A geological telescope through the galaxy?

The prevailing astrophysics paradigm is that after their initial formation, spiral galaxies are rotationally stable over times >6 Gyr, as significant velocity disturbances of stars and gas clouds are rapidly damped on the timescale of their rotation (Marr 2022). Moreover, the presence of supermassive black holes at the centre of many galaxies may contribute to their stability. This equilibrium between gravitational forces and angular momentum ensures the long-term stability of a galaxy. The most significant event a galaxy can experience is a merger with another galaxy. These types of events have the potential to cause minor disturbances or destroy the structure of a spiral galaxy, resulting in a spherical elliptical galaxy with a more randomized movement of stars, unlike the disc-like kinematics of a spiral galaxy. Whereas galactic mergers are thought to be important for the formation and evolution of galaxies, the last known major merger with the Milky Way was around 10 Gyr ago (Helmi 2020). Therefore, the disc of the Milky Way can be assumed to be dynamically stable for the timeframe of Earth's evolution.

Here, we produce local mass distribution models for the Solar System for various grand design barred spiral arm galaxies (Fig. 1; Supplementary Data 1). We employ the same base galactic model as previously used to understand periodicity within Archean zircon Hf isotope time series from the North Atlantic and Pilbara cratons (Kirkland et al. 2022). This model depicts the Milky Way as an evenly spaced four-arm spiral galaxy, with the Solar System located 8 kpc from the galactic centre with an orbital velocity of 240 km s⁻¹ in the galactic disc (Karachentsev and Makarov 1996). The orbital velocity of 210 km s⁻¹ (Camarillo et al. 2018) for the quasistatic spiral structure produces a galactic period (the time for the Solar System to do one complete orbit in the reference frame of the spiral structure) that is much longer than the sidereal period (one orbit about the galactic centre). A galactic period of 748 Myr for our model is consistent with the duration of superchrons, periods of stable magnetic field, with an average spiral arm transit time of 63 Myr (Gillman and Erenler 2019). The local mass density is treated as a normal probability function centred on the spiral arms with a frequency of 187 Myr between the centres of each spiral arm transit (Kirkland et al. 2022). This model is taken from the assumption that the azimuthal density profile is dependent on radial distance from the galactic centre (Junqueira et al. 2013) and that the total density profile across a spiral arm is symmetrical, despite an offset between the gas and stellar components of a spiral arm (Hou and Han 2015). As discussed, disagreements still exist on the structure of the Milky Way and whether it is a two-arm or four-arm spiral (Vallée 2014). Further disputes surround the nature of the local Orion arm and whether it is a true arm or can be considered a transient spur (Hou et al. 2009). Therefore, we create a further five models to test against our data (two armed, Scutum and Perseus, or Sagittarius and Norma), three armed (excluding either the Sagittarius or Scutum arm), and four armed (with and without the Orion spur) (Supplementary Data 1). We treat the Orion spur as a true arm and locate it between the Sagittarius and Perseus arms. To determine the most parsimonious galactic model we use direct cross correlation analysis between the grand design model and zircon isotopic time series (Davis 1986). Additionally, we compare the various time series with a more complex model of the Milky Way based on atomic hydrogen tomography (Jardine 2023).

We begin by considering detrital zircon concordia age distributions over time using the compilation of Puetz et al. (2021). We filter the data to include only those analyses with high degrees of concordance (those specified as groups 1, 2 and 3 by Puetz et al. (2021)); this broadly equates to analyses with less than 10% discordance. We analyse a 2.6–4.1 Ga temporal window as we wish to seek for periodicity within the Earth prior to widespread and demonstrable plate-tectonic processes, when the impact flux may have been higher (Nesvorný et al. 2023). We do not extend the temporal window further into the Hadean as there is a general lack of data, which makes drawing meaningful conclusions difficult. The frequency of zircon concordia ages is counted within 50 Myr bins, although using 10 Myr does not result in different conclusions.

We calculated a periodogram (Schulz and Mudelsee 2002) with an oversampling factor of seven, which does not alter spectral resolution (i.e. the ability to distinguish two peaks in the periodogram) but does enhance the ability to estimate the location of a single peak. To reduce noise, we applied overlapped segment averaging, in which the time series was split into two segments, overlapped by 50%, and averaged (Welch 1967). We also obtain a theoretical red-noise spectrum via the method described by Schulz and Mudelsee (2002) and use the $χ$²-probability distribution to calculate a 95% critical level, which denotes the maximum spectral amplitude expected if the time series was generated by a first-order autoregressive process. Spectral analysis of the zircon time series, via this method, highlights frequency components above the specified critical level (Fig. 2). Isotopic time series are given in Supplementary Data 1.

We also compare the various relative Solar System mass distribution models with zircon $ε$Hf, which reflects a proxy for the average crustal differentiation state of the Earth, within any given temporal bin (Fig. 1). Specifically, we compare fits between relative mass density of the Solar System through its galactic orbit with juvenile zircon Hf change point probability. Change point probability was constructed through applying a cumulative sum mean shift model (e.g. Van Kranendonk et al. 2015) to the zircon Hf data compilation of Cawood et al. (2022). The change point time series is provided in the figshare data repository (see Kirkland 2023). Juvenile zircon Hf change points reflect periods of time when the average composition of zircon-bearing magmas on a global scale stepped towards more radiogenic values, indicative of a greater load of Lu within the primary magmatic source, consistent with greater new crust production from the mantle.

We also compare the different grand design models with zircon oxygen isotopes (Valley 2003). Zircon oxygen isotopes track the broad compositional state of the magma in which they formed. For example, a $δ$¹⁸O value of 5.3 per mil (‰) indicates crystallization from a magma derived from Earth's mantle, which has a relatively uniform $δ$¹⁸O value owing to its long-term isolation from surface processes (Bindeman et al. 2022). On the other hand, the $δ$¹⁸O value of most continental crustal rocks is typically >5 and <10‰ (Liebmann et al. 2023). The higher (isotopically heavier) $δ$¹⁸O value of zircon derived from such crustal magmas may reflect melting of subducted supracrustal material that has seen near-surface fractionation. Lower (lighter) zircon $δ$¹⁸O values imply direct interaction with meteoric or surface-derived fluids or assimilation of hydrothermally altered crust. Hence, oxygen isotope deviations over time are posited to reflect a proxy for depth of melting and inferred to track impact flux (Johnson et al. 2022), perhaps on a galactically influenced frequency (Kirkland et al. 2022). Here we evaluate the distribution structure of zircon oxygen isotopes between 3.3 and 4.1 Ga, as quantified using the skewness metric on a moving 50 Myr temporal bin (Supplementary Data 1) as calculated by Kirkland et al. (2024), using the zircon oxygen compilation presented by Spencer et al. (2022). This time series provides a measure of the asymmetry of a distribution. We reverse scale this metric to highlight periods of time with greater numbers of isotopically light values (negative skewness), which may indicate a greater relative proportion of shallow melting on the planet. We calculate the orbital position for each temporal skewness measure assuming a galactic period of 187 Ma⁻¹ (Gillman and Erenler 2019; Kirkland et al. 2022). This time series of zircon oxygen isotopes is also expressed as the degree of population normality within each temporal bin, which we refer to as zircon oxygen diversity (Fig. 1; see also Kirkland et al. 2024). The zircon oxygen diversity time series is provided via the figshare repository; Kirkland (2023).

In the 2.6–4.1 Ga range, spectral analysis of the number of concordant detrital zircon ages in 50 Myr bins yields two significant periodicities above the critical level. These frequencies are at 270 and 190 Ma⁻¹ (Fig. 2). However, the domination of the number of 2.7 Ga detrital grains in the record means that any further analysis of this time series with respect to the various galactic models is strongly biased toward relationships with the Norma arm, in effect rendering it less suitable to evaluate potential fits.

Nonetheless, direct cross correlation analysis (Davis 1986) can be attempted with geochemical signals in zircon that track compositional averages rather than the direct numbers of grains (Fig. 3). Calculation of cross correlation values for zircon juvenile Hf change points and oxygen diversity, on evenly sampled probability estimates, against the relative Solar System density grand design models, is tabulated in Supplementary Data 1. P values are computed by a t test with n – 2 degrees of freedom. Positive correlations with P < 0.05 are indicated for zircon juvenile Hf change points and the four-armed and the Sagittarius and Norma two-armed models, although the strongest correlation is with the four-armed model. Oxygen isotope diversity, in terms of degree of non-normal distribution, shows statistically significant correlations (P < 0.05) for four-armed, four-armed with Orion and two-armed plus Orion (Sagittarius, Perseus and Orion) models (Fig. 3). Oxygen isotope distribution skewness also increases within the Scutum–Centaurus, Sagittarius and Perseus arms (Fig. 4).

It is well established that zircon age distributions are heterogeneously distributed over time, with various interpretations of this pattern related either to disparities in continental crust production or reworking and/or the influence of the supercontinent cycle (e.g. crustal preservation effects) (Condie 2013). Furthermore, detrital zircon frequency through time will inevitably be subject to preservation and sampling biasing effects (Mulder and Cawood 2021). Nonetheless, here we have identified frequencies, above noise at c. 270 and 190 Ma⁻¹, in a recent comprehensive detrital zircon compilation, that correspond to cycles recognized by other workers using different datasets and spectral analysis tools. For example, Prokoph and Puetz (2015) noted 280 and 180 Ma⁻¹ frequencies, among others, in igneous and detrital zircon datasets. Puetz and Condie (2019) also highlighted a 180 Ma⁻¹ frequency, associating it with the occurrence of superchrons (the period of stable magnetic field) and/or the geomagnetic reversal rate. Frequencies in this range are potentially interesting from an exogenic perspective as they match frequencies in astronomical processes (Gardner et al. 2011; Gillman and Erenler 2019; Kirkland et al. 2022). Specifically, transit through the galactic arms is estimated on a frequency of 180–200 Ma⁻¹ (Gillman and Erenler 2019; Kirkland et al. 2022). Periodicity within Hf and oxygen isotopes has been previously discussed (Kirkland et al. 2022); suffice it to say that frequencies within this range from these isotopic systems have also been equated with long period astronomical forcing via modifications to the inner planet impact flux, itself probably influenced by the galactic tide and spiral arm crossings (Gillman and Erenler 2016, 2019; Gillman et al. 2019; Kirkland et al. 2022). One mechanism of imparting an exogenic periodicity on zircon frequency, which ultimately tracks differentiation processes, would be via impact-related decompression melting and crustal preservation, where more buoyant felsic crust, produced in the wake of an increased impact flux, was preferentially preserved relative to less buoyant, and more likely to be reworked, mafic proto-crust. Arguably, such a model is consistent with greater oxygen isotope diversity as such buoyant crust would have a greater chance to interact with the near-surface environment where oxygen isotope fractionations are large. Furthermore, a wide range of other geodynamic consequences has also been predicted from impacting, including planet-wide tectonic behaviour shifts akin to transient subduction (O'Neill et al. 2019).

In any case, strong positive relative correlation is indicated between zircon juvenile Hf change points and oxygen isotope diversity for the four-armed spiral model (Fig. 3). If enhanced terrestrial impact flux is ultimately driven by Solar System galactic transit, and this manifests itself as more crust production and reworking, then the grand design four-arm model would seem to satisfactorily account for a significant component in the Archean isotopic (but not necessarily absolute zircon numbers) time series variability.

More complex models for the Milky Way galaxy have been proposed based on atomic hydrogen data, which are able to resolve structures that are obscured owing to dust in other frequencies (Kalberla et al. 2005). A composite Milky Way model derived from the Leiden–Argentine–Bonn velocity surveys of atomic hydrogen, corrected where possible using parallax measurements for star formation regions (Jardine 2023), reveals apparently good fits between zircon juvenile Hf change points and regions of enhanced stellar density (Fig. 5). Crustal nuclei production has been linked by some to giant impact events on Earth (Glikson 2013; Johnson et al. 2019; Kirkland et al. 2022). Early juvenile crust production episodes, when average crustal magma compositions shifted towards more mantle-like values, on ancient cratons may therefore show a relationship to impact flux (Kirkland et al. 2022). To explore this concept further, we consider the first juvenile Hf change point across several cratons with the atomic hydrogen-based model of the Milky Way. The first juvenile change point on seven of nine Archean cratons examined sits within mapped spiral arm regions or zones of dense molecular clouds (Figs 4 and 5). Whereas much crust production, at least from the Proterozoic, may be associated with tectonic plate reorganization, the initial juvenile crust production change point for the Zimbabwae Craton (3361 Ma) corresponds to passage in the Norma arm, and to the passage of the Slave (3422 Ma), Wyoming (3500 Ma), Kaapvaal (3511 Ga) and North China (3817 Ma) cratons in the Perseus arm. The Singhbhum Craton's (3661 Ma) first juvenile shift occurs during Solar System transit through an area of high atomic hydrogen that creates a bridge between the Orion spur and the Perseus arm, which runs toward the origin of the Cygnet spur (Fig. 5). The San Francisco (3211 Ma) and Superior (3236 Ma) cratons would conventionally be within the Scutum–Centaurus arm, although on the atomic hydrogen tomographic model they are outside some preferred fits of this arm (Fig. 4). However, at least the San Francisco Craton's change point sits on an ionized atomic hydrogen high and is within a zone of star-forming complexes (Fig. 5), as established from a multi-wavelength study of H II regions, diffuse ionized gas, molecular clouds and OB stars (Russeil 2003).

Although we have included the Orion spur as a third or fifth arm in some of the models it should be noted that spurs, bridges or filaments are generally accepted as transient features and not part of the quasistatic spiral structure (Pérez-Villegas et al. 2015). Instead, these structures are thought to be Giant Molecular Clouds (GMCs) that move into and out of the arms, in the same manner as the Solar System (Dobbs 2013). GMCs are sheared during their exit from an arm owing to the differential rotation in the galactic disc (Miyamoto et al. 2014). This creates elongated filaments, known as Giant Molecular Filaments (GMFs), which appear orthogonal to the spiral arms (Ragan et al. 2014). Observations of GMFs show that they are typically hundreds of parsecs in size with a wavelength separation of approximately three times their size (Ragan et al. 2014). Owing to the nature of spurs, they are not static in reference to the arms and can instead be considered as co-moving traffic with the Solar System. Although spurs, bridges or filaments probably play a role in Solar System perturbations, albeit less significant, it is extremely difficult to translate their transient structure into a time series model as they are not an intrinsic part of the spiral arm structure.

Recent findings regarding the inner and outer arms of the Milky Way offer a potential resolution to the continuing debate surrounding the number of spiral arms (Xu et al. 2023). In their study, Xu et al. (2023) demonstrated that, like other spiral galaxies, the Milky Way exhibits a dual structure consisting of both two primary inner arms and four secondary outer arms. Specifically, they favoured an inner region of the Milky Way characterized by two prominent spiral arms that bifurcate at c. 6.5 kpc from the galactic centre, resulting in the formation of four outer arms. Given the estimated position of the Solar System (8 kpc from the galactic centre, with a potential radial oscillation of c. ± 0.5 kpc), it appears that our Solar System resides just outside the bifurcation where the inner arms split. This spatial relationship, combined with the understanding that the bifurcation of spiral arms evolves as galaxies mature and spiral arms wind up (Wada et al. 2011), suggests that both the two-arm and four-arm models may hold validity at distinct stages in the evolutionary phase of the Solar System.

A key question is what do any correlations between terrestrial isotopic time series (juvenile Hf crust production and diversity in oxygen isotopic signature in magmas) and astronomical structure relate to in terms of terrestrial process. Could they in some way reflect a periodicity established by the galactic tide on the mantle and/or could they reflect impacting? Major impact episodes on Earth are inferred from isotope systematics and correlation to the Moon at c. 3.8 Ga and from terrestrial rocks (and spherules) dated at 3.46 Ga (spherules, Pilbara), 3.25–3.22 Ga (spherule clusters, Kaapvaal and Pilbara cratons), 2.63–2.48 Ga (spherule clusters, Kaapvaal and Pilbara cratons), 2.23 Ga (Yarrabubba crater), 2.02 Ga (Vredefort crater), 1.85 Ga (Sudbury crater), 0.58 Ga (Acraman crater), 0.368–0.359 Ga (late Devonian), 0.214 Ga (late Triassic), 0.145–0.142 Ga (late Jurassic), 0.065 Ga (K–T boundary) and 0.0357 Ga (late Eocene) (Glikson et al. 2016). Many of these events show good correspondence to positions with the galactic arms of the grand design four-arm model or an even better fit when the H I based galactic model is considered (Figs 4 and 5).

An increase in impact flux is a very direct way of enhancing juvenile crust production and a link to the galactic environment has been suggested based on the disturbance of Oort Cloud material with a reduction in comet perihelion towards Earth-crossing orbits (Kirkland et al. 2022). Nonetheless, impacts can lead to secondary or tertiary processes that may, ultimately, be recorded within the magmatic zircon record (Kring and Cohen 2002; Osinski et al. 2020; Schmieder and Kring 2020). In other words, the investigated isotopic time series may not be a direct function of impacting but rather its longer lasting effects in the crust; that is, a response to secondary melt fertility or preservation-related processes. For example, greater plume magmatism has been proposed to have been triggered by impacting on Earth (Jones 2005) and Mars (Reese et al. 2004). Whereas decompression melting within the sub-crater mantle may be almost instantaneous, such large-scale melting may induce long-lived mantle upwelling (Ubide et al. 2017). Furthermore, impact sites with their shallow melt pools and deeper upwelling decompression plumes would generate buoyant crustal nuclei that themselves would have greater preservation potential. These nuclei, and the accreted products of other crustal differentiation processes, would thus be more likely to be preserved through crustal recycling than non-impacted oceanic lithosphere. However, establishing any connection between decompression–plume magmatism and impacting is challenging, as a diversity of processes can result in upwelling mantle (Foulger 2002). For example, a database for ages (3.4–2.0 Ga) of continental large igneous provinces (LIPs; Liebmann et al. 2022), detrended and fitted to counts per 10° Solar System arc, shows little obvious correlation with galactic spiral arm models, despite having a c. 170 Ma⁻¹ periodicity (Prokoph et al. 2004) (Supplementary Data 1; Fig. 4). Nonetheless, there do appear to be some LIPs that correspond in timing to impact events (Fig. 4). This observation implies that the deep time LIP record dominantly reflects heterogeneous endogenic driving processes, which obscure any exogenic signal related to spiral arm transits. Furthermore, Jones et al. (2003) provided evidence to support the idea of Green (1972) that significant degrees of partial melting or even complete melting of mantle peridotite can occur as a result of a substantial impact into oceanic crust. This concept may support a process where some komatiites (rocks with MgO content exceeding 18 wt%) formed through impacting. Based on a compilation of percentage komatiite in total volcanic rocks through time (Condie and O'Neill 2010), higher komatiite proportions erupted at c. 3.5, 3.2 and 2.7 Ga, which has some apparent temporal relationship to the Perseus arm (atomic hydrogen dense region), Norma arm (entry) and Scutum–Centaurus arm (exit), respectively (Supplementary Data 1). Moreover, all three of these orbital positions correspond to spherule or crater evidence of impacting on Earth (Fig. 4).

A potentially clearer correlation in terrestrial time series data and the four-arm grand design model is evident for zircon oxygen isotope diversity (Fig. 4). A similar general probability pattern is also seen for impact-melt clasts in lunar meteorites (Cohen et al. 2000; Daubar et al. 2002; Fernandes et al. 2013), where increased probability is apparent towards each of the four spiral arms. Furthermore, increases in lunar impact-melt probability are not restricted to a single galactic orbit, but occur over multiple orbits through the arms. We consider these observations important as they suggest that the c. 200 Ma⁻¹ periodicity is not simply an endogenic terrestrial process, related to internally mediated early plate tectonic subduction or convective rhythms in the mantle, but rather is forced by external processes to our planet, as transient subduction may follow impacts (O'Neill et al. 2017).

Other periodic geological time series have been linked to impacts on the planet, perhaps one of the most contentious being periods of mass extinction (Bailer-Jones 2009; Gillman et al. 2019). Five major mass extinctions are recognized on the basis that they had the most significant impact on global biodiversity, resulting in the widespread loss of species across various ecosystems. These events were marked by extremely high extinction rates, leading to a substantial reduction in the diversity and abundance of life forms. These extinctions are End-Ordovician (around 443 Myr ago), Late Devonian (around 359 Myr ago), End-Permian (around 252 Myr ago), End-Triassic (around 201 Myr ago) and Cretaceous–Paleogene (around 66 Myr ago). Although there are numerous factors linked to mass extinctions and a single cause has been shown as unlikely for at least the Cretaceous–Paleogene major mass extinction (Bond and Grasby 2017), the correlation with entry and exit of spiral arms (four-arm grand design model) appears rather stark for a periodicity of c. 200 Ma⁻¹, when considered over the Solar System's orbit of Sagittarius A* (Fig. 4). Numerous arguments for and against periodicity in mass extinction, and indeed fossil diversity (Rohde and Muller 2005), have been put forward; nevertheless, it is perhaps relevant to highlight that Goncharov and Orlov (2003) claimed a periodicity of 183 ± 3 Ma⁻¹ for 13 mass extinction events, and Omerbashich (2006) found a similar c. 194 Ma⁻¹ period in genera diversity, matching periodicities in isotopic records of crustal evolution (Kirkland et al. 2022). It can therefore be argued that the more recent biological record of our planet has a similar periodicity to that seen within the deep time isotopic record, as preserved by zircon crystals.

Meteorites come from a variety of sources, with the majority derived from rocky objects orbiting the Sun primarily in the asteroid belt between Mars and Jupiter. The prevailing model of meteorite delivery to Earth implies an episodic cascading process in which break-up of large asteroids produces new populations of fragments that in turn supply the inner Solar System over extended periods of time. Nonetheless, others have argued for a more constant flux (Terfelt and Schmitz 2021). Some meteorites may originate from the Moon or Mars, ejected during impact events on those bodies (Greenwood et al. 2020). Another source is comets, originating from the outer reaches of the Solar System (Vannucchi et al. 2015). It is this latter source that would be more periodic given its reduced gravitational binding to our Sun, and susceptibility to external perturbations, driven by the galactic tide and interactions with stars during the galactic orbit (Matese et al. 1995; Gardner et al. 2011). Models of relative mass density of the Solar System during its galactic orbit have been proposed to approximate periodic Oort cloud comet impact flux on Earth on a c. 200 Ma⁻¹ frequency (Gillman et al. 2019; Kirkland et al. 2022). By fragmenting a single comet into a debris field during passage near the Sun, the impact flux of a fraction of long period comets can be increased (Siraj and Loeb 2021), enhancing the probability of high-velocity collisions with terrestrial planets. Near Earth object meteorite flux would be less periodic, higher frequency, but result in lower energy impacts with less melting than a comet impact, given orbital velocity considerations (Hughes and Williams 2000; Weissman 2006).

The greater variety of crustal reprocessing on a planet with horizontal tectonics means that there will be a more favourable temporal window, between Moon-forming collisions and the commencement of widespread subduction-related plate tectonics, when impact signals may be detected (e.g. <3.8 Ga and >3.2 Ga). If larger impacts are comet related, but more infrequent (Morrison 2006), they would probably have a more periodic flux given the Solar System's radial motion (influencing the galactic tide) and the galactic period (influencing stellar encounters), both of which have been calculated to have a significant effect on dislodging material from the Oort cloud (Gardner et al. 2011). Although on the more recent planet certain biological and impact records (e.g. cratering) may have a similar periodicity to deep time records, geological processes on our tectonically active planet will relatively rapidly (in comparison with refractory mineral isotopic time series) work to expunge that surficial record. This means that, arguably more prima facie evidence of impacting (e.g. craters, shocked quartz and zircon, Cr-rich chondritic spinels, Ni-rich chromites with high Co, V, Zn, platinum group element anomalies, including enrichment in refractory Ir and Ru and depletion in volatiles, detrital chrome-spinel grains originated from micrometeorites) may be biased to more frequent, aperiodic, lower energy, asteroid belt derived collisions. In contrast, isotopic records of comet impact crust production, reflecting major fractionation events in the crust, appear more likely to be periodic. Moreover, although four-arm models provide a simplified approximation of our galaxy, greater complexity is highly likely, consistent with atomic hydrogen observations, which somewhat relaxes the statistical requirements of geological time series to exactly match periodic grand design models for the galaxy's structure. Nonetheless, the geological record, if influenced by exogenic processes, perhaps allows greater nuance in galactic models. In any case, there appear to be strong grounds to suggest that correlating what is, relatively, temporally short (e.g. extinction, impact cratering) terrestrial records into deep time isotopic records will better resolve periodicities, timescales and hence ultimately understanding of any link between catastrophic processes and the modification of Earth's lithosphere and its biological cargo.

The prevailing model for our galaxy is that it rotates around Sagittarius A*, the central supermassive black hole (Fig. 5). Electromagnetic information transfer through this structure has traditionally been regarded as impossible, although quantum entanglement may provide some insight (Hawking 1976; Giddings 1994; Almheiri et al. 2019; 2021). Gravitational lensing, owing to massive objects such as black holes, amplifies light, helping observation and distance measurement of the farthest reaches of the Milky Way (Bozza and Mancini 2004; 2012). In any case, it is the gravitational influence of Sagittarius A* that is much more wide-reaching than the black hole's event horizon itself, creating and sustaining the central bulge and bar, heaping a higher density of stars and dust towards the galactic central region. Hence, for all practical purposes, as observations are significantly obscured towards the galactic centre, our only recourse to understand the far side of Sagittarius A* may be by rocky planet geology, which could preserve in their deep time geological records, a cryptic, undoubtedly biased, legacy of information left when our Solar System was on the other side of the central black hole.

We thank two anonymous reviewers for comments that improved this contribution. We also thank D. Gasser for editorial handling.

CLK: conceptualization (lead), formal analysis (lead), investigation (lead), writing – original draft (lead), writing – review & editing (lead); PS: formal analysis (supporting), writing – review & editing (supporting).

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

All data generated or analysed during this study are included in this published article and its supplementary information files.