Late Paleozoic–Early Mesozoic Granite Magmatism on the Arctic Margin of the Siberian Craton during the Kara-Siberia Oblique Collision and Plume Events

We present new structural, petrographic, geochemical and geochronological data for the late Paleozoic–early Mesozoic granites and associated igneous rocks of the Taimyr Peninsula. It is demonstrated that large volumes of granites were formed due to the oblique collision of the Kara microcontinent and the Siberian paleocontinent. Based on U-Th-Pb isotope data for zircons, we identify syncollisional (315–282 Ma) and postcollisional (264–248 Ma) varieties, which differ not only in age but also in petrochemical and geochemical features. It is also shown that as the postcollisional magmatism was coming to an end, Siberian plume magmatism manifested in the Kara orogen and was represented by basalts and dolerites of the trap formation (251–249 Ma), but also by differentiated and individual intrusions of monzonites, quartz monzonites and syenites (Early–Middle Triassic) with a mixed crustal-mantle source. We present a geodynamic model for the formation of the Kara orogen and discuss the relationship between collisional and trap magmatism.

Thus, two domains (Southern and Northern) represent the passive margins of two paleocontinents. The sediments of Southern Taimyr correspond to the passive margin of the Siberian continent, whereas the Northern domain represents allochthonous continental slope and rise deposits of the Kara microcontinent. The Central Taimyr accretionary belt lies between them.
The formation of the Kara orogen per se began in the Carboniferous, due to the collision between the Kara microcontinent and the Siberian paleocontinent. We can assess the time of its formation from manifestations of regional metamorphism and granitoid magmatism (315-260 Ma) [50][51][52]69], and from dislocations along thrusts and strike-slip faults of various scales [7,[12][13][14]20,35].

Geology of the Granites
Late Paleozoic granites and regionally metamorphosed rocks can be traced from the southwest to the northeast along the northern part of the Taimyr Peninsula for 700 km. Significant areas are occupied by syncollisional granites located in migmatization zones among rocks of the amphibolite  [80][81][82][83] showing the location patterns of faults and Permian postcollisional and Permian-Triassic Siberian LIP igneous rocks.
Thus, two domains (Southern and Northern) represent the passive margins of two paleocontinents. The sediments of Southern Taimyr correspond to the passive margin of the Siberian continent, whereas the Northern domain represents allochthonous continental slope and rise deposits of the Kara microcontinent. The Central Taimyr accretionary belt lies between them.
The formation of the Kara orogen per se began in the Carboniferous, due to the collision between the Kara microcontinent and the Siberian paleocontinent. We can assess the time of its formation from manifestations of regional metamorphism and granitoid magmatism (315-260 Ma) [50][51][52]69], and from dislocations along thrusts and strike-slip faults of various scales [7,[12][13][14]20,35].

Geology of the Granites
Late Paleozoic granites and regionally metamorphosed rocks can be traced from the southwest to the northeast along the northern part of the Taimyr Peninsula for 700 km. Significant areas are occupied by syncollisional granites located in migmatization zones among rocks of the amphibolite facies, especially in the central part of the Northern Domain ( Figure 3). They form intrusions of irregular shape and of various sizes-from lens-like bodies of tens of meters among migmatites to large (several hundred km 2 ) plutons, commonly elongated and concurrent with folds and strike-slip-thrusts (Figures 1 and 3). These bodies can correspond to the parautochthonous varieties described in [49], that is to say, granites that were moved into overlying rocks of epidote-amphibolite and greenschist metamorphism facies with the formation of contact hornfels rims ( Figure 3). There are no syncollisional granites to the south of the Main Taimyr and Diabase strike-slip-thrust sutures, which underlines their exclusive localization in the blocks of the orogen that were exhumed from deep levels to the surface. In contrast, postcollisional granite stocks cut through rocks of the Northern and Central Domain, including unmetamorphosed Paleozoic cover deposits in the Central Taimyr accretionary belt. These plutons have oval and rounded shapes, small sizes from several tens to several hundreds of square kilometers, and are localized mainly in the eastern part of the orogen (Figures 1 and 2).
The relationships between syn-and postcollisional granites can often be observed in the field. The type of contact between postcollisional pink biotite-bearing, medium-coarse-grained granite-porphyries and the coarse-grained, greenish-dark-grey, hornblende-biotite granodiorites was recorded by V.F. Proskurnin in the bed of Oleniy Creek close to Zeeberg Bay ( Figure 4). Minerals 2020, 10, 571 7 of 39 (a) (b) Figure 4. Field photographs of (a) horizontal weakly-wavy jointing of the postcollisional porphyraceous coarse-grained biotite granites (Oleniy Creek, Zeeberg Bay); (b) contact of pink biotitebearing medium-to coarse-grained granite-porphyry and greenish-dark-grey hornblende-biotite coarse-grained granodiorites, with apophyses of fine-grained granite-porphyry and aplites in the upper part of the photo accompanying potassium feldspar porphyroblasts in the host granodiorites.
In the region of the Tessema River mouth (western part of Cape Chelyuskin, Northern Domain), the syncollisional two-mica granites intruding metamorphosed flyschoid deposits are themselves cut by a stock of Permian postcollisional hornblende-biotite granite. Here, both granites are also cut by Early Triassic dolerite dikes [85].
Where the postcollisional plutons intrude weakly metamorphosed or unmetamorphosed rocks, these discordant bodies are surrounded by contact metamorphic rims (of the muscovite-hornfels to amphibole-hornfels facies, often with cordierite and andalusite) that are tens to hundreds of meters wide and even reaching several kilometers [11,13,73,86]. Xenoliths of hornfelsed terrigenous host rocks have been identified both in syncollisional and postcollisional granites and granodiorites ( Figure 5a). Locally, the amount of inclusions of hornfelsed host rocks can be high and they can reach sizes from 10-20 cm for xenoliths to 150-200 m for the blocks [85].
The contact metamorphic rims may have concordant granodiorite and granite apophyses ( Figure 5b). These granitoids are characterized by a widely occurring vein phase, represented by dikes of granite-porphyry and aplite, as well as pegmatite and quartz-feldspar veins. Field photographs of (a) horizontal weakly-wavy jointing of the postcollisional porphyraceous coarse-grained biotite granites (Oleniy Creek, Zeeberg Bay); (b) contact of pink biotite-bearing medium-to coarse-grained granite-porphyry and greenish-dark-grey hornblende-biotite coarse-grained granodiorites, with apophyses of fine-grained granite-porphyry and aplites in the upper part of the photo accompanying potassium feldspar porphyroblasts in the host granodiorites.
In the region of the Tessema River mouth (western part of Cape Chelyuskin, Northern Domain), the syncollisional two-mica granites intruding metamorphosed flyschoid deposits are themselves cut by a stock of Permian postcollisional hornblende-biotite granite. Here, both granites are also cut by Early Triassic dolerite dikes [85].
Where the postcollisional plutons intrude weakly metamorphosed or unmetamorphosed rocks, these discordant bodies are surrounded by contact metamorphic rims (of the muscovite-hornfels to amphibole-hornfels facies, often with cordierite and andalusite) that are tens to hundreds of meters wide and even reaching several kilometers [11,13,73,86]. Xenoliths of hornfelsed terrigenous host rocks have been identified both in syncollisional and postcollisional granites and granodiorites ( Figure 5a). Locally, the amount of inclusions of hornfelsed host rocks can be high and they can reach sizes from 10-20 cm for xenoliths to 150-200 m for the blocks [85].
The contact metamorphic rims may have concordant granodiorite and granite apophyses ( Figure 5b). These granitoids are characterized by a widely occurring vein phase, represented by dikes of granite-porphyry and aplite, as well as pegmatite and quartz-feldspar veins. Minerals 2020, 10 Within the Kara orogen, there are also widespread manifestations of traps of the Siberian plume (Figures 1 and 2). Along with basalts, tuffs, and dolerites sills, there are differentiated intrusions, for example the Dumtalei and Dikarabigai ultrabasic to felsic alkaline layered intrusive complexes with compositions ranging from gabbro-dolerites to monzonites and syenites, as well as individual plutons of syenite, monzonite and quartz syenite [34,42,43,45,46,64,87]. These magmatic bodies are most widespread in the Southern Domain, but they also occur in the Central Domain and even the Northern Domain, including the Severnaya Zemlya archipelago (Figures 1-3) [5,6,11,13], which is very important for the determination of the timeline of the main tectonic events. Moreover, these plutons and dikes intrude both Precambrian and Paleozoic formations, including the syncollisional and postcollisional granites. Early-Middle Triassic syenites, monzonites and quartz syenites form small individual intrusions, often cutting the trap basalts and syncollisional granite plutons ( Figure  3) [34]. They were also identified within and close to stocks of late Permian postcollisional granitoids, forming ring intrusions, in which the outer borders are represented by late Permian granodiorite or granite, and towards the centers of the intrusions the rocks are transitional to Middle-Late Triassic monzogabbro, monzodiorite, and monzonite to syenite, and quartz syenite ( Figure 6) [56]. Within the Kara orogen, there are also widespread manifestations of traps of the Siberian plume (Figures 1 and 2). Along with basalts, tuffs, and dolerites sills, there are differentiated intrusions, for example the Dumtalei and Dikarabigai ultrabasic to felsic alkaline layered intrusive complexes with compositions ranging from gabbro-dolerites to monzonites and syenites, as well as individual plutons of syenite, monzonite and quartz syenite [34,42,43,45,46,64,87]. These magmatic bodies are most widespread in the Southern Domain, but they also occur in the Central Domain and even the Northern Domain, including the Severnaya Zemlya archipelago (Figures 1-3) [5,6,11,13], which is very important for the determination of the timeline of the main tectonic events. Moreover, these plutons and dikes intrude both Precambrian and Paleozoic formations, including the syncollisional and postcollisional granites. Early-Middle Triassic syenites, monzonites and quartz syenites form small individual intrusions, often cutting the trap basalts and syncollisional granite plutons ( Figure 3) [34]. They were also identified within and close to stocks of late Permian postcollisional granitoids, forming ring intrusions, in which the outer borders are represented by late Permian granodiorite or granite, and towards the centers of the intrusions the rocks are transitional to Middle-Late Triassic monzogabbro, monzodiorite, and monzonite to syenite, and quartz syenite ( Figure 6) [56].

Materials-Sampling and Petrography
Samples for petrographic, geochemical and geochronological studies of the granitoids were taken from intrusions of different types in the north and northeast of Taimyr from the Minin Skerries to Cape Chelyuskin (Figure 1). The coordinates and georeferences for studied samples are listed in Appendix A. The syncollisional granitoids, which, after L.V. Makhlaev and N.I. Korobova [49], we assign to the autochthonous and parautochthonous types, form intrusions of irregular shape and of various sizes-from lens-like bodies of tens of meters among migmatites to large (several hundred km 2 ) plutons, commonly elongated and concurrent with folds and strike-slip-thrusts (Figures 1 and

Materials-Sampling and Petrography
Samples for petrographic, geochemical and geochronological studies of the granitoids were taken from intrusions of different types in the north and northeast of Taimyr from the Minin Skerries to Cape Chelyuskin (Figure 1). The coordinates and georeferences for studied samples are listed in Appendix A. The syncollisional granitoids, which, after L.V. Makhlaev and N.I. Korobova [49], we assign to the autochthonous and parautochthonous types, form intrusions of irregular shape and of various sizes-from lens-like bodies of tens of meters among migmatites to large (several hundred km 2 ) plutons, commonly elongated and concurrent with folds and strike-slip-thrusts (Figures 1  and 3). The autochthonous granites are clearly developed in migmatization zones among rocks of the amphibolite facies-migmatized garnet-two-mica sillimanite, garnet-biotite and garnet-hornblende gneisses, where the granites, migmatites and gneisses display gradual transitions. Among these granitoids, granodiorites and porphyroblastic gneissic granites have been identified as well as equigranular "cloud" granites [13,49,88]. These rocks retain relict gneissic banding, skialiths and typically have replacement structures and textures. In contrast, the parautochthonous varieties intrude rocks of the epidote-amphibolite and greenschist facies, usually metasandstones, metasiltstones and metapelites, forming hornfelsed rims (Figures 3 and 7).
Syncollisional granitoids also include coarse-grained (up to pegmatoid) porphyroblastic biotite-amphibolite and two-mica granite (samples 203004 and 203009). These granites have a massive, spotted, sometimes gneissic texture. The spotted texture is due to aggregates of rust-red-brown biotite associating with graphite. The porphyroblasts, which can occupy up to 60-70% of the rock volume, are represented by microcline and plagioclase. Tabular feldspar porphyroblasts (up to 5-10 cm long) are commonly subparallel and, together with biotite, form the gneissic foliation. Plagioclase, perthitic potassium feldspar and quartz with undulose extinction are present in close quantities. In addition to these minerals, the groundmass of the rock contains hornblende, muscovite, garnet, titanite and other accessory minerals.

Major and Trace Element Analyses
The results of X-ray fluorescence and ICP-MS analyses for the studied rocks are given in Appendix B. The analyses were performed at the Central laboratory of A.P. Karpinsky Russian Geological Research Institute (VSEGEI) in Saint Petersburg (Russia) according to their proprietary methodology.

Major and Trace Element Analyses
The results of X-ray fluorescence and ICP-MS analyses for the studied rocks are given in Appendix B. The analyses were performed at the Central laboratory of A.P. Karpinsky Russian Geological Research Institute (VSEGEI) in Saint Petersburg (Russia) according to their proprietary methodology.
X-ray fluorescence measurements were obtained using an ARL-9800 XRF spectrometer (Thermo Fisher Scientific (Ecublens) SARL, Switzerland) with an X-ray tube fitted with a Rh anode target disk.

U-Pb SIMS Analytical Technique
The zircons were cast with Buehler EpoKwick resin, along with the Temora [90] and 91500 standards [91]. Then the zircons were half-sectioned and finally polished. Transmitted and reflected light images, as well as BSE and CL images were taken for guidance during analysis.
In-situ U-Pb analyses were performed using the SHRIMP-II SIMS in the Center of Isotopic Research at VSEGEI, applying a secondary electron multiplier in peak-jumping mode following the procedure described in [92]. The analytical data were acquired for nine mass-stations (from 196 Zr 2 O to 254 UO 2 ), with four to five mass-spectra per individual analysis. Each fourth to fifth analyses was done on the Temora reference zircons. The results were processed using the software SQUID v1.13 [93] and ISOPLOT/Ex 3.22 [94]. Common lead was corrected using the measured 204 Pb/ 206 Pb and the model values of Stacey and Kramers [95]. The results are presented in Appendix C. The ages discussed in the text and error ellipses in the diagrams are at 95% confidence, ratios and corresponding ages in the Table are given with 1σ errors.

Ar-Ar Dating
Ar-Ar analyses were performed following the procedure described in [96]. The studied monomineral fractions were wrapped in Al foil, placed into quartz ampoules, which were then pumped out and welded. Irradiation was done in the Cd-coated channel of the (BBP-K type) reactor at Tomsk Polytechnic University. Weighted standard samples of monomineral biotite fractions MCA-11 and LP-6 were used as mineral monitors. The gradient of the neutron flux did not exceed 0.5% of the sample size. The experiments using stepwise heating were carried out in a quartz reactor with an external heater. Released Ar was purified using two successive ZrAl SAES getters. The isotopic composition of Ar was measured on a Noble Gas 5400 mass spectrometer (Micromass, UK) in the Analytical Center of IGM SB RAS (Novosibirsk, Russia). The blank for 40 Ar in the stepwise heating unit was no higher than n × 10 -10 ncm 3 . The results of Ar/Ar measurements and calculations are presented in Appendix D.  (Figure 10), they correspond to rocks of the calc-alkalic and alkali-calcic magmatic series, with one exception occurring in the calcic series. On the ANK vs ACNK diagram ( Figure 11) the granitoids and diorites plot in the peraluminous field with an ACNK index range from 1.04 to 1.55. These samples show similarities in La, Ce, HREE, Hf, Zr, Nb, Ta, Sm, Tb, and Y concentrations with upper and lower continental crust (Figure 12a,b).

Major and Trace Element Geochemistry
The postcollisional intrusive rocks are less diverse in their contents of SiO 2 (66.7-74.5 wt.%) and have high K 2 O (4.44-5.33 wt.%) and total alkalis contents (7.99-8.99), moderate Na 2 O contents, low Al 2 O 3 contents (13.2-15.5 wt.%) and Na 2 O/K 2 O ratios in the range 0.58-0.83 (Appendix B). On the TAS diagram these samples plot in the fields of granites and quartz monzonites (Figure 9). Most of them correspond to rocks of the alkali-calcic magmatic series, with individual samples extending into the calc-alkalic series ( Figure 10). These rocks are weakly peraluminous (ACNK = 1.03-1.08) or weakly metaluminous (ACNK = 0.98) ( Figure 11). All postcollisional rocks are depleted in TiO 2 (0.58-0.06 wt.%). These samples demonstrate a similarity in concentrations of REE, Hf, Zr, Nb, Ta, Sm, Tb, and Y with the upper continental crust, although their concentrations of HREE are closely similar to the lower continental crust (Figure 12c,d).
Major and trace elements compositions of the studied igneous rocks are in good agreement with published data for the late Paleozoic collisional granites of Northern Taimyr [54,66] (Figures 9-12). Taking these data into account, it seems that in addition to alkali-calcic types, the postcollisional suite also contains alkalic rocks, similar to the plume-related Early and Middle Triassic monzodiorites, monzonites, syenites and quartz syenites of Northern Taimyr [34,56] (Figure 10). The latter (granites, quartz monzonites and syenites) are metaluminous varieties of the alkalic series with higher concentrations of Ba, K, Sr, La, Ce, Nd, P, Sm, Tb, and Yb than in the majority of the postcollisional rocks ( Figures 10-12).   The postcollisional intrusive rocks are less diverse in their contents of SiO2 (66.7-74.5 wt.%) and have high K2O (4.44-5.33 wt.%) and total alkalis contents (7.99-8.99), moderate Na2O contents, low Al2O3 contents (13.2-15.5 wt.%) and Na2O/K2O ratios in the range 0.58-0.83 (Appendix B). On the TAS diagram these samples plot in the fields of granites and quartz monzonites ( Figure 9). Most of them correspond to rocks of the alkali-calcic magmatic series, with individual samples extending into the calc-alkalic series ( Figure 10). These rocks are weakly peraluminous (ACNK = 1.03-1.08) or weakly metaluminous (ACNK = 0.98) ( Figure 11). All postcollisional rocks are depleted in TiO2 (0.58-0.06 wt.%). These samples demonstrate a similarity in concentrations of REE, Hf, Zr, Nb, Ta, Sm, Tb, and Y with the upper continental crust, although their concentrations of HREE are closely similar to the lower continental crust (Figure 12c,d).

Zircons Morphology and Internal Structure
The morphology and internal structures of the analyzed grains are displayed on Figures 13-15. Sample 13103. The zircon population comprises mainly subhedral grains, the majority of which are fractured and contain numerous inclusions. The zircons are colorless to brown, transparent to translucent, and faceted mainly by prisms and pyramids, whereas bipyramids are subordinate. . ANK (molar Al2O3/(Na2O + K2O)) versus ACNK (molar Al2O3/(CaO + Na2O + K2O)) diagram after [99] for the studied felsic rocks.

Zircons Morphology and Internal Structure
The morphology and internal structures of the analyzed grains are displayed on Figures 13-15. Sample 13103. The zircon population comprises mainly subhedral grains, the majority of which are fractured and contain numerous inclusions. The zircons are colorless to brown, transparent to translucent, and faceted mainly by prisms and pyramids, whereas bipyramids are subordinate.  [54,66]. Yellow fields are postcollisional subalkaline granites and granodiorites from [13]. Grey fields are Siberian LIP alkaline syenites from [34]. Dark grey lines are Siberian LIP felsic rocks from [56]. UCC and LCC-upper and lower continental crust data from [100]. Chondrite and primitive mantle values are from [101] and [102], respectively.
Major and trace elements compositions of the studied igneous rocks are in good agreement with published data for the late Paleozoic collisional granites of Northern Taimyr [54,66] (Figures 9-12). Taking these data into account, it seems that in addition to alkali-calcic types, the postcollisional suite also contains alkalic rocks, similar to the plume-related Early and Middle Triassic monzodiorites, monzonites, syenites and quartz syenites of Northern Taimyr [34,56] (Figure 10). The latter (granites, quartz monzonites and syenites) are metaluminous varieties of the alkalic series with higher concentrations of Ba, K, Sr, La, Ce, Nd, P, Sm, Tb, and Yb than in the majority of the postcollisional rocks (Figures 10-12).

Zircons Morphology and Internal Structure
The morphology and internal structures of the analyzed grains are displayed on Figures 13-15.

Zircon U-Pb Dating
Sample 13103. Among the ten analytical spots, no. 3.1 yielded a significantly young age, having the lowest U content and the highest Th. This analysis was excluded from the calculations. The other nine analyses form a concordant cluster with an age of 309 ± 1 Ma (MSWD value of 0.79 and a probability of 0.38), which is taken as the crystallization age of the rock (Figure 13a).
Sample 18032. The cluster of 10 analytical spots yields 207 Pb/ 206 Pb and 238 U/ 206 Pb ratios that are concordant. The concordia age is 305 ± 3 Ma (MSWD = 0.58 and probability = 0.44). The cluster can be Sample 204064. The zircons are dominantly euhedral to subhedral, transparent, colorless to pink and elongated. Variably shaped polyphase melt inclusions are common, some fractures follow the crystallographic forms, marking concentric growth zoning. CL images display fine to moderately broad bands of concentric growth zoning, with convolute zones around large inclusions. Some structurally disconformable patches are observed, probably as a result of a grain break up, although others may be somewhat rounded fragments of xenocrysts.
Sample 203004. The zircon population comprises euhedral to subhedral grains, transparent to translucent, pinkish to brown in color. The grains are fractured and contain some solid-phase inclusions. CL images demonstrate the presence of disconformable domains surrounded by rims of variably pronounced fine oscillatory zoning, with the outermost bands being CL-dark (non-zoned). Some of them contain inherited cores. The zircon grains features imply their magmatic genesis.
Sample 203009. The zircons are euhedral to subhedral elongated bi-pyramid-prismatic grains with both pyramidal and dipyramidal faces, transparent to translucent, and colorless to brown. The zircons contain various inclusions and have fractures, marking oscillatory growth zoning as well as some irregular fractures. CL imaging reveals the weakly pronounced growth zoning of oscillatory concentric and patchy types, suggesting a magmatic origin.
Sample 203025. Only six fragments of transparent colorless, apparently grains were obtained from this sample. They demonstrate rather weak to no internal CL zoning, suggesting a rapid growth, which is supported by the presence of relatively large melt inclusions. These features imply a magmatic origin.
Sample 554. This sample is dominated by euhedral to subhedral, short to moderately long grains, faceted by combinations of two prisms, pyramids and bipyramids. The zircons are commonly transparent and light-colored. Faint concentric growth zoning is locally observed in transmitted light, whereas bubble-and needle-shaped inclusions are common, but fractures are rather sparse. CL images reveal the prevalence of medium-broad concentric growth zoning, with some central parts having somewhat rounded outlines that most probably resulted from partial resorption of the early zircon nuclei. These features suggest their growth from a melt.
Sample 204025-4. The majority of the zircons are of elongate habit, transparent, colorless to pink. Facets are formed by two prisms with (101) pyramids and bipyramids. The zircons contain numerous acicular and stubby mineral inclusions along with melt inclusions of various shapes. Fine to intermediate broad-banded oscillatory growth zoning supports their magmatic origin.
Sample 204001. The zircons are dominantly euhedral, transparent, light to brownish-colored with two prisms and (101) pyramids. Acicular mineral and variably shaped melt inclusions are common. The inner CL images commonly reveal concentric, fine to moderately broad banding, indicative of their magmatic origin.

Zircon U-Pb Dating
Sample 13103. Among the ten analytical spots, no. 3.1 yielded a significantly young age, having the lowest U content and the highest Th. This analysis was excluded from the calculations. The other nine analyses form a concordant cluster with an age of 309 ± 1 Ma (MSWD value of 0.79 and a probability of 0.38), which is taken as the crystallization age of the rock (Figure 13a).  1, 1.1, 4.1, 2.1, 3.1, 6.1) form a concordant cluster with a slightly older age of 309 ± 3 Ma, an MSWD of 0.78 and a probability of 0.38, although this overlaps within error of the total population. Since there is no obvious anomaly in any of the results, and since the difference between two ages falls in the margin of error, we take total concordia age to be the best estimate of the crystallization age of the rock (Figure 13b).
Sample 204064. Analytical spot 1.1 has a very high 206 Pb* content and yielded a middle Paleoproterozoic age. The zircon's morphology clearly indicates it is a xenocryst. Ten analytical spots from 10 other zircon grains yielded concordant 207 Pb/ 206 Pb and 238 U/ 206 Pb ratios and form a cluster with an MSWD equal to 4.4 and probability equal 0.036 and a concordia age of 303 ± 2 Ma, which is taken as the crystallization age of the rock (Figure 13c).
Sample 203004. All ten analytical spots plot on concordia in terms of 207 Pb/ 206 Pb and 238 U/ 206 Pb ratios (Figure 14a). However, there is a significant scatter of ages along the curve. Analyzes 7.1 and 4.1 have to be excluded as they are significantly older than the other results. The remaining eight spots form a cluster with a concordia age of 287±2 Ma, MSWD of 0.34 and probability of 0.56. Excluding two other analyses with the largest error ellipses gives nearly the same age of 287 ± 3 Ma, but a slightly lower MSWD of 0.26 and a probability of 0.61.
Sample 203009. Nine of ten analyses plot on concordia and form a cluster with an age of 283 ± 1 Ma, MSWD of 0.16 and probability of 0.69 (Figure 14b). Spot 8.1 has an anomalously high 207 Pb/ 206 Pb ratio and % 206 Pb c and was excluded from the calculation.
Sample 203025. Seven analytical spots from six zircon grains yielded 207 Pb/ 206 Pb and 238 U/ 206 Pb ratios close to concordant and form a cluster with a concordia age of 282 ± 2 Ma (Figure 14c). The MSWD value is 1.2, and the probability is 0.26. The concordia age is taken as the crystallization age for this rock.
Sample 554. Ten analytical spots yielded 207 Pb/ 206 Pb and 238 U/ 206 Pb ratios close to concordant and form a cluster with a concordia age of 256 ± 5 Ma (Figure 15a). The MSWD value is 0.36, with a probability of 0.55. There are no significant deviations in isotope ratio values, so we determine this date as the crystallization age for this rock.
Sample 204025-4. Ten analytical spots yielded 207 Pb/ 206 Pb and 238 U/ 206 Pb ratios close to concordant and form a cluster with a concordia age of 255 ± 2 Ma (Figure 15b). Spots 2.1, 3.1 and 4.1 (all taken from zircon rims) show elevated U contents. The MSWD value of 3.5 is somewhat high with a probability of 0.062. The concordia age is taken as the crystallization age for this rock.
Sample 204001. Among the twelve analytical spots, three (6.1, 4.1, 1.1) yielded anomalously young ages. The other nine spots form a concordant cluster with a concordia age of 252 ± 2 Ma, MSWD value of 0.099 and a probability of 0.75 (Figure 15c).

Discussion
For many years, there have been discussion on how and when did one of the key Arctic structures form at the junction between the Siberian Craton and the Kara microcontinent. New geostructural, geochemical and geochronological data for igneous rocks of the Kara orogen present us with the opportunity to have a deeper understanding of the tectonic processes that manifested along the northern margin of the Siberian Craton in the late Paleozoic to early Mesozoic time. We consider that the late Paleozoic granites and associating acidic and intermediate igneous rocks are, among others, the most useful source of information in this respect.
Our new data have strengthened our opinion that the Kara orogen granites can be divided into syncollisional and postcollisional types [13,32]. Moreover, we are confident that the first group is represented by anatectic granites formed by melting of the crust. It is also clear that various types of material melted in the upper and lower crust-calcareous-greywacke, greywacke, pelitic and basic. This is demonstrated by the variations of petrographic composition of the rocks-from biotite and biotite-hornblende granodiorite and diorite to biotite and two-mica granite and quartz porphyriesas well as by variations in their geochemical composition. These syncollisional granodiorites and granites are peraluminous and correspond to rocks of the calc-alkalic and alkali-calcic magmatic series with Na content exceeding K. That the formation of the syncollisional granites is related to crustal melting is confirmed by the R1-R2 discrimination diagram after [103] (Figure 17), on which these granites plot in the zone of anatectic granitic melts.
In contrast, the postcollisional biotite and biotite-amphibole granodiorites, subalkaline granites, granite-porphyries, quartz syenites and monzonites are weakly peraluminous to weakly metaluminous and belong to the alkali-calcic and alkalic magmatic series and are enriched in K and total alkalis, and slightly enriched in Ba and Sr. On the same diagram ( Figure 17) the postcollisional rocks from this study and from published data [13,50] clearly form a trend from the field of anatectic granitic melts towards monzonite and monzodiorite, with elevated alkalinity.
On the MALI vs SiO2 diagram by [98] (ref. Figure 10), the syncollisional Kara orogen granites plot mostly in the S-type field, and the postcollisional ones-mostly in the A-type field, however both groups significantly overlap the I-type field.

Discussion
For many years, there have been discussion on how and when did one of the key Arctic structures form at the junction between the Siberian Craton and the Kara microcontinent. New geostructural, geochemical and geochronological data for igneous rocks of the Kara orogen present us with the opportunity to have a deeper understanding of the tectonic processes that manifested along the northern margin of the Siberian Craton in the late Paleozoic to early Mesozoic time. We consider that the late Paleozoic granites and associating acidic and intermediate igneous rocks are, among others, the most useful source of information in this respect.
Our new data have strengthened our opinion that the Kara orogen granites can be divided into syncollisional and postcollisional types [13,32]. Moreover, we are confident that the first group is represented by anatectic granites formed by melting of the crust. It is also clear that various types of material melted in the upper and lower crust-calcareous-greywacke, greywacke, pelitic and basic. This is demonstrated by the variations of petrographic composition of the rocks-from biotite and biotite-hornblende granodiorite and diorite to biotite and two-mica granite and quartz porphyries-as well as by variations in their geochemical composition. These syncollisional granodiorites and granites are peraluminous and correspond to rocks of the calc-alkalic and alkali-calcic magmatic series with Na content exceeding K. That the formation of the syncollisional granites is related to crustal melting is confirmed by the R1-R2 discrimination diagram after [103] (Figure 17), on which these granites plot in the zone of anatectic granitic melts.
In contrast, the postcollisional biotite and biotite-amphibole granodiorites, subalkaline granites, granite-porphyries, quartz syenites and monzonites are weakly peraluminous to weakly metaluminous and belong to the alkali-calcic and alkalic magmatic series and are enriched in K and total alkalis, and slightly enriched in Ba and Sr. On the same diagram ( Figure 17) the postcollisional rocks from this study and from published data [13,50] clearly form a trend from the field of anatectic granitic melts towards monzonite and monzodiorite, with elevated alkalinity.
On the MALI vs SiO 2 diagram by [98] (ref. Figure 10), the syncollisional Kara orogen granites plot mostly in the S-type field, and the postcollisional ones-mostly in the A-type field, however both groups significantly overlap the I-type field. An even more significant overlap of the studied compositions of granitoids is observed on the discrimination diagrams from [105,106] (Figure 18), that use relationships between contents of Nb and Y, Ta and Yb, Rb and (Nb + Y). On the (Nb + Y) vs Rb diagram, the granitoids mostly occupy the field of postcollisional granites, which can indicate the diversity of their crustal magmatic sources. The latter is corroborated by the similarity in concentrations of most high field strength elements, including Ta, Nb, Y and Yb in the studied intrusive rocks with those concentrations from the upper and lower crust by [100] as seen by their distributions (Figure 12b,d).
Our new U-Pb data for zircon (SHRIMP-II SIMS) obtained for the Kara orogen leads us to divide the granitoids into three age groups: late Carboniferous syncollisional with ages of 309 ± 1 Ma, 305 ± 3 Ma and 303 ± 2 Ma; early Permian with ages of 287 ± 3 Ma, 283 ± 1 Ma and 282 ± 2 Ma; and late Permian postcollisional with ages of 256 ± 5, 255 ± 2 Ma and 252 ± 2 Ma. Taking into consideration the previously published data for the ages of these granitoids [13,32,[50][51][52][53], the following age groups can be determined: 315-304 Ma, 287-282 Ma and 264-248 Ma. Thus, the interval between the three main pulses of granitoid magmatism is approximately 20 m.y. In our opinion, this corresponds to a single continent-microcontinent collisional process.
Granitic magmatism began due to the collision of continental masses, with thickening of the crust, accompanied by folding that migrated to the south (in present-day geographic coordinates). However, strike-slip processes were dominating, and the interaction between Siberia and the Kara microcontinent was an oblique collision [20,22,35,67]. The Kara microcontinent continued to rotate counterclockwise relative to Siberia and to slide along the margin of Siberia (Figure 19), which is what probably caused the second magmatic pulse. The formation of postcollisional intrusions took place after the Kara microcontinent had slowed down significantly in relation to the Siberian paleocontinent. Note that the syncollisional granites were emplaced in the western and central part of the region, and the postcollisional ones were mainly in the eastern part (Cape Chelyuskin). Judging from the Ar-Ar isotope data, the cooling of the collisional granites in the orogen due to the closing temperature for the Ar-Ar system for biotite and amphibole (350-500 °C) lasted until 246-245 Ma (see Figure 16). An even more significant overlap of the studied compositions of granitoids is observed on the discrimination diagrams from [105,106] (Figure 18), that use relationships between contents of Nb and Y, Ta and Yb, Rb and (Nb + Y). On the (Nb + Y) vs Rb diagram, the granitoids mostly occupy the field of postcollisional granites, which can indicate the diversity of their crustal magmatic sources. The latter is corroborated by the similarity in concentrations of most high field strength elements, including Ta, Nb, Y and Yb in the studied intrusive rocks with those concentrations from the upper and lower crust by [100] as seen by their distributions (Figure 12b,d).
Our new U-Pb data for zircon (SHRIMP-II SIMS) obtained for the Kara orogen leads us to divide the granitoids into three age groups: late Carboniferous syncollisional with ages of 309 ± 1 Ma, 305 ± 3 Ma and 303 ± 2 Ma; early Permian with ages of 287 ± 3 Ma, 283 ± 1 Ma and 282 ± 2 Ma; and late Permian postcollisional with ages of 256 ± 5, 255 ± 2 Ma and 252 ± 2 Ma. Taking into consideration the previously published data for the ages of these granitoids [13,32,[50][51][52][53], the following age groups can be determined: 315-304 Ma, 287-282 Ma and 264-248 Ma. Thus, the interval between the three main pulses of granitoid magmatism is approximately 20 m.y. In our opinion, this corresponds to a single continent-microcontinent collisional process.
Granitic magmatism began due to the collision of continental masses, with thickening of the crust, accompanied by folding that migrated to the south (in present-day geographic coordinates). However, strike-slip processes were dominating, and the interaction between Siberia and the Kara microcontinent was an oblique collision [20,22,35,67]. The Kara microcontinent continued to rotate counterclockwise relative to Siberia and to slide along the margin of Siberia (Figure 19), which is what probably caused the second magmatic pulse. The formation of postcollisional intrusions took place after the Kara microcontinent had slowed down significantly in relation to the Siberian paleocontinent. Note that the syncollisional granites were emplaced in the western and central part of the region, and the postcollisional ones were mainly in the eastern part (Cape Chelyuskin). Judging from the Ar-Ar isotope data, the cooling of the collisional granites in the orogen due to the closing temperature for the Ar-Ar system for biotite and amphibole (350-500 • C) lasted until 246-245 Ma (see Figure 16). It should be noted that at the same time as the postcollisional magmatism was coming to an end (252-248 Ma according to published research and our new U-Pb and Ar-Ar isotope data), the main phase of Siberian plume trap magmatism manifested in the Kara orogen as early as 251-249 Ma (see Figure 1). The age of dolerites and basalts from the Tunguska syneclise and of differentiated trap intrusions from Taimyr [45,64] fully correspond to this main phase of the Siberian plume. These bodies, together with individual plutons of syenite and monzonite intruding the trap basalts and collisional granites, formed in the Kara orogen lithosphere that was well-heated by collisional processes in the Early-Middle Triassic [34,42,43,56]. In contrast to the collisional intrusive rocks of the Kara orogen, these monzonites, quartz monzonites and syenites that formed due to the Siberian plume are characterized by a mixed crustal and mantle sources, and are metaluminous and the most enriched in Ba and Sr (see Figures 9-12).
Such a space-time connection of collisional and plume magmatism in the northwestern (Arctic) margin of the Siberian Craton allows us to determine that the plumbing system for magmatic melts of the plume was emplaced due to the oblique collision of the Siberian paleocontinent and the Kara microcontinent in the late Paleozoic along the sutures of the Kara orogen and the Yenisei-Khatanga rift, being the weakest zones of the region. That is to say, this collision became a "trigger" for the outpouring of the Siberian traps at the Permian-Triassic boundary. The structural association of the late-stage Siberian LIP alkaline felsic plutons with old suture zones has been previously described for the Neoproterozoic orogen in the Yenisei Ridge [107].
At the same time, there is still a discussion in the literature on the questions concerning the time of collision between the Siberian paleocontinent and the Kara microcontinent that were formulated over 30 years ago (e.g., [12,13,16,17,19,20,23,25,27,53,63,73,[108][109][110][111]). The authors participating in the discussion believed that up until the Middle Jurassic, the South Anyui basin-a bay of the Paleo-Pacific Ocean (Panthalassa)-existed and separated the Chukotka and Kara continental units from Siberia. Chukotka and Kara both were parts of the Arctida continent at the time, as proposed by It should be noted that at the same time as the postcollisional magmatism was coming to an end (252-248 Ma according to published research and our new U-Pb and Ar-Ar isotope data), the main phase of Siberian plume trap magmatism manifested in the Kara orogen as early as 251-249 Ma (see Figure 1). The age of dolerites and basalts from the Tunguska syneclise and of differentiated trap intrusions from Taimyr [45,64] fully correspond to this main phase of the Siberian plume. These bodies, together with individual plutons of syenite and monzonite intruding the trap basalts and collisional granites, formed in the Kara orogen lithosphere that was well-heated by collisional processes in the Early-Middle Triassic [34,42,43,56]. In contrast to the collisional intrusive rocks of the Kara orogen, these monzonites, quartz monzonites and syenites that formed due to the Siberian plume are characterized by a mixed crustal and mantle sources, and are metaluminous and the most enriched in Ba and Sr (see Figures 9-12).
Such a space-time connection of collisional and plume magmatism in the northwestern (Arctic) margin of the Siberian Craton allows us to determine that the plumbing system for magmatic melts of the plume was emplaced due to the oblique collision of the Siberian paleocontinent and the Kara microcontinent in the late Paleozoic along the sutures of the Kara orogen and the Yenisei-Khatanga rift, being the weakest zones of the region. That is to say, this collision became a "trigger" for the outpouring of the Siberian traps at the Permian-Triassic boundary. The structural association of the late-stage Siberian LIP alkaline felsic plutons with old suture zones has been previously described for the Neoproterozoic orogen in the Yenisei Ridge [107].  At the same time, there is still a discussion in the literature on the questions concerning the time of collision between the Siberian paleocontinent and the Kara microcontinent that were formulated over 30 years ago (e.g., [12,13,16,17,19,20,23,25,27,53,63,73,[108][109][110][111]). The authors participating in the discussion believed that up until the Middle Jurassic, the South Anyui basin-a bay of the Paleo-Pacific Ocean (Panthalassa)-existed and separated the Chukotka and Kara continental units from Siberia. Chukotka and Kara both were parts of the Arctida continent at the time, as proposed by [16,17]. These authors were of the opinion that this part of Arctida collided with Paleo-Siberia only in the Mesozoic, which led to the formation of the fold-and-thrust structure of Southern Taimyr, to deformation of the Early Triassic traps and of the Jurassic sediments at Cape Tsvetkov in Eastern Taimyr.
As we have shown above, the occurrence of Early Triassic dolerite dikes of the Siberian plume traps in the entire orogen (Southern, Central and Northern tectonostratigraphic domains), including Northern Taimyr and Severnaya Zemlya, as well as the existence of Early-Late Triassic syenites and monzonites in Northern Taimyr, introduce additional constraints for paleogeodynamic reconstructions of this region. This information, together with paleomagnetic data [19,20], contradicts the proposed existence of an oceanic bay between the Siberian paleocontinent and the Kara microcontinent until the Middle Jurassic. Our studies show that by the Permian-Triassic boundary, when the Siberian plume magmatism was initiated, the Siberian paleocontinent and the Kara microcontinent already formed a single unit.
The causes of deformation of the trap complex and the Jurassic sediments of Cape Tsvetkov can probably be found in the kinematics of adjacent structures. In the Late Jurassic, deformation in the Verkhoyansk fold belt had begun due to the collision of the Kolyma-Omolon superterrane with the Siberian paleocontinent [15]. These events might have caused compressional deformation in the Taimyr-Severnaya Zemlya fold-and-thrust belt [112,113]. The Amerasian Basin is considered to have opened during the Early Jurassic-Early Cretaceous [114][115][116], whereas from the other side of Siberia, the West Siberian basin began forming due to a system of north-south trending rift systems, causing compressional deformation in the Kara orogen.

Conclusions
Our new geostructural, geochemical, geochronological data for igneous rocks of the Kara (Taimyr-Severnaya Zemlya) orogen, when combined with the analysis of published research results, allow us to present a number of conclusions based on all of these studies: 1.
The late Carboniferous-Permian granitoids widely represented in the Kara orogen were formed due to oblique collision of the Siberian paleocontinent and the Kara microcontinent; 2.
The syncollisional granites (315-282 Ma) were formed in areas of anatectic melting of crustal material and were localized mainly in zones of amphibolite metamorphism and migmatization. They are mostly peraluminous biotite-hornblende granodiorites or porphyroblastic biotite-amphibole and two-mica granites, with characteristics of the calc-alkalic and alkali-calcic magmatic series; 3.
The postcollisional granites (264-248 Ma) form mainly small stocks intruding metamorphic rocks and syncollisional granites, along with unmetamorphosed Paleozoic cover deposits in the Central Domain. They are biotite-amphibole granodiorites, subalkaline granites, granite-porphyries, quartz syenites and monzonites, weakly peraluminous to weakly metaluminous, and belong to the alkali-calcic and alkalic magmatic series. These rocks are enriched in K and total alkalis and slightly enriched in Ba and Sr. 4.
The granites of the Kara orogen can be divided into three age groups: The duration of granite formation corresponds to one single collisional process of the continent-microcontinent type. We explain the 20 m.y. time gap between magmatic pulses in the late Carboniferous and the early Permian to the effect of the continuing counterclockwise rotation of the Kara microcontinent relative to Siberia, while at the same time sliding along its margin.

5.
At the same time as the postcollisional magmatism was ending, in the Early Triassic, differentiated and individual intrusions related to the Siberian plume began intruding the lithosphere of the Kara orogen that was well-heated through collisional processes. They are metaluminous monzonites, quartz monzonites and syenites with a mixed crustal-mantle source, characteristic of the calc-alkalic and alkalic magmatic series and are the most enriched in Ba and Sr. 6.
The oblique collision of the Siberian paleocontinent and the Kara microcontinent at the end of the Paleozoic led to the opening of the plumbing system for the magmatic melts of the Siberian plume along the sutures of the Kara orogen and the Yenisei-Khatanga rift because these were the weakest zones. Therefore, this collision became a "trigger" for the outpouring of the Siberian traps at the Permian-Triassic boundary. 7.
The occurrence of Early Triassic dolerite dikes of the Siberian plume traps across the entire orogen (Southern, Central and Northern tectonostratigraphic domains), including Northern Taimyr and Severnaya Zemlya, as well as the existence of Early-Late Triassic syenites and monzonites in northern Taimyr, introduce additional constraints for the paleogeodynamic reconstruction of this region. Our studies show that by the Permian-Triassic boundary, when the Siberian plume magmatism was initiated, the Siberian paleocontinent and the Kara microcontinent had already formed a single unit.  Note: * volume of argon in cm 3 occupied under normal conditions, ** parameter characterizing the neutron flux intensity.