. The large sizes possible for these objects means that, while none was probably a sterilizing impact, many may have severely heated the oceans and atmosphere, boiled off the upper layer of seawater. The 3.8-3.2 Ga development of the Earth’s surface environment and life may have been constrained largely by the continuing flux of large impactors. Only as that flux declined in the Late Archean were stable surface systems established within which non-thermophilic organisms and a stable geodynamic system could develop and evolve.


  • Dmitri February 3, 2015, 12:07

    Fresh presentation that hooks with the topic & discussion made by professor Sarah Stewart fr Harvard University where she discusses in details her work published in December 2013 journal Nature on new research made into similar origin of Earth & Moon, how to make the findings fit with the models that could explain the observations.

    In short – current models don’t support any of the findings of Earth & Moon similarity. If the very strict angular momentum limits are lax and the models are redone on super-fast rotating Earth (2-3 hours rotation over its axis) and simulated collision with a Mars size object then the debris cloud & fluid super critical conditions allowing model the more consistent numerical results for describing Moon formation from proto Earth. Despite the fact the whole upper mantel of the Earth is vaporized by the impact the terrestrial water does not evaporate, leaving pretty much it to Earth.

    The theory is such a neat one that this alone needs return missions to Moon being as important & challenging as the Rosetta mission to comet.

    Her paper & cover of Nature.

    Presentation in SETI Institute on Jan 28 2015. Ignore the bad audio, concentrate on her talk.