New palaeopr oter ozoic palaeomagnetic data fr om Central and Northern Finland indicate a long-lived stable position for Fennoscandia

The Svecofennian gabbro intrusions coincide temporally with the global 2100–1800 Ma oro-gens related to the amalgamation of the Mesoproterozoic supercontinent Nuna. We provide a new reliable 1891–1875 Ma palaeomagnetic pole for Fennoscandia based on rock magnetic and palaeomagnetic studies on the Svecofennian intrusions in central Finland to ﬁll gaps in the Palaeoproterozoic palaeomagnetic record. By using the new pole together with other global high-quality data, we propose a new palaeogeographic reconstruction at 1885 Ma. This, to gether with pre vious data, supports a long-li ved relati vel y stable position of Fennoscandia at low to moderate latitudes at 1890–1790 Ma. Similar stable pole positions have also been obtained for Kalahari at 1880–1830 Ma, Siberia at 1880–1850 Ma, and possibly India at 1980– 1775 Ma. A new reconstruction at the beginning of this period indicates the convergence of several cratons at 1885 Ma in the initial stages of the amalgamation of the Nuna supercontinent at low to moderate latitudes. The close proximity of cratons at low to moderate latitudes is fur ther suppor ted by global and regional palaeoclimatic indicators. Stable position of several cratons could indicate a global period of minimal apparent drift at ca . 1880–1830 Ma. Before this period, the global palaeomagnetic record indicates large back-and-forth swings, most prominently seen in the high-resolution 2020–1870 Ma Coronation loops of the Slave craton. These large back-and-forth movements have been explained as resulting from an unstable geomagnetic ﬁeld or basin-or local-scale vertical-axis rotations. Ho wever , the most likely explanation is inertial interchange true polar wander (IITPW) events, which is in line with the suggestion of large amplitude true polar w ander e vents during the formation of the supercontinent


I N T RO D U C T I O N
High-quality palaeomagnetic poles (Buchan et al. 1994(Buchan et al. , 2000 ) ) are an essential tool used to reconstruct the palaeo geo g raphic histor y and relative lateral movement of different continental blocks (i.e.plate tectonics) in geological deep time by establishing the ancient latitudes and azimuthal orientations recorded in rock formations (see re vie w in Pesonen et al. 2021 ).Most of the global high-quality poles are derived from mafic intrusions, which can be precisely dated and are reliable magnetic recorders.Ho wever , due to the inherently episodic nature of mafic intrusions, there are large gaps in the record (Evans et al. 2021 ).The majority of the Palaeoproterozoic palaeomagnetic data from Fennoscandia are derived from gabbro intrusions located in Finland and Sweden.Several of these gabbro intrusions have been radiometrically dated and they form age clusters at ca .1885 Ma, 1860 Ma, 1800 Ma and 1780 Ma (poles re vie wed in Salminen et al. 2021a ), but of these, only the 1786 Ma Hoting gabbro provides a key pole (Buchan et al. 2000 ) with a proven primary magnetization (Elming et al. 2009 ).The other Palaeoproterozoic key pole for Fennoscandia was obtained from the 1976 Ma Onega dykes and sills (Lubnina et al. 2016 ).Our aim in this work is to provide a new reliable pole to fill gap at ca .1930-1870 Ma in the palaeomagnetic record of Fennoscandia, targeting ca .1885 Ma gabbros located in central Finland and ca .1866 Ma gabbros located in nor ther n Finland.Some of the ca .1885 Ma gabbros have pre viousl y been studied b y Pesonen & Stigzelius ( 1972 ), but their sampling scheme at the time produced only one sample from each gabbro intrusion, which does not fulfil modern statistical requirements and moreover may have not averaged out the palaeosecular variation (Meert et al. 2020 ).
The Svecofennian gabbros temporally coincide with the global 2100-1800 Ma orogens related to the formation of the Mesoproterozoic supercontinent Nuna (e.g.Hoffman 1988Hoffman , 1989 ; ;Zhao et al. 2002 ;Meert & Santosh 2022 ), also known as Columbia (see Wang et al. 2021 for a discussion of supercontinents, megacontinents and supercratons; andMeert 2012 andEvans 2013 for the nomenclature of Nuna versus Columbia).During the transition from Palaeoproterozoic supercratons (Salminen et al. 2021b ) to the Nuna supercycle (Elming et al. 2021 ), the palaeomagnetic poles of several cratons show large back-and-forth oscillating swings (Mitchell et al. 2010 ;Antonio et al. 2017 ;Gong & Evans 2022 ).This is most prominently seen in the 2020-1870 Ma Coronation loops obtained from the palaeomagnetic poles of the Slave craton (McGlynn & Irving 1978 ;Mitchell et al. 2010 ;Gong & Evans 2022 ).The Coronation loops have been interpreted as products of basin-or local-scale vertical axis rotations (Bingham & Evans 1976 ;Irving et al. 2004 ;Gong et al. 2018 ), as the result of switching of the geomagnetic field between a dominant axial dipole field and an equatorial dipole field (i.e. as Ediacaran; Abrajevitch & Van der Voo 2010 ), or by inertial interchange true polar wander (IITPW) events (Mitchell et al. 2010 ;Gong & Evans 2022 ).IITPW can be considered as a long-term process by which the solid body of the planet shifts relative to its spin axis (Gold 1955 ).IITPW events are expected in any planet with active geodynamics (Perron et al. 2007 ;Mitchell 2014 ) and they should have a global ef fect (Ev ans 1998(Ev ans , 2003 ) ). Mitchell et al. ( 2010 ) and Antonio et al. ( 2017 ) also proposed temporally overlapping large swings of poles for Fennoscandia, Amazonia, Superior, Kalahari and India at 1880-1840 Ma, thus supporting the global nature of IITPW.Ho wever , the 1880-1840 Ma swings of poles for Fennoscandia proposed by Antonio et al. ( 2017 ) were later contradicted by Elming et al. ( 2021 ).In this work, we explore the global ca .2100-1790 Ma high-quality data covering both the proposed IITPW event and its aftermath to examine whether the period of minimal apparent drift, following the period of large back-and-forth movements, was a global phenomenon.
A total of 38 samples were taken from five outcrops of the Kiuruvesi gabbrodiorite (Fig. 2 ) using a portable gasoline field drill.The samples were oriented using both sun and magnetic compasses.

Ylivieska (1883 Ma) and Mertuanoja gabbros
The Ylivieska area for ms par t of the accretionary arc complex of central and western Finland (Korsman et al. 1997 ).Grano-and quartzdiorite massifs cover hundreds of square kilometres in the area, but various schists and gneisses are also common (Salli 1961 ).Some more mafic plutonic rocks with gabbrodioritic or peridotitic composition are also present (Fig. 2 ; Salli 1961 ).According to Salli ( 1961 ), gabbrodioritic rocks follow the schists conformably, whereas granitoids cut them, indicating that the gabbros are younger than the schists but older than the granitoids (Salli 1961 ;Pesonen & Stigzelius 1972 ).
Two separate gabbro units were sampled in the Ylivieska area (Fig. 2 ).A total of 33 samples were taken with a portable gasoline field drill from six outcrops of the well-dated (U-Pb zircon age 1883 ± 8 Ma; Patchett & Kouvo 1986 ) Ylivieska gabbro, which covers an area of 45 km 2 (Salli 1961 ;Pesonen & Stigzelius 1972 ).A total of 37 samples were taken from eight outcrops of a smaller gabbro unit located near a narrow stream Mer tuanoja, nor theast of the Ylivieska gabbro (Fig. 2 ).The samples were oriented using both sun and magnetic compasses.
The Jalokoski gabbro was sampled at five outcrops by Mertanen & Pesonen ( 1992 ), with four of the outcrops located in Finland and one in Sweden.A total of 23 samples were collected.Because of the strong magnetic anomaly in the area, only sun-compass directions were used for orientation.

Nilsi ä lampr oph yre (1790 Ma) and micr otonalite d yk es
The Nilsi ä study area is located in the nor ther n Savo region in central-eastern Finland.Several microtonalite and lamprophyre dykes intrude the late Archaean (3100-2700 Ma; Sorjonen-Ward & Luukkonen 2005 ) basement gneiss of the Karelian Province (Fig. 2 ).These dykes intrude into the large-scale normal fault structures crosscutting early Svecofennian (1890-1875 Ma) regional deformation structures, more or less parallel to the NW-SE-trending palaeosuture zone between the Karelian Province and the Proterozoic Svecofennian domain (Woodard et al. 2014 ).Well-dated lamprophyre dykes (1790 ± 3 Ma; Woodard et al. 2014 ) in the area show no evidence of deformation or metamorphism (Woodard et al. 2014 ).Based on crosscutting relationships, microtonalite dykes predate the lamprophyre dykes in the area, but are younger than the 1860 Ma granite intrusion (Huhma 1981 ).
Sampling in Nilsi ä area (Fig. 2 ) included a total of 13 samples taken from a lamprophyre dyke on the island of Pieni-Vinkki in  lake Vuotj ärvi and a total of 74 samples from 13 microtonalite dykes in the surrounding area.Sampling was conducted with a portable gasoline field drill.Samples were oriented using both sun and magnetic compasses.

M E T H O D S
The rock magnetic and palaeomagnetic measurements for this study were conducted in the Geophysical Laboratory of the Geological Surv e y of Finland (GTK) for samples taken from Jalokoski gabbro and Nilsi ä, and in the Solid Earth Geophysics Laboratory of the University of Helsinki (UH) for samples taken from the Ylivieska and Kiuruvesi gabbros.
Stepwise alternating field (AF) demagnetizations of the natural remanent magnetization (NRM) of the samples were performed up to a maximum field of 160 mT using a three-axis demagnetizer coupled with a cryogenic 2G DC SQUID magnetometer (UH, GTK), or up to a maximum field of 100 mT using a Schonstedt AF demagnetizer (GTK).Selected sister specimens were thermally demagnetized up to 680 • C using an ASC Scientific model TD-48SC furnace (UH) or a custom-designed furnace (GTK).After each demagnetization step, a 2G DC SQUID magnetometer or an Agico JR4 spinner magnetometer (GTK) was used to measure remanent magnetization.Vector components were visually identified using stereographic and orthogonal projections and demagnetization decay cur ves (Zijder veld 1967 ).The mean directions were calculated using the least squares method (Kirschvink 1980 ). Mean remanent directions were determined for each component using Fisher statistics (Fisher 1953 ).
The magnetic mineralogy of the samples from the Ylivieska and Kiur uvesi intr usions was investigated using ther momagnetic analyses and hysteresis measurements of selected powdered whole-rock samples.An Agico KLY-3S-CS3 Kappabridge system was used to analyse ther momagnetic proper ties.During the procedure, the bulk susceptibilities were measured while heating the samples from room temperature to 700 • C and cooling them back to room temperature.Curie temperatures were determined using Cure v al 8.0.2 softw are ( http://www.agico.com).Hysteresis properties were measured with a Princeton Measurement Corporation Micromag TM 3900 model vibrating sample magnetometer (VSM) to identify the domain states of the magnetic carriers.
The Lowrie-Fuller test (Lowrie & Fuller 1971 ) was conducted for selected samples to investigate the domain structure of magnetic minerals.First, the NRM was AF demagnetized in steps up to 160 mT with a cryogenic 2G DC SQUID magnetometer (UH).Next, anhysteretic remanent magnetization (ARM) was imparted with an A GICO LD A3 alternating field demagnetizer, with a peak field of 100 mT and a biasing field of 50 μT in the z -direction.ARM w as pro gressi vel y demagnetized using the same sequence as was used for NRM demagnetization.Finally, an MMPM10 pulse magnetizer was used to induce isothermal remanent magnetization (IRM) in a 1.5 T field along the z -direction.IRM was progressi vel y demagnetized using the same sequence as in the previous steps.The coercivity spectra of NRM, ARM and IRM were then compared.The Lowrie test (Lowrie 1990 ) was used to examine the different coercivity fractions of IRM of the selected samples from Ylivieska and Kiuruvesi gabbros.IRM was imparted in samples with an MMPM10 pulse magnetizer in three axes: 1.2 T in the z -direction, 0.4 T in the y -direction and 0.12 T in the x -direction.IRM in these three axes is diagnostic for hard, intermediate and soft magnetic carriers, respecti vel y.After inducing IRM, samples were thermally demagnetized in steps up to 680 • C and magnetization along each orthogonal axis was then plotted against temperature.

Rock magnetic and palaeomagnetic results for the 1886 Ma Kiuruvesi gabbrodiorite
Ther momagnetic cur ves for selected samples from the Kiur uvesi gabbrodiorite display a significant decrease in magnetic susceptibility at 580 • C, indicating the presence of almost pure magnetite (Fig. 3 ).We interpret the minor drop in magnetic susceptibility at ca .380 • C to be related to the presence of maghemite, which transforms to magnetite during the heating.Other options could be titanomagnetite, impurities or changes in the magnetic grain size.Heating and cooling curves are irreversible, with cooling curves lacking the lower-temperature phase.The Lowrie-Fuller test indicated that single domain (SD) magnetite is a remanence-carrying mineral (Fig. 4 ).Ho wever , the Lo wrie test sho wed that magnetite is present in soft, intermediate and hard components for the Kiuruvesi gabbrodiorite samples (Fig. 4 ).In addition, a very narrow hysteresis loop indicates the presence of multidomain (MD) magnetite (Fig. 4 ).
Palaeomagnetic data from the 1886 ± 5 Ma (Kouvo 1978 6 , Table 1 ).Thermal demagnetization of the specimens produced square-shouldered decay curves with unblocking temperatures in a narrow range below 580 • C, indicating SD magnetite (Fig. 5 ).AF demagnetization showed typical coercivities of magnetite, with a median destructive field (MDF), which is the demagnetizing field required to halve the magnetization (Dunlop 1983 ), below 50 mT and no sign of a higher coercivity component with different magnetization directions (Fig. 5 ).Most of the specimens revealed a viscous component below 10 mT AF or 300 • C, but the direction of this viscous component was highly scattered and it is likely that the viscous component is carried by MD magnetite grains (Fig. 5 ).

Rock magnetic and palaeomagnetic results for ca . 1885 Ma Ylivieska and Mertuanoja gabbros
Thermomagnetic analyses of the selected samples from the Ylivieska and Mertuanoja gabbros yielded curves similar to the Kiuruvesi gabbrodiorite, although the susceptibility decrease at ca .380 • C w as significantl y steeper in some of the Yli vieska gabbro samples (Fig. 3 ).The Lowrie-Fuller test demonstrated significantly harder ARM compared to IRM, indicating that magnetization is carried by SD magnetite (Fig. 4 ).The Pesonen & Stigzelius ( 1972 ), whose sampling scheme only included one sample from each outcrop (Figs 5 and 6 , Table 1 ).The palaeomagnetic directions from Pesonen & Stigzelius ( 1972 ) form two groups, whereas the new palaeomagnetic directions obtained in this study form a single group displaying an elongation in the N-S direction and overlapping these two groups (Fig. 6 ).Thermally demagnetized specimens showed a square-shouldered shape with unblocking temperatures slightly below 580 • C, indicating SD magnetite (Fig. 5 ).AF demagnetized specimens displayed typical coercivities of SD magnetite, with an MDF below 50 mT, and no sign of a higher coercivity component with a different magnetization direction.Most of the samples from the Ylivieska gabbro displayed remanence with a viscous component, probably carried by MD magnetite.This component was isolated below 10 mT or 300 • C, with highly scattered directions.
Palaeomagnetic data from the undated Mertuanoja gabbro display a ChRM direction with a NW declination and intermediate downward inclination ( D = 339 • , I = 44 • with α95 = 5 • and k = 106), similar to Ylivieska and Kiuruvesi gabbros (Figs 5  and 6 , Table 1 ).The remanent magnetization of the samples from the Mertuanoja gabbro behaved similarly to the samples from the Ylivieska gabbro during thermal and alternating field demagnetization.

P alaeomagnetic r esults f or the 1866 Ma Jalokoski gabbro
Palaeomagnetic data from the 1866 ± 6 Ma (V ä än änen & Lehtonen 2001 ) Jalokoski gabbro display a ChRM direction with a NW declination and intermediate downward inclination ( D = 336 • , I = 41 • with α95 = 8 • and k = 90) (Figs 5 and 6 , Table 1 ).In thermally demagnetized specimens, the decay curves were squareshouldered, with unblocking temperatures in a narrow range below 580 • C, which indicates a low-Ti SD magnetite as the carrier of magnetization (Fig. 5 ).AF demagnetized specimens displayed typical coercivities of magnetite with an MDF below 50 mT, and no sign of a higher coercivity component with different magnetization directions (Fig. 5 ).Most specimens also yielded a viscous component revealed after demagnetization below 10 mT AF or 300 • C.This viscous component displayed a direction close to that of the Earth's present field.

P alaeomagnetic r esults f or the Nilsi ä 1860-1790 Ma microtonalite and 1790 Ma lamprophyre dykes
Palaeomagnetic data from 13 microtonalite dykes in the Nilsi ä study area display a ChRM direction with a NNW declination and intermediate downward inclination ( D = 345 • , I = 35 • with α95 = 2 • and k = 320; Figs 5 and 6 , Table 1 ).Thermally demagnetized microtonalite specimens yielded decay curves with a square-shouldered shape and unblocking temperatures close to 400 • C (Fig. 5 ).In addition, their behaviour during AF demagnetization was typical for (titano)magnetite (Fig. 5 ).These curves indicate that the magnetic carrier is probably titanomagnetite with a low to medium Ti content.The dated 1790 ± 3 Ma (Woodard et al. 2014 ) lamprophyre dyke in Nilsi ä displayed a ChRM direction with a NNW declination and intermediate downward inclination

Quality of the data
The obtained magnetization directions of the studied Svecofennian intrusions together with previous data from the same intrusions are presented in Fig. 6 .These data, together with site-mean poles and their R-scores (Meert et al. 2020 ), are compiled in Table 1 .
The pole calculated for the Kiuruvesi gabbrodiorite fulfils four of the seven revised Van der Voo ( 1990 ) criteria (Meert et al. 2020 ).It has a well-determined radiometric age of 1886 ± 5 Ma (Kouvo 1978, as cited in Marttila 1981 ; R1: 1).The Kiuruvesi mean pole does not statistically average out the PSV, as the observed A 95 = 2 • is significantly lower than the minimum value A 95 min = 5 • of the reliability envelope (Deenen et al. 2011(Deenen et al. , 2014 ) (R2: 0).This is also indicated by a precision parameter for pole K = 607 (Fisher 1953 ).Rock magnetic analysis were used to identify magnetic carriers (R3: 1).Neuvonen et al. ( 1981 ) reported diabase dykes crosscutting the Kiuruvesi gabbrodiorite, but despite our sampling efforts, none of these intrusions were found in this study.Therefore, it was not possible to perform baked contact tests (Everitt & Clegg 1962 ;R4: 0).According to Neuvonen et al. ( 1981 ) and Marttila ( 1987 ), the Kiuruvesi gabbrodiorite is crosscut by several subvertical dykes, indicating that no tilting has occurred after intrusion (R5: 1).Finally, only normal polarity was observed (R6: 0), and the pole calculated from the Kiuruvesi gabbrodiorite data does not resemble younger poles (R7: 1).
The pole calculated for the Ylivieska gabbro fulfils three of the seven criteria, most importantly having a well-determined radiometric age of 1883 ± 8 Ma (Patchett & Kouvo 1986 ) (R1: 1).The Ylivieska mean pole averages out the PSV, as the observed A 95 = 6 • is within the reliability envelope of Deenen et al. ( 2011Deenen et al. ( , 2014 ; A 95 min = 5 • , A 95 max = 18 • ).Ho wever , only six of the eleven studied outcrops have more than three samples, so the Ylivieska data thus do not meet the statistical requirements (R2: 0).Rock magnetic methods were used to identify magnetic carriers (R3: 1), but no palaeomagnetic field tests were performed (R4: 0).Crosscutting dykes were not observed, so possible tilting cannot be ruled out (R5: 0).Only normal polarity was observed (R6: 0).The pole calculated from Ylivieska gabbro data does not resemb le y ounger poles (R7: 1).
The pole calculated from the Mertuanoja gabbro fulfils only two of the seven R-criteria.It lacks radiometric dating (R1: 0) but is assumed to be related to the intrusion of the Ylivieska gabbro due to its geographical proximity.Eight outcrops were sampled for the Mertuanoja gabbro, with three or more samples from each.In addition, based on the reliability envelope of Deenen et al. ( 2011Deenen et al. ( , 2014) ) , data from the Mertuanoja gabbro average out the PSV ( A 95 = 6 • , A 95 min = 5 • , A 95 max = 22 • ), but the precision parameter for the pole (Fisher 1953 ) is high ( K = 97 • ), and the Mertuanoja gabbro data therefore fail criterion R2 (R2: 0).Rock magnetic methods were used to identify magnetic carriers (R3: 1).Due to the lack of crosscutting dykes and contacts to host rock, no palaeomagnetic field tests were performed (R4: 0) and possible tilting could not be ruled out (R5: 0).Re- versed polarities were not observed (R6: 0).The pole calculated from the Mertuanoja gabbro data does not resemble younger poles (R7: 1).The Jalokoski gabbro fulfils three of the seven R-criteria, with a well-determined radiometric age of 1866 ± 6 Ma (V ä än änen & Lehtonen 2001 ; R1: 1).Because the data only originate from five outcrops and the precision parameter for the pole (Fisher 1953 ) is high (K = 139), the pole from the Jalokoski gabbro fails the second R criterion (R2: 0).The AF and thermal demagnetization curves have been used in the identification of magnetic carriers (R3: 1).There were no observations of cross-cutting dykes or contacts to host rock, and therefore no palaeomagnetic field tests were performed (R4: 0) and possible tilting could not be ruled out (R5: 0).Similar to most Svecofennian intrusions, only normal polarity was observed (R6: 0).The pole calculated from the Jalokoski gabbro data does not resemble younger poles (R7: 1).
The pole calculated from the Nilsi ä microtonalite dykes fulfils three of the seven R-criteria.Due to the inherited zircon, there are no reliable radiometric ages for these dykes (R1: 0).According to Huhma ( 1981 ), the dykes predate the well-dated lamprophyre dykes in the area (1790 ± 3 Ma; Woodard et al. 2014 ), but are younger than the ca .1860 Ma granite intrusion, providing a poorly constrained age estimate of 1860-1790 Ma.The data from the Nilsi ä microtonalite dykes do not statistically average out the PSV, as the observed A 95 = 2 • is significantly lower than the minimum value A 95 min = 4 • of the reliability envelope (Deenen et al. 2011(Deenen et al. , 2014 ; ;R2: 0).This is also indicated by an exceptionally high precision parameter for the pole ( K = 328; Fisher 1953 ).The shapes of the AF and thermal demagnetization curv es hav e been used in the identification of magnetic carriers (R3: 1).No palaeomagnetic field tests were performed (R4: 0).The 1860-1790 Ma Nilsi ä microtonalite dykes are subvertical (Neuvonen et al. 1981 ), as are also the 1790 Ma lamprophyre dykes in the area (Paavola 1984 ).Moreover, the lamprophyre dykes display no evidence of deformation or metamorphism (Woodard et al. 2014 ).These observations indicate that no tilting has occurred since the dykes intruded (R5: 1).Only normal polarity was observed (R6: 0).The pole calculated from Nilsi ä microtonalite data does not resemb le y ounger poles (R7: 1).

Site
Site name  Lat, Lon: site latitude and longitude; B: number of analyzed sites; N: number of analyzed samples; n: number of analyzed specimen; * marks the level used in mean calculations; D: declination; I: inclination; α95: the radius of the 95% confidence cone in Fisher ( 1953 ) statistics; k: Fisher ( 1953 ) precision parameter; Plat, Plon: latitude and longitude of the virtual geomagnetic pole; A95: radius of the 95% confidence cone of the pole; K: Fisher ( 1953 ) precision parameter of pole; R: revised Van der Voo ( 1990 ) reliability criteria proposed by Meert et al. ( 2020 ) with a R-score (number of criteria fulfilled by the pole) A95 min , A95 max : minimum and maximum values of the reliability envelope (Deenen et al. 2011(Deenen et al. , 2014 ) ); A : data obtained from Neuvonen et al. ( 1981 ).B : data obtained from Pesonen & Stigzelius ( 1972 ).

Origin of the magnetization
The poles from this work, calculated from the Kiuruvesi gabbrodiorite, the Ylivieska gabbro, the Mertuanoja gabbro, the Jalokoski gabbro and the Nilsi ä microtonalites, form a spread cluster located at ca .230-250 • E and 40-50 • N (Fig. 7 ).The A95 confidence cones of these poles overlap one another with the exception of the Ylivieska pole.Ho wever , the Kiuruvesi and Nilsi ä poles do not statistically average out the PSV, and therefore may be considered as virtual geomagnetic poles (VGP).According to McElhinny & McFadden ( 2000 ), an individual VGP may differ by even 20 • from the rotation axis of the palaeomagnetic reference frame, therefore indicating the Kiuruvesi and Nilsi ä poles might not be distinguishable from the Ylivieska pole.It is possible that the tight clustering of the studied ca .1885-1790 Ma poles could arise from slow cooling of the large gabbro intrusions combined with minimal drift, as magnetization ages ca .10-20 Myr (Buchan et al. 2000 ;Pesonen et al. 2003 ) and even 100 Myr (Elming et al. 2009 ) younger than the crystallization ages of Fennoscandian gabbros have been proposed.Ho wever , we consider slo w cooling unlikely, as a more recent study of the ca .20-km-long and 4-km-wide Turinge gabbrodiabase showed that it intruded at great depth, indicating relati vel y rapid cooling with a magnetization age close to the baddeleyite U-Pb age of the rock (Elming et al. 2019 ).In addition, poles calculated in this study overlap with a high-quality palaeomagnetic pole derived from the rapidly cooled ca .1870 Ma Keuruu diabase dykes (Klein et al. 2016 ; Fig. 7 ), indicating magnetization ages close to this age.
Palaeomagnetic data from Ylivieska gabbro, show a N-S elongation (Fig. 6 ).According to Beck Jr ( 1999 ) elongation could be related to the slow cooling combined with rapid continental drift.Ho wever , in the case of the Ylivieska gabbro, this is an unlikely explanation, as tight clustering of 1885-1790 Ma Fennoscandian poles prove against relati vel y rapid polar wander required for the development of the elongation in response to slow cooling of gabbro.Fur ther more, the Lowrie test demonstrates that magnetite is present in three coercivity fractions (Fig. 4 ).It is probable that the intermediate coercivity fraction has overlapping coercivity spectra with hard coercivity fraction, and therefore secondary component might be incompletely separated, which could explain the observed elongation.Ho wever , this direction is not close to commonly obser ved secondar y components (e.g.component B in Mer tanen et al. 2006b ), and therefore this is not a likely explanation, and the cause of elongation remains unsolv ed.Nev ertheless, the pole calculated for the Ylivieska gabbro is within the tight cluster of 1870-1790 Ma Fennoscandian poles (Fig. 7 ), indicating that the direction of the suggested secondary component does not greatly differ from the direction of the primary magnetization.
We suggest that the magnetization ages are close to the crystallization ages, based on comparisons with the high-quality data and cooling rates of intrusions of approximately the same size.The cause of N-S elongation observed in the palaeomagnetic data from the Ylivieska gabbro remains unsolved, but the mean direction of Ylivieska gabbro seems to be similar to other sites, and therefore considered to present primary magnetization.

A new reliable 1891-1875 Ma palaeomagnetic pole
By combining the data from the similarly aged Kiuruvesi gabbrodiorite and the Ylivieska gabbro (Kouvo 1978, as cited in Marttila 1981 ;Patchett & Kouvo 1986 ) and the presumabl y coe v al Mertuanoja gabbro, we calculated a new reliable ∼1885 Ma mean pole at Plat: 47 • N and Plong: 238 • E (with B: 27, A95: 3 • , K: 66; Table 2 ).Considering that the studied rock units are relati vel y large gabbro intrusions, each outcrop has been thought to represent a separate cooling unit.
This new mean pole satisfies five of the seven R-criteria (Table 2 ).It has a well-determined radiometric age of 1891-1875 Ma (R1: 1).The palaeosecular variation is statistically averaged out, as the observed A 95 = 3.5 • is within the reliability envelope ( A 95 min = 3.2 • , A 95 max = 10.3 • ), and there are statistically sufficient number of sites and samples (R2: 1).Rock magnetic methods were used to identify the magnetic carriers (R3: 1).Subvertical dykes that are reported to crosscut the Kiuruvesi gabbrodiabase (Marttila 1987 ) indicate that no tilting has occurred after the intrusion (R5: 1).The pole does not resemb le y ounger poles (R7: 1).
The new mean pole fills the 60 Myr gap in the palaeomagnetic record of Fennoscandia at ca .1930-1870 Ma.It overlaps with ca .1870, 1860, 1800 and 1790 Ma high-quality poles and ca .1840 Ma and 1810 Ma poor-quality poles of Fennoscandia (Fig. 7 ), indicating a relati vel y stable position for Fennoscandia at ca .35-20 • N latitudes (Salminen et al. 2021a ) during the time period from 1885 to 1790 Ma.Coe v al occurrences of e v aporites, stromatolite-like structures and lateritic palaeosols in Finland and Sweden support these palaeolatitudes (see compilation of palaeoclimatic indicators for Baltica and its subcratons in Salminen et al. 2021a ).

Large global oscillations in palaeomagnetic data at 2060-1880 Ma
During the period of break-up of the proposed Palaeoproterozoic supercratons (Salminen et al. 2021b ) and the onset of the amalgamation of the Nuna supercontinent (Elming et al. 2021 ), the palaeomagnetic poles of several cratons display large back-andforth oscillating swings (Fig. 8 ).This is most prominently seen in the palaeomagnetic poles of the Slave craton of Laurentia at 2020-1870 Ma (McGlynn & Irving 1978 ;Mitchell et al. 2010 ;Gong & Evans 2022 ).These swings of up to ∼110 • in the apparent polar wander path (APWP) of the Slave craton, called the Coronation loops (McGlynn & Irving 1978 ), are found in the high-resolution palaeomagnetic record derived not only from mafic intrusions but also from sedimentary rocks from the Great Slave, Coronation, and Kilohigok Basins (Mitchell et al. 2010 ;Gong & Evans 2022 ).During the Coronation loops, the Slave craton rifted from the proposed Sclavia supercraton (Bleeker 2003 ;French & Heaman 2010 ) and amalgamated with other Laurentian Archaean cratons.Such palaeomagnetic data would indicate rapid latitudinal drift and rotation of the Slave craton with plate velocities higher than for moderntype plate tectonics (Zahirovic et al. 2015 ), which therefore argues South India 1766 ± 1, 1764.5 ± 0.9 Shankar et al. ( 2018 ) 43 .0320 .0 10 .31111101 6 A Meert et al. ( 2021 ) Code refers to codes used in Figs 7, 8 and 9; Plat, Plon: pole latitude and longitude; A95: 95% confidence circle of the pole; R 1234567 : revised Van der Voo ( 1990 ) reliability criteria proposed by Meert et al. ( 2020 ) with a R-score (number of criteria fulfilled by the pole); Grade: grade of the pole.A, B: high-quality poles, * : not graded yet, -lower-quality pole.
against apparent polar wander alone being able to explain the observed swings.
Competing models have been proposed to explain the Coronation loops: local vertical axis rotation of small blocks of the Slave craton (Irving et al. 2004 ;Gong et al. 2018 ), an unstable magnetic field involving a considerable horizontal dipole (i.e. as Ediacaran; Abrajevitch & Van der Voo 2010 ) and inertial interchange true polar wander (IITPW; Mitchell et al. 2010 ;Gong & Evans 2022 ).Ho wever , using high-resolution sedimentary data from the Goulbur n superg roup of the Slave craton, Gong & Evans ( 2022 ) noted that although basin-scale rotations were able to bring some correlative poles closer, some other poles were simultaneously displaced even further apart.Therefore, although basin-scale rotations may affect some poles, they alone cannot explain the observed backand-forth mo vement.Moreo ver, antipodal directions present in the high-resolution Slave craton data indicate that the geomagnetic field should not have been significantl y dif ferent from a geocentric axial dipole (GAD) field during the Coronation loops (Gong & Evans 2022 ).Gong & Evans ( 2022 ) summarized these results by suggesting that the IITPW probably controls the overall pattern of the Coronation loops.
Large swings in high-quality Palaeoproterozoic palaeomagnetic data have been obtained globally, including for the Amazonia, Kalahari, India and Superior cratons (Mitchell et al. 2010 ;Antonio et al. 2017 ).Antonio et al. ( 2017 ) also proposed a large 1880-1850 Ma swing for Fennoscandia, but this was later contradicted by Elming et al. ( 2021 ), who only used poles with proven primary magnetization for Fennoscandia in their analyses, unlike Antonio et al. ( 2017 ).Here, we extend the time period with reliable poles to cover the duration of the Coronation loops and explore the global high quality 2060-1870 Ma palaeomagnetic data.
pole from the Mashonaland dolerites (Bates & Jones 1996 ) and the 1875 Ma pole from the post-Waterberg older sills (Hanson et al. 2004 ).Palaeomagnetic data from the Superior craton indicate a ca .40 • distance between the ca .2000 Ma Minto pole (Buchan et al. 1998 ;Evans & Halls 2010 ) and the 1880 Ma Molson pole (Halls & Heaman 2000 ;Evans & Halls 2010 ).Palaeomagnetic data from the Amazonian craton are scarce, but show relati vel y large distances between poles at 2080-1880 Ma: ca .60 • at 2080-2065 Ma, and ca .120 • at 2065-1880 Ma (Fig. 8 ).Similar to the Amazonian craton, data for the Indian craton are scarce, but large distances between the poles are obtained at 2220-1980 Ma, indicating an apparent drift of ca .180 • (Fig. 8 ).In summary, the back-and-forth swings in the 2060-1880 Ma AP-WPs of the Fennoscandia, Slave, Amazonia and Kalahari cratons form hairpin shapes with different lengths and ages of their apexes.In addition, high-quality poles from India indicate significant drifting without back-and-forth movement.We believe that this may be due to a lack of coe v al and high-resolution data and propose that the oscillation may indicate IITPW events (Evans 2003 ;Gong & Evans 2022 ).This is in line with the suggestion of large-amplitude TPW events during the formation of a supercontinent (Evans 1998(Evans , 2003 ; ;Mitchell et al. 2012Mitchell et al. , 2021 ; ;Mitchell 2014 ).

A period of stable pole positions in the aftermath of a proposed IITPW event
High-quality 1890-1790 Ma palaeomagnetic data for Fennoscandia show the pre viousl y mentioned tight cluster of poles between longitudes of 210-240 • E, indicating a period of stable pole positions after the significant drifting at 2060-1930 Ma (Fig. 8 ).Djeutchou et al. ( 2021 ) compiled high-resolution 1890-1830 Ma palaeomagnetic data for the Kalahari craton from the Black Hills dyke swarm and coe v al regional magmatic provinces (F ig. 8 , Tab le 2 ), indicating minimal apparent drift for the Kalahari craton at 1880-1830 Ma, similarly to Fennoscandia (Figs 7 and 8 ).In addition, Djeutchou et al. ( 2021 ) provided 1890-1830 Ma magnetostratigraphy, displaying periods of normal, reversed and mixed polarity.One polarity reversal is shown at ca .1870-1860 Ma, correlating well with the only known reversed-polarity data from Fennoscandia between 1890 and 1830 Ma, recording dual polarity in the ca .1870 Ma Keuruu dykes (Klein et al. 2016 ).In addition, dual polarities are observed in the Superior craton from the ca .1880 Ma pole obtained from the Molson dykes (Halls & Heaman 2000 ;Evans & Halls 2010 ) and the ca .1870 Ma Haig-Flaherty-Sutton mean pole (Schmidt 1980 ;Schwarz et al. 1982 ).The Kalahari data also indicate reversed polarity at ca .1830 and 1850 Ma (Djeutchou et al. 2021 ), but no high-quality palaeomagnetic data are available for Fennoscandia during those times.Correspondingly, the ca .1840 Ma Haukivesi intrusions indicate a normal polarity period for Fennoscandia at that time, but no 1840 Ma data exist for the Kalahari craton.
Siberian data are scarce, but three ca .1880-1850 Ma palaeomagnetic poles are within 10 • of each other , and sho w reversed or dual polarities, therefore supporting stable palaeopole locations (Fig. 8 ).The palaeomagnetic data from the Indian craton are too scarce to draw conclusions, but three high-quality 1980-1770 Ma poles may indicate a relati vel y stable position for India during that period (Fig. 8 ).In contrast, two high-quality palaeomagnetic poles from Amazonia show a ca .100 • distance between 1880 and 1850 Ma (Antonio et al. 2017 ).

A palaeogeographic reconstruction at 1885 Ma
We provide a new palaeogeog raphical reconstr uction at 1885 Ma, at the beginning of the suggested period of minimal apparent drift following the proposed IITPW event (Fig. 9 ).The reconstruction largely follows the 1860 Ma reconstruction provided by Elming et al. ( 2021 ), with equatorial positions for India, Siberia, Sarmatia and Volgo-Uralia, and Fennoscandia and the Superior craton at moderate latitudes.At that time, the Slave craton had already amalgamated with Rae and Hearne cratons (Hoffman 2014 ), and is reconstructed to low to moderate latitudes, further east of Superior, with the Manike w an Ocean separating these cratons (Stauffer 1984 ).Contradicting the reconstruction by Elming et al. ( 2021 ), we prefer a position of Kalahari in the nor ther n hemisphere following the 1880 Ma reconstruction by Djeutchou et al. ( 2021 ).Djeutchou et al. ( 2021 ),, based this position for Kalahari on their magnetostratigraphic correlations with coe v al re versed polarity data from Fennoscandia and Superior, and the alignment of the Black Hills and Mazowe dyke swarms of the Kalahari craton with the ca .1880 Ma Circum-Superior large igneous province magmatic centre of the superior craton (Fig. 9 ), originally proposed by Minifie et al. ( 2013 ).Global low to moderate latitudes are supported by various types of palaeoclimatic evidence at the global and regional scale.The abundance of global black shale deposits and banded iron formations at ca .2000-1700 Ma indicates the presence of huge landmasses at low to moderate latitudes (Condie et al. 2001 ).This is further supported by the increase in stromatolite deposits observed in the marine carbonates of North America (Peters et al. 2017 ), the extensive ca .2000-1800 Ma red bed successions in the Kalahari region (Dorland et al. 2006 ;Moen 2006 ;Geng et al. 2014 ; van Niekerk & Beukes 2019 ;de Kock et al. 2021 ) and the ca .1890-1840 Ma e v aporites, stromatolite-like structures and lateritic palaeosols observed in Fennoscandia (Lager 2001 ;Lahtinen & Nironen 2010 ;Dahlin et al. 2012 ).
Together with the previous 1870-1790 Ma palaeomagnetic data (e.g.Elming et al. 2021 ;Salminen et al. 2021a ), this new reconstr uction suppor ts a long period of minimal apparent drift for Fennoscandia at low to moderate latitudes.Several cratons had already assembled together at this time, possibly explaining the period of stable positions.
for Fennoscandia re vie wed in Salminen et al. ( 2021a ) and verify that Fennoscandia was located at low to moderate latitudes at 1885-1790 Ma, further supported by the occurrences of evaporites, stromatolite-like structures and lateritic palaeosols in Finland and Sweden at that time.
(ii) Unlike the previous studies on Kiuruvesi (Neuvonen et al. 1981 ) and Ylivieska gabbros (Pesonen & Stigzelius 1972 ), the new palaeomagnetic data meets the present-day statistical requirements for high-quality data.New data are also statistically shown to average out the palaeosecular variation.
(iii) The palaeomagnetic data from cratons in Fennoscandia, the Kalahari, Siberia and possibly India indicate a global period of minimal drifting at ca .1880-1830 Ma, following the period of large craton movements at ca .2060-1880 Ma.We agree on the previous interpretations of these movements being related to the IITPW events (Mitchell et al. 2010 ;Antonio et al. 2017 ;Gong & Evans 2022 ), thought to be related to the amalgamation of supercontinent (Evans 1998(Evans , 2003 ; ;Mitchell et al. 2012, 2021Mitchell 2014 ).
(iv) The new 1885 Ma palaeo geo g raphic reconstr uction at the beginning of this period indicates that most of these cratons were located at low to moderate latitudes at this time, further supported by global and regional palaeoclimatic indicators.Reconstruction shows that Kalahari, Fennoscandia, India, Siberia, Sarmatia, Superior and Slave cratons were at close proximity.Observed stable positions of Fennoscandia, Kalahari, Siberia and possibly India could therefore indicate that these cratons were drifting as a stable joint landmass, after the amalgamation indicated by the proposed Palaeoprotorozoic IITPW event.

A C K N O W L E D G M E N T S
The authors would like to thank the re vie wers, Michiel de Kock and Silvana Geuna, for their comments, which greatly improved this paper.We would also like to thank David Whipp from the University of Helsinki for his helpful comments for the article.This work is related to the PhD study of TL and was funded by the Doctoral Programme in Geosciences (GEODOC) of the University of Helsinki.The fieldwork was funded with Academy of Finland grant #288277 to JS and with a grant from the Petter and Margit F orsstr öm F oundation to TL.

AU T H O R C O N T R I B U T I O N
Research was conceptualized by Johanna Salminen and Toni Luoto.Sampling was mainly conducted by Toni Luoto, Johanna Salminen and Satu Mertanen.Satu Mertanen also provided results from earlier unpublished studies.Toni Luoto processed the data and analysed the results.The work was supervised by Johanna Salminen.Toni Luoto wrote the original paper.Johanna Salminen, Satu Mertanen, Sten-Åke Elming and Lauri Pesonen re vie wed and edited the paper.

C O N F L I C T O F I N T E R E S T
The authors declare that the y hav e no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.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.

D ATA AVA I L A B I L I T Y
Site level data are available in the paper.Sample/specimen level data available by request from corresponding author.

Figure 3 .
Figure3.High-temperature thermomagnetic curves for selected samples from Kiuruvesi gabbrodiorite and two gabbro units near Ylivieska.Heating (cooling) curves are drawn with a solid black (dashed grey) line.

Figure 5 .
Figure 5. Examples of demagnetization behaviour from alternating field (AF) and thermal demagnetizations (TH).Filled (open) symbols represent the horizontal (vertical) plane in orthogonal projections and downward (upward) directions in stereoplots.Arrows indicate the direction of the characteristic remanent magnetization (ChRM) component.Numbers in orthogonal projections indicate the applied AF field in mT or temperature in • C.
Lowrie test indicated that

Table 1 .
Palaeomagnetic results for intrusions of Svecofennian age.
and the 1990 Ma pole from the Waterberg unconformity-bounded sequence II (de Kock et al. 2006 ), and a ca .35 • distance between the 1880 Ma