1 Introduction

Continental reconstruction became, notwithstanding some earlier fanciful attempts, a serious geological problem after the advent of Wegener's theory of continental drift and some very successful reconstructions of the peri-Atlantic continents were made before the rise of plate tectonics (e.g. Wegener 1922; Argand 1924; DuToit 1927, 1937; Choubert 1935; Carey 1958). Plate tectonics made not only a detailed, quantitative reconstruction of the continental positions possible but also the establishment of displacement paths of individual continents with respect to one another, largely thanks to the presence of ocean-floor magnetic anomalies and fracture zone trends in places via plate circuit constructions. However, this method could be applied only as far back in time as the age of the oldest magnetic anomalies on extant ocean floors, which is the medial Jurassic (on the basis of the generally accepted M25 anomaly, which is late Oxfordian, or the suggestion by Greene et al. 2017, to identify M42, which is about 168.5 Ma old, i.e. early Bathonian). For reconstructions earlier than the medial Jurassic, the geologist has so far had three main methods to fix the positions of continents in the past (Fig. 1): palaeomagnetism (Fig. 1A, B), palaeoclimatology (Fig. 1C, D) and palaeobiogeography (Fig. 2).

Fig. 1
figure 1

Two of the most widely used methods in making continental reconstructions in the past. A and B illustrate the properties of the present magnetic field of the earth relevant to using palaeomagnetism for continental reconstructions. A shows that the magnetic and geographic poles are not coincident, and B shows a map of declinations for 2010 (from http://geokov.com/education/magnetic-declination-inclination.aspx). Since the magnetic poles wander, so do the lines of equal declination. C and D are maps of precipitation (from https://earthobservatory.nasa.gov/global-maps/TRMM_3B43M) and temperature distribution today on the face of the earth. In D1 (from https://blogs.egu.eu/divisions/ts/tag/temperature/), AC is Alaska Current, HC Humboldt Current, GS Gulf Stream, CC Canaries Current, BC Benguela Current, T Tibetan Plateau. D2 is a satellite image of a small part of the Gulf Stream, showing its turbid, meandering nature. Note that in all cases, the boundaries are so blurred that using any of these methods gives a minimum uncertainty of 1000 km for palaeomagnetism and several thousand km for palaeoclimatology

Full size image
Fig. 2
figure 2

A highly simplified map showing the Kipchak Arc and the locations of faunal assemblages that Fortey and Cocks (2003, Fig. 15) think to justify placing the Chu-Ili, Chingiz and Tien Shan areas close to Gondwana-Land as independent continental entities. The map shows our alternative suggestion of placing them all along the Kipchak Arc, as originally suggested by Şengör et al. (1993). Kipcpak locations are equally compatible with the faunal assemblages used by Fortey and Cocks (2003). That is why Şengör et al. (2014b, 2018) kept them as parts of the Kipchak Arc in contrast to the suggestion by Fortey and Cocks (2003)

Full size image

2 Commonly used tools in continental reconstructions

Palaeomagnetism proved itself a crucial quantitative tool when the continental drift and sea-floor spreading concepts needed evidence for the mobility of Earth's lithospheric plates before the medial Jurassic. Although this method has the ability to reveal relative motions between plates based on the geocentric axial dipole (GAD) hypothesis, the wandering of geomagnetic poles in a small wobbling path caused the migration of the south magnetic pole about 10–15 km per year only during the past century. These kinds of palaeosecular variations can partially be neutralised by taking the average of the geomagnetic field over 105 or 106 years, which are obtained from palaeomagnetic sites (Tauxe 2010). However, cumulative errors during the collection and measuring steps, undetected inclination shallowing in sedimentary rocks, the uncertainty of palaeohorizontal in magmatic rocks combined with palaeosecular variations, geomagnetic field reversals and superchrons can create errors of the order of 5° or even more in direction, which have resulted in at least 500 km displacement errors in palaeogeographic reconstructions (Tarling 1988). On top of that, palaeomagnetic methods cannot determine palaeolongitudes and the constraints provided by the marine magnetic anomalies are absent prior to the medial Jurassic (Van der Voo 1987). Finally, the widely used apparent polar wander paths suffer from the averaged data sets that contain palaeomagnetic poles with questionable reliability (Vaes et al. 2023). Despite these shortcomings, palaeomagnetism is a powerful tool for establishing the position of a continent along a meridian with an error margin of 500–1000 km and its orientation with respect to the magnetic North Pole. But it is not possible to tell how far away from one another two continents may have been at any given time along the magnetic parallels. Only marine magnetic anomalies can inform us about the amount of separation between two continents along magnetic parallels provided no substantial subduction has occurred along either of the separated continents. Most of the Atlantic Ocean provides a good example of such a case. Construction of hypothetical synthetic oceanic isochrons to track the opening of the Neo-Tethys as done by Stampfli and Borel (2002) is doomed to failure because of the lack of knowledge about the amount and direction of subduction and the uncertainty of the pre-collisional geometries of the continental margins to establish the shapes of now-vanished oceans.

Palaeoclimatology is a much less precise tool (Fig. 1C, D). Even if one assumes fixed continents throughout geological time, the boundaries of the present-day climatic zones are so diffuse that not much precision below a few 1000 km can be hoped for. World climate is based on a variety of orbital parameters and the distribution of oceans and continents on the earth's surface, plus elevation or depth below sea level. In other words, one must know the continental distribution and the terrestrial relief to be able to erect a halfway reliable climate model. The frequent failure of the climate models even for the present-day earth is a dire warning against excessive confidence in palaeoclimates for continental reconstructions.

Palaeobiogeography is useful to indicate the places of barriers and migration paths for organisms. It is very reliable if fossil collections of relevant organisms are comprehensive and abundant in areas of interest and if the barriers separating them are extensive (whole oceans, major mountain systems). With inadequate fossil collections and in case of inefficient barriers, it is much less reliable. Where reliable, it is a useful tool to test palaeomagnetic observations and palaeoclimatological inferences. Figure 2 shows a case of equivocal indications provided by fossil assemblages.

Another source of error in palaeobiogeographic considerations, when used for reconstruction purposes, is the difference between 'total’ or ‘flush suturing' (Fig. 3A), ‘suturing via transform faulting' (Fig. 3B), 'contact via island arcs' (Fig. 3C) and 'point contact' between two continents (Fig. 3D). In all these cases, exchange of terrestrial fauna between the three continents seen in Fig. 3E will be possible. Their conjunction will form a barrier for the marine faunas, but the tectonic significance of all three cases is entirely different from one another. Case B cannot lead us to believe that a total suturing of the continents a, b and c occurred simply because the terrestrial faunas could go from one continent to the others. Despite this exchange, there is still a lot of ocean to be consumed between those continents after the onset of faunal exchange for terrestrial organisms and obstacles for marine organisms. The situation in Fig. 3A may be obtained by a head-on collision or through the cases illustrated in Fig. 3B, D.

Fig. 3
figure 3

Various scenarios of continental collision. A With flush sutures x–x′ and y–y′, that may or may not be the results of original head-on collisions. B Side collisions resulting from transform-fault-juxtaposition. C Indirect collisions via island arc bridges, D Point collisions. E Uncollided state of three continents a, b and c

Full size image

Moreover, biogeographic data are as yet largely limited to about the last 540 Ma, i.e. since the beginning of the Cambrian; hence, they are not applicable to Precambrian reconstructions (e.g. Şengör et al. 2020).

All three methods enabling reconstructions discussed above are much better at longitudinal than in latitudinal control. Notice that here we did not list structural geological observations on strain as constraints for continental reconstructions because they are almost never used. It is, in fact, the purpose of this paper to show how important they may be.

3 Detailed geology of orogenic belts separating continents and the help they provide for continental reconstructions

Figures 4 and 5 illustrate attempts at reconstructing the history of two major orogenic belts, namely the Altaids and the Saharides. Figure 4 displays the inferred history of the Altaids (Şengör et al. 2014a, b, 2018) and surrounding areas (Şengör and Natal'in 1996; Natal'in and Şengör 2005). The methods used in these reconstructions are presented in some detail in Şengör et al. (1993) and Şengör et al. (2014a, b, 2018). What the reader should notice in these figures (A–L) is the evolution of the Kipchak Arc. In frames A–C, there is not much that the three popular reconstruction methods mentioned above cannot deal with. Major motions are along the meridians, and continental continuity can be reliably established by fossils. But when we come to frames E–G, without the detailed strain observations, i.e. the progressive change of shape of the rock masses between the two continents and the cumulative displacements shown by them, the motions of the Russian and Siberian cratons along the parallels cannot be inferred. The same is true when we consider the evolution in frames I–L. We are thus obliged to do the structural geology of orogens soldering two previously separate cratons in great detail, down to a scale of 1/50 to 1/25 thousand, and to date the prominent steps in the structural evolution with high-resolution zircon dating and/or high-resolution biostratigraphy. In places, the geologist may have to work with smaller scales. In that case, detailed hypotheses should be erected about the strain evolution and ways must be indicated where further mapping in larger scales might falsify the original hypothesis.

Fig. 4
figure 4

Tectonic evolution of the Altaids taken from Şengör et al. (2014b, 2018). Heavy black arrows show the motion of the Russian Craton and some peripheral units with respect to Siberian Craton: A–L. M shows the black arrows in A–L in sequence coloured according to the International Stratigraphic Chart colours for their corresponding time intervals. In N, the displacement path revealed in M is smoothed

Full size image
Fig. 5
figure 5

Tectonic evolution of the Saharides from Şengör et al. (2020)

Full size image

Figure 4M shows the incremental motions of the Russian Craton with respect to the Siberian Craton. It is very rare that continents move with abrupt changes of direction, save for small escaping fragments as in Turkey or Central, East and Southeast Asia during continental collisions. Thus, we tried to smooth Russia's displacement path with respect to Siberia. Still, we had to accept two abrupt changes of direction in its motion: one during the medial Ordovician and the other during the early and late Permian transition. Curiously, both times coincide with rapid field reversals and accelerated sea-level rises. Both times represented preludes to widespread extinction events in the biosphere (medial and end-Permian and end-Ordovician). Since these latter observations are not unique to the cases considered here, we refrain from attaching any causal relationship between them and the times of rapid reversal of continental motion. What is important in Fig. 4M, N is that they show that it has been possible to infer the total displacement path of the Russian Craton with respect to the Siberian Craton.

The methodology outlined above can also be applied to Precambrian orogenic belts, provided familiar plate tectonic environments can be identified. Figure 5 illustrates the case of the Saharides, a late Precambrian to early Cambrian orogenic belt under the Sahara and in the Arabian–Nubian Shield (Şengör et al. 2020, 2021). In this case, the Tuareg island arc was detached from the Congo Craton sometime during the early Tonian (Fig. 5A), migrating towards the West African Craton (Fig. 5B) and initiating the Pharusian Orogeny towards the end of the Cryogenian. As the two cratons continued approaching one another, the Tuareg Arc became sliced and folded around vertical axes exactly as in the case of the Kipchak Arc in the Altaids. To document this stage, it was important to look at the structural geology of the remnants of the Tuareg arc both in the Arabian Peninsula and in the Sahara. We carefully mapped the trend lines (cf. Suess 1883) of the Saharides using the available field observations, Google Earth images and magnetic anomaly lineations. Palinspastic restoration of them was done using the former plate tectonic environments, what Şengör et al. (2021) called the essential units, such as magmatic arc fronts, subduction–accretion complexes, fore- and back-arc basins, flexural molasse basins, etc. These essential units now lie disrupted by syn- and post-orogenic structures before the geologist, dissolved into what Şengör et al. (2021) called the accidental units. Once the accidental units are returned to their original positions using mainly structural geology and stratigraphy, the trend lines automatically acquire their pre-collisional geometries. In the case of the Saharides, the convergence of the Congo Craton with the West African Craton parallel to subparallel with the latitudes is constrained by the length of the Tuareg arc, established by mapping out its trend lines and reconstructing them to their pre-collisional positions, after the collision of its 'tip' with the West African Craton during the Pharusian orogeny. There was indeed a non-negligible component of motion parallel with the lines of latitude, especially during the Ediacaran. This was indeed the time of maximum strike-slip activity within the Saharides (Fig. 5C).

4 Conclusions

Continental reconstructions cannot represent even an acceptable approximation to what may have been the case in the past unless the very large strains caused by diverse and large amounts of displacements in orogenic belts are taken into consideration. Although numerous large- and medium-scale structures have been reported from orogenic belts sandwiched between two continents none have been shown how far they influenced the displacement paths of the two continents with respect to one another. This is largely because the strain and displacement along them have not been adequately constrained, despite the fact that very large displacements along orogenic belts in Pacific-type continental margins have long been known from palaeomagnetic observations. Dewey (1975a, b, 1976, 2019) has repeatedly emphasised that plate tectonics destroy records and thus the geologist is robbed of the means of making complete reconstructions. Despite the loss of the former ocean floors before the medial Jurassic, detailed studies on the displacements and strains in orogenic belts seem to offer a possible way to cheat the almighty plate tectonics in its insatiable Moloch-like appetite for ocean floors.