Reconstruction Methods

There are several types of data or information that can be used to reconstruct the continents and these data can generally be divided into quantitative (one can with ease calculate a pole of rotation, i.e. an Euler pole) and semi-quantitative methods.  Below we list the most common methods; some of these are dealt with in detail later.








A widely used method of reconstructing plates relative to a fixed mesosphere utilises linear chains of volcanoes that display age progression and are thought to have been caused by focused spots of melting in the upper mantle assumed to be fixed relative to each other over geologically long periods of time (fixed-hotspot hypothesis).  


From Early Cretaceous times (c. 130 Ma), hotspot reconstructions offer a unique contribution to palaeoreconstructions since they provide palaeolongitudinal control, otherwise absent in palaeomagnetic data, for plate positions.  A fundamental assumption with hotspot reconstructions is that the hotspots are stationary, or move at insignificant speeds, relative to plate-tectonic velocities.  Based on dated volcanic hotspot tracks from the North American, South American, African, and Indian-Australian plates, Müller et al. (1993) estimated best-fit plate rotations relative to present-day hotspots since the Early Cretaceous (M10, c. 130 Ma).  In EasyGmap we use this hotspot frame.  Although hotspot reconstructions can be used back to 130 Ma it is recommended that they are only used for the last 90-95 million years (Torsvik et al. 2001a).



  1. Volcano chain geometry determines plate movement/direction

  2. Plate velocity can be calculated if volcano tracks are dated

  3. Oldest hotspot related volcanoes are c. 130 Ma

  4. Hotspot reconstructions are absolute (latitude and longitude) assuming that hotspots are stationary with respect to the mantle (prior to 84-90 Ma hotspots appears stationary, but for older times hotspot movements between the Indo-Atlantic and Pacific hotspots are indicated by the available data)



Magnetic anomalies (sea-floor spreading)


Magnetic anomaly data, globally applicable from mid-Jurassic times (c. 175 Ma), provide relative fits between continents.  Several and different digital versions of magnetic anomalies are available for the North Atlantic region.  We have compared several of these data-sets with the most recent magnetic anomaly map of Norway and adjacent areas (Olesen et al. 1997 ), and found that a digital data-set used by Skogseid and co-workers (e.g. Skogseid et al. 2000) is the best and most detailed.  Normal polarity anomalies (5-8,13,18,20-24) from the North Atlantic are displayed in EasyGmap, and ages, based on the time-scale of Cande & Kent (1995) are listed in Fig. 3.  The cruder, world magnetic anomaly data-set is that of Müller et al (1997)


Magnetic anomaly fits have been used extensively for North Atlantic reconstructions, beginning with anomaly A33old (c. 80 Ma) and A24 (52.364-53.347 Ma), the oldest normal polarity chrons identified in the Labrador Sea and the NE Atlantic, respectively.  For older times, magnetic anomaly fits must be considered minimum fits since they do not account for pre-drift extension.  For example, the majority of Pangea  reconstructions are Late Triassic/Early Jurassic or younger reconstructions and assume insignificant intra-plate deformation since the Permian.  Lottes & Rowley (1990) incorporated some Late Triassic-to-younger intra-plate deformation and the seafloor spreading history in their Pangea reconstructions but domains of Late Palaeozoic and Early Mesozoic extensional deformation in the North Atlantic, including the North Sea, East Greenland and the Barents Shelf, remained unincorporated in their work.  In EasyGmap we have attempted to account for pre-drift intra-plate deformation and Euler poles for initial opening fits, interpolated stage-poles and magnetic anomaly fits in the North Atlantic are detailed in Torsvik et al (2001b)



  1. Oldest magnetic anomaly is c. 175 Ma (’preserved’ seafloor)

  2. Superimposing magnetic anomalies with the same age across a spreading ridge is a powerful reconstruction tool

  3. Magnetic anomaly reconstructions are relative

  4. We can calculate spreading velocities between two plates




Palaeomagnetic Data

Under ideal circumstances, when rocks are formed, they acquire a remanent (permanent) magnetization that records a declination (angle with respect to the Greenwich meridian) and an inclination (angle with the horizontal plane) that copies the local Earth magnetic field at that site.  The inclination varies with latitude and is the main feature of interest in palaeomagnetic reconstructions.  Palaeomagnetic data therefore yield palaeolatitudal and rotational constraints for continents throughout geological history. In EasyGmap we use a master apparent polar wander path for North America that is modified from Torsvik et al. (1996) for most of the Silurian and Devonian, and Torsvik et al. (2001b) for Carboniferous (300 Ma) and younger times.




Geometric matching of continental borders

Several reconstructions of the Atlantic (prior to rifting and seafloor spreading) exist in the literature, but the Bullard et al. (1965) fit, based on least-square fitting of 500 fathom (c. 900 meter) contours across the Atlantic, is the most well-known.  This reconstruction is superior to all others in matching North American and European paleopoles from mid-Palaeozoic to Late Triassic times (Torsvik et al. 2001b).




Matching of stratigraphic sections and crustal provinces

The primary geometry and continental fits for the Neoproterozoic Rodinia Supercontinent are to a large extent governed by an attempt to match the 1100-1300 Ma Grenvillian-Sveconorwegian-Kibaran crustal provinces; this alone does not provide a unique solution since crustal provinces can either be interpreted as having formed continuous belts or as conjugate margins.




The faunal principles have not changed since they were first set out by Cocks & Fortey (1982, fig. 1). Before any animal can be used, both its age and its individual ecology must be assessed correctly.


The distribution of those animals with a planktonic, pelagic or nektonic (swimming) lifestyle is controlled by oceanic currents and temperature and thus, although the deposits in which their fossils are found usually form part of one or more of the terranes which we recognise as separate, they are of no significance when assessing the closeness or individuality of such terranes. However, since the great majority of such animals are dependant on temperature, then their occurrence may  be generally correlated with the latitude in which the animals lived (Cocks & Verniers 2000). In the Lower Palaeozoic, these planktonic groups are best represented by graptolites, a minority of the trilobites such as the pelagic Carolinites, cephalopods, chitinozoa, acritarchs and apparently conodonts (the detailed ecology of which are still poorly understood). Since the dispersal of these animals was very often rapid, the quickly-evolving members of these groups include most of the best fossils for international correlation.


In contrast, those animals with a benthic lifestyle, which were and are confined to the sea-floor for their adult life, such as brachiopods, most trilobites, bivalves, gastropods and most ostracods, and which were also temperature dependant.  These may be divided into two groups; the majority, which lived in shallower-water seas on the continental margins, and which were therefore both latitudinally-related and also confined to particular terranes; and a smaller number which lived below the thermocline and which were thus independent of palaeolatitude, and which were distributed on the deeper parts of continental shelves and on the ocean floors.  It is the former, larger, group of benthic animals upon which we rely most strongly to provide the faunal support for the terrane reconstructions presented here.


Despite the fact that the adult brachiopods and most trilobites were confined to relatively small sites, their larval stage was planktonic for shorter or longer periods and thus the genera dispersed as time progressed. As a working rule of thumb, although different animals have and had very different dispersal rates (McKerrow & Cocks 1986), many faunas appear to be separable if the oceanic width is above 1000km. If two terranes are at the same latitude, then the composition of the benthos will be largely the same if the terranes are close to one another. However, if the terranes drift apart, then the larvae of the descendant species of the original benthos will not cross the intervening deeper ocean after a certain period of time: the discriminating palaeontologist will thus be able to identify the two separate terranes as different and separate. Comparably, if two terranes at the same latitude but with different benthos drift towards each other, then the two terranes recognised by the palaeontologist in the earlier period will merge into a single faunal province in the later period, but without providing any certain evidence that the two terranes actually collided. In addition, the largest palaeocontinents, such as Gondwana in our time period, are so substantial that they cover many degrees of latitude, and thus the benthos at the northern and southern extremes of the continent can be quite different. However, between these two palaeolatitudinal extremes the faunas at intermediate latitude should show a gradation or cline, comparable to the cline seen today in the benthic molluscan faunas along the western seaboard of North America, which stretches from the tropics of Panama to the high latitudes of Alaska. This is allied to another principle; that the lower the latitude, the larger the number of different species and genera which will be found, i.e. the bio-diversity is generally greater (in equivalent ecological situations) the closer one gets to the Equator.