Hadaean to Palaeoarchaean stagnant-lid tectonics revealed by zircon magnetism

Plate tectonics is a fundamental factor in the sustained habitability of Earth, but its time of onset is unknown, with ages ranging from the Hadaean to Proterozoic eons1–3. Plate motion is a key diagnostic to distinguish between plate and stagnant-lid tectonics, but palaeomagnetic tests have been thwarted because the planet’s oldest extant rocks have been metamorphosed and/or deformed4. Herein, we report palaeointensity data from Hadaean-age to Mesoarchaean-age single detrital zircons bearing primary magnetite inclusions from the Barberton Greenstone Belt of South Africa5. These reveal a pattern of palaeointensities from the Eoarchaean (about 3.9 billion years ago (Ga)) to Mesoarchaean (about 3.3 Ga) eras that is nearly identical to that defined by primary magnetizations from the Jack Hills (JH; Western Australia)6,7, further demonstrating the recording fidelity of select detrital zircons. Moreover, palaeofield values are nearly constant between about 3.9 Ga and about 3.4 Ga. This indicates unvarying latitudes, an observation distinct from plate tectonics of the past 600 million years (Myr) but predicted by stagnant-lid convection. If life originated by the Eoarchaean8, and persisted to the occurrence of stromatolites half a billion years later9, it did so when Earth was in a stagnant-lid regime, without plate-tectonics-driven geochemical cycling.

I can follow the line of arguments developed in the study.The interpretation is reasonable and might stand the test of time.The 'plate tectonic test' (to make it highly unlikely that plate motion would have occurred during 3.9-3.4Ga) is performed with Phanerozoic plate motion models.This is logical since plate tectonics for sure was active during the entire Phanerozoic.However, the BGS and JH sample locations are associated with the Kaapvaal and Jilgarn-Pilbara cratons for which the oldest APWP has been proposed for a period between ~3.1 and ~2.7 Ga: they would be juxtaposed and constitute the Vaalbara super-craton (e.g.Zegers et al., 1998;de Kock et al., 2009).OK, this is younger than the time span considered in the present study and the APWP is not beyond discussion (e.g.Evans and Muxworthy, 2019).Some discussion on how the paleolatitudes compare and what would be the paleolatitude trajectory of the Vaalbara super-craton between ~3.1 and ~2.7 Ga would add.Plate motion of several cm per year to over 1 m per year have been argued for (the latter is very high for present-day standards).Is the change over from stagnant-lid tectonics to (present-day-style?) plate tectonics occurring within a few hundreds of millions of years?What if present-day NW Australia and South Africa were juxtaposed to each other already before ~3.1 Ga?OK, they would be stationary (latitudinally) based on the paleointensity results presented here, but is it then warranted to infer a global signal?What are thoughts on when plate tectonics (implicitly present-day style?) would have initiated?
Below are specific comments.

First paragraph
With the onset of plate tectonics present-day style plate tectonics is implied?In the stagnant-lid regime there was volcanism as well?So, nutrient supply (not per se recycling) was possible?Main text P2 first paragraph.I can follow very well the line of reasoning developed but there is quite some generic paleomagnetic knowledge assumed implicitly.That the field is dipolar and that inclination and intensity thus scale with latitude.This probably might be presumed knowledge.Does the same also pertain to that individual detrital grains in a sediment cannot provide directional information on paleolatitude (probably clear) but that paleointensity still can do?This might warrant a concise explanation.
References 16 and 17 refer to much younger rocks than dealt with here?What is implied with 'higher grades'?The younger sedimentary units have not been or barely metamorphosed?Second paragraph.Zircon by itself cannot provide paleomagnetic information, it is the magnetite inclusions that are responsible for that.
Third paragraph.That the inclusions have not been magnetically reset is crucial.Perhaps it is wise to briefly recapitulate why the grains are not reset.Playing devil's advocate: the reason for the nonvarying paleointensity between 3.9 and 3.4 Ga is that the crystal radiometric ages vary between those two values, but the magnetic age would be 3.4 Ga for all of them.P3 top paragraph.The isolated iron particles are not consisting of metallic iron isn't it?Be specific as to their magnetite nature I would say.
Second paragraph and throughout text: recalculate emu to Am^2 since that is the proper SI unit.
The reason to adopt the 565°C paleointensity protocol in the Jack Hills studies is to see through later medium grade metamorphism.Metamorphic temperatures that prevailed require laboratory heating to 565°C for demagnetization and remagnetization experiments to obtain meaningful results.What is the discriminative power of the paleointensity at that temperature?The 565-580°C portion of the blocking/unblocking temperature spectrum is small.In the Barberton Greenstone Belt lower metamorphic temperatures prevailed so that in principle a larger portion of the demagnetization spectrum can be used for meaningful partial thermoremanent magnetization acquisition experiments: how low a laboratory temperature is required to see through the metamorphism in the Barberton Greenstone Belt?Do the paleointensity results at let's say 520°C deliver the same outcome as those at 565°C?Third paragraph: the zircons with a much younger rim (how much is much?) presumably do not qualify.OK, the paleointensity is representative of the youngest age.
P4 discussion.Perhaps it is wise to provide some more information on plate tectonics vs stagnant lid tectonics.Are these seen as the only two options?What is the difference between a mantle plume and a heat pipe?P5 top.But just before the constant paleointensity has been used to argue for no (significant) plate motion and here it is related to geodynamo efficiency.How can somebody tell apart the difference?The reasoning in the final bits may be a little quickly developed.
P5 sample preparation and selection.What is the reason for picking a NRM cut-off in the first place?Just practical issues?If the samples are not sufficiently magnetized they will deliver meaningless results, or is there yet another reason?P6 geochronology.Curiosity, where does the geochronological SQUID3 acronym stand for? Figure 1 caption.Fe is not metallic Fe, name it magnetite or Fe-oxide?Figure 2. In case of zoned zircons, the youngest growth rim has magnetically reset all older portions of the respective grain?Zircon grows at higher temperatures than the maximum magnetite (un)blocking temperature?What are the demagnetization diagrams of the dated grains shown in c), f) and i)?In a) and b) (added vector difference sum decay curve?) the 565°C value is almost 0.5 of the starting NRM, what is that for c), f) and i)? Show the age dating of BGZ5-z1? Figure 3.Why are the other zircon paleointensity data shown in grey?Are they from yet another region?They seem to be 'high-ish'.How are the uncertainty ranges in the paleointensity calculated? 1 sigma, 2 sigma?It seems that the means of binned intervals do not come with an uncertainty, yet in extended data table 4 most bins contain more than a single value.
Extended data figures 1-3.Put the respective zircon grain labels on the appropriate panels.Australia) aims to use paleointensity data from Hadean-to Mesoarchean-age single zircons to infer that plate/mobile lid tectonics (as we know it) was not operating at this time.Instead, they argue that stagnant lid tectonics was operational.To do this, they obtained many detrital zircons from the so-called green sandstone layer in the Barberton Greenstone Belt of South Africa, where numerous Hadean to Mesoarchean detrital zircons grains have already been found.They performed paleointensity measurements on these grains, and those which produced acceptable intensities; they did follow up petrography via SEM (BSE and CL, and of course EDS), to find magnetite inclusions, followed by U-Pb geochronology via SHRIMP.This methodology is, of course, sound, and is similar to studies already published from the Jack Hills area of western Australia, where the authors have already worked and published data from.The paleo-intensity measurements appear to be near identical between the two sites, and are apparently constant between ⁓3.9 Ga and ⁓3.4 Ga, suggesting these two pieces of the crust did not move significantly along an apparent polar wander path, which appears to invalidate having plate tectonics, but suggesting instead that stagnant lid tectonics was dominant at this time.
Originality and significance.The results are indeed significant; if true, it would allow us to infer that plate tectonics was not operating before the Mesoarchean.This is much debated in the geological community, with estimates for when plate tectonics started varying between the end of the Proterozoic and the Hadean.Most scientists usually estimate either a Meso-or Neoarchean onset, but even this remains moot.This, of course, will have implications for life on our planet and others.As for novelty, this is less certain; this methodology has been presented before and has been vigorously debated in the geological community.Hence the authors need to find primary magnetite inclusions, as well as consistent ages throughout a single grain.Here, the only thing that has changed is from talking about the existence of a magnetic field around the Earth and the state of the core, to plate tectonics itself.
Data and methodology.The methodology is sound, as stated above.It is a valid approach, showing the paleointensity values in the zircons, the existence of magnetite inclusions in them, and their documentation by BSE, EDS and CL, coupled with U-Pb geochronology.The devil, however, is in the detail.There are not enough samples and data to form a solid conclusion, which leads to the Appropriate use of statistics and treatment of uncertainties.I am no specialist in paleointensity, but I can assess this: U-Pb geochronology.The SHRIMP U-Pb shows (approximately) 76 analyses from 35 grains.Zircons grains are notorious heterogeneous in terms of age (Pb loss), hence the advantages of spot dating of various zones and the documentation by BSE and CL.Therefore, it is important to document this heterogeneity by imaging and age dating.For statistics, a good number of analyses, therefore, is between three and five, with five being preferable.This means that of the 35 grains in this study, only seven have enough analyses.Of these seven: BGS1z9 has 13 analyses (13 concordant), with concordant 207Pb/206Pb ages between 3804 Ma and 4115, a time of over 300 Myr, not suitable.
BGS1z11 has eight analyses (eight concordant), with concordant 207Pb/207Pb ages between 3721 Ma and 3784 Ma, a time of less than 100 Myr, suitable.
BGS2z2 has seven analyses (seven concordant), with concordant 207Pb/206Pb ages between 4017 Ma and 3468 Ma, a time of over 500 Myr, not suitable.BGS2z10 has six analyses (six concordant), with concordant 207Pb/206Pb ages between 3226 Ma and 4056 Ma, a time over 700 Myr, unsuitable.
BGS2z14 has three analyses (three concordant), with concordant 207Pb/206Pb ages between 3748 Ma and 3907 Ma, a time over 100 Myr, not suitable.
BGS2z15 has three analyses (three concordant), with concordant 207Pb/206P ages between 3388 Ma and 3769 Ma.A time of almost 400 Myr, is not suitable.
BGS5z12 has seven analyses (six concordant), with concordant 207Pb/206Pb ages between 3691 Ma and 3782 Ma, a time of less than 100 Myr, suitable.
Therefore, in my opinion, of the seven grains with good paleointensity, and sufficient analyses, only two are actually suitable.Additionally, if we assume a 100 Myr-age bin, both present a snapshot of only the 3700-3800 M age bin.You will still need an additional zircon to have an MSWD Reduced chisquared statistic), for the one age bin, and I actually recommend an additional two age bins, preferably from more than one location.The authors have also published such data from Jack Hills and used it in this study, but I did not evaluate that study herein.So, on this, I find this study unsuitable and far-fetched.
In conclusion, it is actually very difficult to evaluate this study.There are several other geological points to consider: there is a lot of geological debate around the existence of plate tectonics against stagnant lid tectonics, and what the evidence is for and against it.The authors completely ignore this.See some references below.Paleomagnetic data does not exist alone, this needs to be considered.It is easy to refer the authors to the papers of Nutman and Friend on the subject.Additionally, these are detrital data, so in a sense, they are not 'in-situ'.Therefore, trying to evaluate paleo-latitude in my mind is a bit of nonsense.These zircons come from blocks which do not exist any longer.The oldest evidence of stable crust is from 3.1 Ga in the case of the Kaapvaal Craton, and 2.7 Ga in the case of the Yilgaran Craton.These greenstone belts on them are not formed on the stable crust, and we have no knowledge about how 'exotic' they may be, and what the basement truly was.They could be nappes transported over large distances, for example.Any existence of the movement of the blocks needs to be considered, and actually, there is no good evidence.There is a hypothetical 'superblock' called Isukasia documented mostly based on 3.6 Ga magmatism and metamorphism worldwide.How do the authors account for such widespread magmatic and metamorphic ages without discussion?Paleomagnetic studies also account for so-called van der Voo criteria, especially in the Precambrian.Paleointensity in this study seems to try to bypass all this, but actually, it is still relevant.Adequate numbers of sites, field tests, rock magnetic studies on the zircon grains, cooling ages, and ages of magnetism.All of this seems to be ignored in the quest for a good story.Suggested improvements.Please consider the advice above, generate at least three different age bins between 3.0 Ga and 4.1 Ga, with between three and five suitable zircons.These zircons have to have good paleo-intensity measurements, documentation by SEM in BSE, EDS and CL of magnetite grains, and with between three and five U-Pb geochronology ages that are concordant (preferably more than 95% concordant), with ages falling within this 100 Myr age window.Consider the van der Voo criteria and the geology.This is a huge task, I know, but if you are going to make huge, revolutionary and ground-breaking research, this is what is needed; there is no other way.
References.Generally sufficient, although almost all of this type of work is mostly done only by this research team.However, some more citations of the limitations, i.e., Evans (2018) -RESEARCH FOCUS: Probing the complexities of magnetism in zircons from Jack Hills, Australia.Geology 46, 479-480, would be good, or Nutman et al. ( 2021) -Fifty years of the Eoarchean and the case for evolving uniformitarianism.Precambrian Research 367, 106442.cClarity and context.Generally, very good, but as I said, the devil is in the detail.One can present all the supportive statistics one wants, but you cannot ignore the fact: there is not enough data from enough zircons even to begin to consider your case.Additionally, to disprove plate tectonics in the Archean, one has to look further and deeper in the literature.
I conclude with a quote: Deciphering the Eoarchean to Paleoarchean geological record must come from observation, not modelling.

November 14, 2022
Reviewer 1: Comment: "I can follow the line of arguments developed in the study.The interpretation is reasonable and might stand the test of time.The 'plate tectonic test' (to make it highly unlikely that plate motion would have occurred during 3.9-3.4Ga) is performed with Phanerozoic plate motion models.This is logical since plate tectonics for sure was active during the entire Phanerozoic." Response: We thank the reviewer for this assessment.
Comment: "However, the BGS and JH sample locations are associated with the Kaapvaal and Jilgarn-Pilbara cratons for which the oldest APWP has been proposed for a period between ∼3.1 and ∼2.7 Ga: they would be juxtaposed and constitute the Vaalbara super-craton (e.g.Zegers et al., 1998;de Kock et al., 2009).OK, this is younger than the time span considered in the present study and the APWP is not beyond discussion (e.g.Evans and Muxworthy, 2019).Some discussion on how the paleolatitudes compare and what would be the paleolatitude trajectory of the Vaalbara super-craton between ∼3.1 and ∼2.7 Ga would add.Plate motion of several cm per year to over 1 m per year have been argued for (the latter is very high for present-day standards).Is the change over from stagnant-lid tectonics to (present-day-style?) plate tectonics occurring within a few hundreds of millions of years?
Response: We agree with the reviewer that the previous work on the hypothesized Vaalbara super-craton is not beyond discussion.As is also clear in the papers cited by the reviewer (specifically Z egers e t a l., 1 998, a nd d e K ock e t a l., 2 009, v s.E vans a nd M uxworthy, 2019), there is debate on the existence of a single craton involving the Kaapvaal and Pilbara.The involvement of the Yilgarn craton is another issue of debate; geologic ties between the Pilbara and Yilgarn for these ages are not straightforward.However, we are comfortable with the close geological association marked by the Ventersdorp Supergroup (Kaapvaal craton) and Fortesque Group (Pilbara craton), but these date to 2.7-2.8Ga.Thus, there is a 700 millionyear-long age gap between this association and the time when our Jack Hills and Barberton paleointensity/paleolatitude records start to diverge (∼3.4 Ga).This duration is roughly equivalent to that we examine to constrain the median maximum latitudinal motion characteristic of modern plate tectonics.Given that the median maximum latitudinal motion of 0-600 Ma plates is ∼76 o , it follows that even very large paleolatitude differences between t he hypothesized 'Vaalbara' super-craton at 2.7-2.8Ga, and the ∼3.4 Ga JH/BGS zirons could be reconciled by plate motion rates comparable to those of modern plate tectonics.We have added a paragraph in a new section 6 to the supplement with the references noted above, and discuss this issue for context.However, we prefer to restrict our main text to the test of moving vs. stationary lithosphere.
Comment: "What if present-day NW Australia and South Africa were juxtaposed to each other already before ∼3.1 Ga?OK, they would be stationary (latitudinally) based on the paleointensity results presented o C. Another is to be well above the metamorphic temperature for our paleointensity estimate.
The reviewer is correct in noting that there is a difference with the Jack Hills, which have experienced a higher metamorphic temperature.However, the other related reason -which we did not explicitly state but showed Figure 2b to illustrate-is to be within the unblocking of those grains most likely to preserve the primary paleointensity value.With single domain grains, we can expect unblocking closer to the Curie temperature, as compared to 520 o C.And Figure 2b shows a break in slope from a gradual demagnetization curve to a sharp drop in magnetic intensity; 565 o C is within this sharp drop and is better suited for a paleointensity estimate because we can have more confidence that this is within the single domain range.We now include this additional justification in the revised Methods.
Comment: "Third paragraph: the zircons with a much younger rim (how much is much?) presumably do not qualify.OK, the paleointensity is representative of the youngest age." Response: Yes, we assign the youngest age to the magnetization age.We have expanded on our exploration of different areas within the zircon for geochronology by describing the analyses conducted on each of the zircons in the revised supplement.We provide 77 additional age analyses to further bolster the age accuracy.
Comment: "P4 discussion.Perhaps it is wise to provide some more information on plate tectonics vs stagnant lid tectonics.Are these seen as the only two options?What is the difference between a mantle plume and a heat pipe?" Response: There are multiple ways crust might be constructed by stagnant lid tectonics, but the specific test here is of a plate tectonic style of mobility.However, we now add an addition reference to a review paper (Nutman et al., 2021) that discusses salient issues.In typical usage, heat pipes describe transfer of heat by advection along narrow conduits in the lithosphere, together with downward advection of cold lithosphere.A mantle plume is generally thought to transfer heat from deeper in the mantle, and there is no substantial downward flow near the mantle plume location.We feel the references to Moore and Webb (2013) and Nutman et al. ( 2021) are sufficient to address this for the reader.
Comment: "P5 top.But just before the constant paleointensity has been used to argue for no (significant) plate motion and here it is related to geodynamo efficiency.How can somebody tell apart the difference?The reasoning in the final bits may be a little quickly developed." Response: Yes, we agree.We have added a new section to the supplement to provide context.During the last 200 myr, paleomagnetic data (directions and intensity) and dynamo models show variations that probably reflect changes in the pattern of core-mantle boundary heat flux that in turn influence dynamo efficiency.If there had been large changes in paleointensity betwen 3.4 and 3.9 Ga, the solution would be non-unique.We would be unable to separate changes in dynamo efficiency from plate motion.But we do not see such changes, and therefore we can make inferences on latitudinal stability.
Comment: "P5 sample preparation and selection.What is the reason for picking a NRM cut-off in the first place?Just practical issues?If the samples are not sufficiently magnetized they will deliver meaningless results, or is there yet another reason?" Response: The cutoff is a practical issue of being able to measure accurately the NRM and the demagnetized value.NRMs less than the cutoff value can be measurable, but their demagnetized value could fall beneath the sensitivity of the ultra-high sensitivity SQUID magnetometer at the University of Rochester.We have added a line to the Methods to better explain this.
Comment: "P6 geochronology.Curiosity, where does the geochronological SQUID3 acronym stand for?" Response: Technically SQUID3 stands for nothing, other than the name of the software.We now add this to the Methods.This is the name used in the literature; it is capitalized but it is not an acronym; instead it seems to be a play on words, following SHRIMP (which is an acronym) to follow a seafood theme!In any case, this is the common usage in the geochronology community.
Comment: "Figure 1 caption.Fe is not metallic Fe, name it magnetite or Fe-oxide?"Response: Corrected.
Comment: "Figure 2. In case of zoned zircons, the youngest growth rim has magnetically reset all older portions of the respective grain?Zircon grows at higher temperatures than the maximum magnetite (un)blocking temperature?
Response: Yes, this is our interpretation.We have further explained this in the revised Methods, and in the description of each grain.
Comment: "What are the demagnetization diagrams of the dated grains shown in c), f) and i)?In a) and b) (added vector difference sum decay curve?) the 565 o C value is almost 0.5 of the starting NRM, what is that for c), f) and i)? Show the age dating of BGZ5-z1?" Response: In the 565 o C technique, magnetometer measurements are not made at intermediate temperatures to keep the total number of temperature treatments to a minimum, minimizing the risk of sample alteration.The NRM values for these grains, and NRM after treatment at 565 o C, are provided in Table 2, and of course will be available in the MagIC database.There is no significant trend in the NRM/NRM 565 value with age (e.g.Pearson Correlation coefficient of -0.06).However, because the NRM/NRM 565 value can be influenced by the directional characteristics when detrital zircons are the subject of study (i.e., varying angles between a low temperature overprint and a high temperature primary component), we do not advocate it as an indicator of rock magnetism.Instead, we feel the multidomain checks, also summarized in Table 2 and are all at low values, are better indicators of the viability of the rock magnetic recorders.
Comment: "Figure 3. Why are the other zircon paleointensity data shown in grey?Are they from yet another region?They seem to be 'high-ish'.How are the uncertainty ranges in the paleointensity calculated? 1 sigma, 2 sigma?It seems that the means of binned intervals do not come with an uncertainty, yet in extended data table 4 most bins contain more than a single value." Response: The grey data, identified in the caption, are not zircon data but are paleointensities from other single crystals from extant whole rocks.We now explicitly state they are from extant whole rocks.We had previously discussed these results relative to the zircon results at the end of Supplementary section 3.0 (now 4.0) and for convenience repeat that text here: "Paleointensities from the SCP analyses of Nondweni dacite samples are within the value of BGS values of the same age, but SCP of Barberton dacites at 3.45 Ga are higher.Both the Nondweni and Barberton SCP values are plotted as individual results from two relatively shallow intrusions and these might be expected to sample higher frequency variations of the geomagnetic field than the zircons."The uncertainty assignments on the individual zircon results follow the empirical approach used in Tarduno et al. (2015Tarduno et al. ( , 2020)), based on the comparison of full Thellier results and 565 o C results.Uncertainties of the combined BGS/JH data set are standard errors and are identified in the figure caption of Figure 4. We do not plot the uncertainties of the individual age bins in some other figures because that would obscure other data, but as noted these standard errors are available in Extended Data Table 4.
Comment: "Extended data figures 1-3.Put the respective zircon grain labels on the appropriate panels." Response: Done.
Comment: "P12.Extended data figure 2. It is not metallic Fe inclusions.Extended data figure 4.Here paleointensity data are plotted without uncertainty intervals." Response: We have modified the figure to read Fe-oxide.We feel it is better to plot the data without the uncertainties to see the trends and fitting.The uncertainties are plotted in the main text figure.
Comment: "Extended data figure 5. What is the red diamond in panel b?" Response: The point where the maximum latitudinal distance is reached.This was accidentally omitted and has been added.We thank the reviewer for catching this.
Comment: "Extended data table 2. Delta565 is the angle between the TRM vector and the applied field.It shows values up to 29 o .Ideally it should be close to zero.What is the underlying reason for the measured deviations?The amount of MD effects seems very low." Response: The deviations could have several sources.The rock magnetic source could be an anisotropy of the collection of magnetic particles in a given zircon.However, these are extremely challenging experiments and we can not exclude small error contributions related to the alignment of the laser beam and a zircon.We have noted this in the revised text.
Comment: "Supporting information.P3.Third paragraph.What is the meaning of acronym PDM? Check other acronyms throughout the document." Response: Thank you for catching this.Paleomagnetic dipole moment.We have noted this in the revised manuscript.
Comment: "P5.Third paragraph.Presumably the likelihood of observing two sites with no motion on separate plates is low.Is that written as such?" Response: Yes.We have clarified the sentence and thank the reviewer for catching this.

Reviewer 2
Comment: "Summary.The manuscript submitted by Tarduno et al. (2022-05-07225; Hadean to Paleoarchean stagnant lid tectonics recorded by the paleomagnetism of single zircons of South Africa and Australia) aims to use paleointensity data from Hadean-to Mesoarchean-age single zircons to infer that plate/mobile lid tectonics (as we know it) was not operating at this time.Instead, they argue that stagnant lid tectonics was operational.To do this, they obtained many detrital zircons from the so-called green sandstone layer in the Barberton Greenstone Belt of South Africa, where numerous Hadean to Mesoarchean detrital zircons grains have already been found.They performed paleointensity measurements on these grains, and those which produced acceptable intensities; they did follow up petrography via SEM (BSE and CL, and of course EDS), to find magnetite inclusions, followed by U-Pb geochronology via SHRIMP.This methodology is, of course, sound, and is similar to studies already published from the Jack Hills area of western Australia, where the authors have already worked and published data from.The paleo-intensity measurements appear to be near identical between the two sites, and are apparently constant between ?3.9 Ga and ?3.4 Ga, suggesting these two pieces of the crust did not move significantly along an apparent polar wander path, which appears to invalidate having plate tectonics, but suggesting instead that stagnant lid tectonics was dominant at this time.""Originality and significance.The results are indeed significant; if true, it would allow us to infer that plate tectonics was not operating before the Mesoarchean.This is much debated in the geological community, with estimates for when plate tectonics started varying between the end of the Proterozoic and the Hadean.Most scientists usually estimate either a Meso-or Neoarchean onset, but even this remains moot.This, of course, will have implications for life on our planet and others.
Response: We thank the reviewer for this assessment.
Comment: "As for novelty, this is less certain; this methodology has been presented before and has been vigorously debated in the geological community... Hence the authors need to find primary magnetite inclusions, as well as consistent ages throughout a single grain.Here, the only thing that has changed is from talking about the existence of a magnetic field around the Earth and the state of the core, to plate tectonics itself." Response: We note that our work is the first to make conclusions on mobile versus non-mobile lithosphere based on data constraining latitudinal motion for the Eoarchean to Paleoarchean interval.As discussed below and further examined with new data, we have probed different crystal areas within a zircon and obtain weighted mean ages.When a younger rim is present, we equate its age with the time of magnetization.
Comment: "Data and methodology.The methodology is sound, as stated above.It is a valid approach, showing the paleointensity values in the zircons, the existence of magnetite inclusions in them, and their documentation by BSE, EDS and CL, coupled with U-Pb geochronology.The devil, however, is in the detail.There are not enough samples and data to form a solid conclusion, which leads to the Appropriate use of statistics and treatment of uncertainties." Response: We thank the reviewer for this assessment of our use of statistics.As detailed Therefore, in my opinion, of the seven grains with good paleointensity, and sufficient analyses, only two are actually suitable.Additionally, if we assume a 100 Myr-age bin, both present a snapshot of only the 3700-3800 M age bin.You will still need an additional zircon to have an MSWD Reduced chi-squared statistic), for the one age bin, and I actually recommend an additional two age bins, preferably from more than one location.The authors have also published such data from Jack Hills and used it in this study, but I did not evaluate that study herein.So, on this, I find this study unsuitable and far-fetched." Response: We thank the reviewer for raising this issue which led to our recognition of an error in our supplementary data table.With this correction, one of the large age ranges discussed by the reviewer is removed.However, the larger point is that we have probed different areas within a crystal to obtain weighted mean ages.When a younger "rim" is identified, we assign that to the magnetization age.However, we recognize that we did not describe this fully, and we did not have multiple age analyses for the younger grains.To address these issues, we have made the following additional measurenents and associated revisions: • We now more clearly state in the manuscript that when a younger "rim" is identified, we assign that to the age of magnetization.conclusions).
• We have expanded the Methods, and in the Supplement we now include a description of the age analyses and assignment for each zircon.We note that our assumption that a single analysis would be sufficient to represent the age of the younger zircons is confirmed by our new data, but the additional analyses add rigor.
Comment: "In conclusion, it is actually very difficult to evaluate this study.There are several other geological points to consider: there is a lot of geological debate around the existence of plate tectonics against stagnant lid tectonics, and what the evidence is for and against it.The authors completely ignore this.See some references below.Paleomagnetic data does not exist alone, this needs to be considered.It is easy to refer the authors to the papers of Nutman and Friend on the subject." Response: Our first three citations discuss the geologic debate.However, as discussed below, we now specifically reference the review paper by Nutman et al. (2021) to further alert readers to the geologic data and interpretations.
Comment: "Additionally, these are detrital data, so in a sense, they are not 'in-situ'.Therefore, trying to evaluate paleo-latitude in my mind is a bit of nonsense.These zircons come from blocks which do not exist any longer.The oldest evidence of stable crust is from 3.1 Ga in the case of the Kaapvaal Craton, and 2.7 Ga in the case of the Yilgaran Craton.These greenstone belts on them are not formed on the stable crust, and we have no knowledge about how 'exotic' they may be, and what the basement truly was.They could be nappes transported over large distances, for example." Response: If the zircons come from highly varied sources, rather than 2 more or less restricted areas, this further bolsters our interpretation because if more areas of the globe have been sampled, the prediction of plate tectonics is that they should have greater latitudinal variation.
Comment: "Any existence of the movement of the blocks needs to be considered, and actually, there is no good evidence.There is a hypothetical 'superblock' called Isukasia documented mostly based on 3.6 Ga magmatism and metamorphism worldwide.How do the authors account for such widespread magmatic and metamorphic ages without discussion?" Response: In our manuscript we are of course not excluding the growth of continental crust during the Hadean to Eoarchean.We are saying that our data do not point to plate mobility typical of modern plate tectonics.We have emphasized this point by referencing the Nutman et al. review paper.
Comment: "Paleomagnetic studies also account for so-called van der Voo criteria, especially in the Precambrian.Paleointensity in this study seems to try to bypass all this, but actually, it is still relevant.Adequate numbers of sites, field tests, rock magnetic studies on the zircon grains, cooling ages, and ages of magnetism.All of this seems to be ignored in the quest for a good story.Suggested improvements.Please consider the advice above, generate at least three different age bins between 3.0 Ga and 4.1 Ga, with between three and five suitable zircons.These zircons have to have good paleo-intensity measurements, documentation by SEM in BSE, EDS and CL of magnetite grains, and with between three and five U-Pb geochronology ages that are concordant (preferably more than 95% concordant), with ages falling within this 100 Myr age window.Consider the van der Voo criteria and the geology.This is a huge task, I know, but if you are going to make huge, revolutionary and ground-breaking research, this is what is needed; there is no other way." Response: The Van der Voo criteria apply to paleomagnetic directions, so we have developed a different set of criteria that leads us to select only ∼3.5% of the zircons initially separated.These are very strict criteria, and are summarized in the Methods together with the success statistics.These criteria start with light microscopy, and extend to several different types of paleointensity data checks, to SEM analyses and finally U-Pb analyses.Field tests have already been published (e.g.Usui, Tarduno et al., 2009).Ultimately, the time signature of paleointensities from zircons at different global localities is important, and this is the first study to make this comparison.
As noted above, for the Eoarchean zircons our ages are based of weighted means of several spots.We have now better documented this.We have returned to the mount and added 77 new age analyses.The new ages are now weighted means based on at least 3 analyses as requested by the reviewer and are supplied in Table 1 (there are a few exceptions, explained in the Supplement).All the values and related statistics for paleointensity and inferred paleolatitude have been recalculated.Changes are quite minor and all of the conclusions we draw in the original manuscript are unchanged.However, we appreciate the reviewer's insistence on further tests of the accuracy of our ages.Comment: "Clarity and context.Generally, very good, but as I said, the devil is in the detail.One can present all the supportive statistics one wants, but you cannot ignore the fact: there is not enough data from enough zircons even to begin to consider your case.Additionally, to disprove plate tectonics in the Archean, one has to look further and deeper in the literature.I conclude with a quote: Deciphering the Eoarchean to Paleoarchean geological record must come from observation, not modelling." Response: We agree with the reviewer's closing quote.We are not attempting to disprove plate tectonics on geological grounds.We are simply conducting a test of latitudinal mobility for the first time based on data, not modeling.By referring to the geologic work documenting shortening in the main text, we feel our revision now further emphasizes geologic data (also discussed in our first three references/citations), and we thank the reviewer for the suggestion.
P12. Extended data figure 2. It is not metallic Fe inclusions.Extended data figure 4.Here paleointensity data are plotted without uncertainty intervals.Extended data figure 5. What is the red diamond in panel b? Extended data table 2. Delta565 is the angle between the TRM vector and the applied field.It shows values up to 29°.Ideally it should be close to zero.What is the underlying reason for the measured deviations?The amount of MD effects seems very low.Supporting information.P3.Third paragraph.What is the meaning of acronym PDM? Check other acronyms throughout the document.P5.Third paragraph.Presumably the likelihood of observing two sites with no motion on separate plates is low.Is that written as such?*********************************************************** Referee #2: Summary.The manuscript submitted by Tarduno et al. (2022-05-07225; Hadean to Paleoarchean stagnant lid tectonics recorded by the paleomagnetism of single zircons of South Africa and Comment: "References.Generally sufficient, although almost all of this type of work is mostly done only by this research team.However, some more citations of the limitations, i.e., Evans (2018) -RESEARCH FOCUS: Probing the complexities of magnetism in zircons from Jack Hills, Australia.Geology 46, 479-480, would be good, or Nutman et al. (2021) -Fifty years of the Eoarchean and the case for evolving uniformitarianism.Precambrian Research 367, 106442."Response: We are happy to add the Nutman et al. 2021 reference and to refer to it in the main text.