A Rhaetian 40Ar/39Ar age for the Rochechouart impact structure (France) and implications for the latest Triassic sedimentary record


Corresponding author. E-mail: martin.schmieder@geologie.uni-stuttgart.de


Abstract–40Ar/39Ar dating of potassium feldspar (primary spherulitic-blocky and secondary idiomorphic K-feldspar) separated from impact-metamorphosed gneiss found near Videix in the western central part of the Rochechouart impact structure (NW Massif Central, France) yielded a Rhaetian combined age of 201 ± 2 Ma (2σ), indistinguishable within uncertainty from the age of the Triassic/Jurassic boundary. Ballen quartz intergrown with the primary K-feldspar indicates post-shock temperatures exceeding approximately 1000 °C that affected the precursor gneiss. Geochemically, both feldspar types represent essentially pure potassium end-members. Apart from the approximately 15 km diameter impact deposit area, the youngest crystallization age known for basement rocks in this part of the Massif Central is approximately 300 Ma. No endogenic magmatic-thermal events are known to have occurred later in this region. The K-feldspar recrystallized from local feldspar melts and superimposed post-shock hydrothermal crystallization, probably within some thousands of years after the impact. It is, therefore, suggested that the 40Ar/39Ar age for the Videix gneiss (as a potassic “impact metasomatite”) dates the Rochechouart impact, in consistence with evidence for K-metasomatism in the Rochechouart impactites. The new age value is distinctly younger than the previously obtained Karnian–Norian age for Rochechouart and, thus, contradicts the Late Triassic multiple impact theory postulated some years ago. In agreement with the paleogeographic conditions in the western Tethys domain around the Triassic/Jurassic boundary, the near-coastal to shallow marine Rochechouart impact is compatible with the formation of seismites and tsunami deposits in the latest Triassic of the British Isles and possible related deposits in other parts of Europe.


The approximately 20–25 km in diameter, complex Rochechouart impact structure (France, départements Haute-Vienne and Charente; centered at 45°50′N, 0°56′E) is hosted by Precambrian to Paleozoic (Variscan) crystalline rocks of the northwestern French Massif Central, predominantly gneisses, leptynites, and granites, as well as some granitic to dioritic dikes (e.g., Chèvremont et al. 1996; Lambert 2010). As the Rochechouart impact structure is deeply eroded down to the crater floor, no primary crater morphology is preserved, and estimates for the pre-erosional original diameter of the impact structure are approximately 40–50 km (Fig. 1) (Lambert 1977, 2010). The shocked crater basement is characterized by the occurrence of shatter cones, as well as lithic and pseudotachylitic breccia dikes. Crater-filling impact lithologies comprise an approximately 60 m pile of basal lithic impact breccias (∼30 m in thickness) overlain by lenses of impact melt rocks (up to ∼10 m) and suevite (∼20 m), finally topped by a fine-grained, ash-like impactoclastic layer. The Rochechouart impactites show a distinct enrichment of potassium with respect to the unshocked surrounding target rocks, which is ascribed to impact-induced K-metasomatism (e.g., Lambert 1977, 2010; Reimold et al. 1984). A geologic summary of the Rochechouart impact structure and associated impactites is given by Lambert (1977), Reimold and Oskierski (1987), Kelley and Spray (1997), and Lambert (2008, 2010). Most recently, geochemical analyses of impact melt rocks suggested a nonmagmatic iron meteorite as the Rochechouart impactor (Tagle et al. 2009). Although Rochechouart was early confirmed as of impact origin by Kraut (1969) and studied through decades, the Rochechouart impact age has been a matter of debate for a long time. Previous investigations, including K–Ar (Kraut and French 1971; Lambert 1974), Rb–Sr (Reimold and Oskierski 1987), apatite and glass fission track (Wagner and Storzer 1975), as well as paleomagnetic dating (Pohl and Soffel 1971; Carporzen and Gilder 2006), resulted in a broad (Middle Triassic to Late Jurassic) time window for the Rochechouart impact. 40Ar/39Ar laser dating of pseudotachylites from the Champagnac quarry in the northeastern part of the impact structure yielded a Karnian to Norian 214 ± 8 Ma age (Kelley and Spray 1997; see also Fig. 2 for summary), which promoted the idea that Rochechouart could be a member of a ∼214 Ma and ∼4500 km long terrestrial impact crater chain that includes the impact structures of Rochechouart, Manicouagan (Québec, Canada), Lake Saint Martin (Manitoba, Canada), Obolon (Ukraine), and Red Wing Creek (North Dakota, USA), as well as possibly Wells Creek (Tennessee, USA) and Newporte (North Dakota, USA; Spray et al. 1998; Walkden et al. 2002). However, the formation of crater chains is considered highly improbable to occur on Earth (Bottke et al. 1997; Melosh 1998; Richardson et al. 1998), and the Late Triassic multiple impact theory has been questioned by some authors (Kent 1998; Schmieder and Buchner 2008; Wartho et al. 2009). Precise and accurate dating is critical to this question and, furthermore, essential regarding paleoenvironmental considerations for the Rochechouart impact scenario (i.e., proximity and possible influence of the nearby Triassic to Jurassic shoreline and/or sediments upon crater formation and vice-versa). In this study, we present new 40Ar/39Ar age data for the Rochechouart impact structure and discuss the results with respect to high-energy paleoenvironmental events (earthquakes and tsunamis) and possible distal impact ejecta in the latest Triassic sedimentary record of Europe.

Figure 1.

 Position of the Rochechouart impact structure at the northwestern margin of the French Massif Central and surrounding sedimentary basins (see map inset for position of the scene). Solid circle: present diameter of the eroded impact structure (23 km); dotted circle: possible pre-erosional diameter (∼40 km). R = city of Rochechouart; V = village of Videix. Scene computed from Shuttle Radar Topographic Mission data.

Figure 2.

 Ages so far reported for the Rochechouart impact structure. Stage boundary ages after Ogg et al. (2008).

Samples and Analytical Techniques

Petrography and Geochemistry

Optically fresh aggregates of spherulitic-blocky (primary) and idiomorphic (secondary) K-feldspar separated from impact-metamorphosed gneiss found near the village of Videix in the western central part of the Rochechouart impact structure (see Fig. 1 for location) were chosen as material for dating. Sampling was done during a field trip in summer 2005 (E. B. and M. S.). The modal and geochemical composition of the rock was determined by optical microscopy, energy dispersive X-ray (EDX) measurements (CamScan™ SC44 scanning electron microscope-EDAX™ PV 9723/10 system; 15 kV; Institut für Planetologie, Universität Stuttgart), electron microprobe analyses (CAMECA™ SX 100 microprobe; acceleration voltage 15 kV; beam current 10 nA; beam diameter 5 μm; microprobe silicate compound standards: Na: albite; K: orthoclase; Mg: periclase; Al: corundum; Ba: barite; Ti: rutile; Mn: rhodonite; Ca and Si: wollastonite; Fe: hematite; Institut für Mineralogie and Kristallchemie, Universität Stuttgart), and supplementary X-ray diffractometry (XRD; Bruker-AXS D8 affiliation; Institut für Mineralogie and Kristallchemie, Universität Stuttgart).

The reddish shocked gneiss shows a partially fluidal texture that overprints a relict foliation fabric (Figs. 3A and 3B) and is mainly composed of potassium feldspar (Figs. 3C and 3D), silica (Figs. 3E and 3F), and some accessory minerals. Potassium feldspar occurs as both “primary” spherulitic to blocky crystals (Figs. 3C and 3D) intergrown with each other and silica within the rock groundmass, as well as “secondary” idiomorphic K-feldspar of the morphological adularia type grown along rims of cavities (Figs. 3D, 4, and 5). Silica forms commonly vesicular microcrystalline masses of α-quartz that largely retained the original clast shape and display relict shock metamorphic features (Fig. 3E) and domains of toasted appearance that locally consist of ballen quartz (Figs. 3F and 5). Accessories predominantly comprise hematite crystallites (partially distributed within spherulitic-blocky K-feldspar and accounting for elevated iron content; Figs. 3D and 6) and zircon. EDX measurements under defocused beam yielded a siliceous-potassic whole-rock composition of the Videix gneiss (∼66 wt% SiO2; ∼11 wt% K2O; ≤0.2 wt% Na2O; <0.1 wt% CaO), similar to that of impact melt rocks found in the central part of the Rochechouart impact structure (e.g., Oskierski and Bischoff 1983; Lambert 2010). Both spherulitic-blocky primary (Or98–100Ab0-2) and idiomorphic secondary (Or99–100Ab0-1) K-feldspar types represent essentially pure potassium end-members, as revealed by microprobe analyses (Tables 1 and 2). Silica locally contains ≤1 wt% in Al2O3. The characteristic peaks of the Videix gneiss K-feldspars match those of (natural and pure synthetic) sanidine, adularia, α-quartz, and hematite in the X-ray diffractogram (see Fig. S1, Supporting Information; also compare Kroll et al. 1986).

Figure 3.

 Impact-metamorphosed gneiss from Videix. A) Polished slice of reddish gneiss with larger domains of recrystallized K-feldspar (center left). B) Overall thin section image showing a partially fluidal relict gneiss fabric (plane polarized light). C) Intergrown crystals of primary groundmass K-feldspar (Kfs; cross polarized light). D) Locally spherulitic aggregate of primary K-feldspar (Kfs) in the rock groundmass, microcrystalline hematite (Hem), and secondary idiomorphic crystals of K-feldspar (adularia; Adu) grown in a cavity (cross polarized light). E) Clast of recrystallized microcrystalline α-quartz (Qtz) with relict shock features (planar fractures; cross polarized light). F) Vesicular aggregate of silica with domains of microcrystalline α-quartz (a) and ballen α-quartz in locally toasted domains (b) (cross polarized light).

Figure 4.

 Scanning electron microscope (secondary electron) image of a cavity filled with idiomorphic adularia; rim is mainly spherulitic-blocky K-feldspar and microcrystalline quartz.

Figure 5.

 Electron microprobe (backscattered electron) image of the Videix gneiss showing primary spherulitic-blocky K-feldspar (Kfs; light gray) partially intergrown with microcrystalline quartz (darker gray), larger domains of microcrystalline quartz (Qtz), and secondary idiomorphic K-feldspar (adularia; Adu) grown in cavities.

Figure 6.

 Geochemical composition (microprobe data; compare Tables 1 and 2) of primary and secondary K-feldspar in the Videix gneiss. Note the slightly higher sodium and iron content in the primary K-feldspars (triangular plots range from approximately 80 to 100 wt% K2O; compare geochemical data given in Tables 1 and 2).

Table 1.   Major element composition of primary spherulitic-blocky K-feldspar in the impact-metamorphosed Videix gneiss (10 electron microprobe measurements; values in wt%).
TiO20.01n.d.n.d.n.d.n.d.0.090 .09n.d.0.010.310.05
Table 2.   Major element composition of secondary idiomorphic K-feldspar in the impact-metamorphosed Videix gneiss (10 electron microprobe measurements; values in wt%).

40Ar/39Ar Dating

Aggregates of spherulitic-blocky and idiomorphic K-feldspar (up to ∼0.5 mm in size, with K contents of ∼2 and ∼7 wt%, respectively, as measured within the extraction line) were handpicked from the crushed rock on the basis of their fresh optical appearance and rinsed with ethanol and deionized water in an ultrasonic cleaner. 40Ar/39Ar dating was done at the Institut für Geowissenschaften, University of Heidelberg (see Trieloff et al. 2005; Schwarz and Trieloff 2007a, for technical details), using the decay constants determined by Steiger and Jäger (1977); due to miscalibration of the K-decay constants compared to the U-decay series constants, the 40Ar/39Ar ages have to be approximately 1% higher in the age range of the Mesozoic with respect to those calculated with the constants recommended in Steiger and Jäger (1977; see also Villeneuve et al. 2000; Ramezani et al. 2005; Mundil et al. 2006; Schwarz and Trieloff 2007b; Jourdan et al. 2009a). The samples were wrapped in Al-foil, packed in quartz ampoules (12 × 55 mm), Cd-shielded (to avoid isotopic interferences), and irradiated for 5 days at the GKSS Geesthacht, Germany. After irradiation, the samples were heated in an inductively heated Mo-furnace up to approximately 1500 °C in several temperature steps. The gas was cleaned with Zr-, Ti-, Cu/CuO-, and Al-Getters. Argon was measured using a Finnigan MAT GD150 mass spectrometer (Thermo Scientific–Finnigan MAT, Bremen, Germany). The argon results were corrected for mass discrimination, argon decay, isotopic interferences (all information listed in the isotopic Tables 3 and 4), and blanks from the extraction line (glass and furnace). The argon isotopic composition of blanks was always atmospheric within uncertainty. Ages were calculated against the BMus/2 standard with an age of 328.5 ± 1.1 Ma, which was calibrated against five further international Ar-Ar dating standards; for further information about the analytical procedure see Schwarz and Trieloff (2007a). All error assignments given in this study are within 2σ error limits, unless stated differently.

Table 3.   Isotopic argon data set and ages for primary K-feldspar separated from impact-metamorphosed gneiss, Videix, Rochechouart impact structure; errors in parentheses include error on the standard; bold notations indicate data included in the plateau calculation.
Step #Temp (°C)36Ar (fA)±1σ37ArCa (fA)±1σ38Ar (fA)±1σ39ArK (fA)±1σ40Ar (fA)±1σ40Ar* (fA)±1σ36Ar/40Ar (×10−6)±1σ39Ar/40Ar (×10−5)±1σ39Ar (%)Age (Ma)±1σ
 15007.760.13n.d.n.d.1.510.061.380.112312.21.518.738.233575659.64.90.1567135 (135)
 26003.120.0725.870.521.110.0423.300.171628.33.6705.321.01918441431112.52146.74.3 (4.3)
 36902.150.2010. (1.9)
47401.660.1511.530.233.150.15161.800.527376.626.26886.951.12252021931017.53203.01.6 (1.7)
57600.960.0615.620.311.930.1793.060.044225.814.43942.821.2227142202810.08202.11.0 (1.3)
68001.140.08n.d.n.d.1.850.0891.790.254235.814.83897.625.3270182167109.95202.51.3 (1.5)
78601.280.06n.d.n.d.2.830.26125.380.205630.915.45252.918.2227102227713.59200.00.7 (1.0)
89300.680.12n.d.n.d.1.710.11101.510.324477.428.94277.741.21512722671611.00201.11.9 (2.1)
910000.810.09n.d.n.d.1.450.1563.730.122884.723.52645.018.2281312209186.91198.21.3 (1.5)
1011001.450.1318.720.391.500.1566.030.213197.040.02767.514.9455412065277.15200.01.2 (1.4)
11120010.480.25n.d.n.d.2.450.1432.870.154498.967.51403.032.5232965731113.56203.54.6 (4.6)
12135042.370.529.630. (19)
Weighted plateau: 201.4 ± 1.2 (1.4) Ma (1σ), n = 8 steps, MSWD = 1.07, P = 0.38J = 0.0027979 ± 0.0000015Std: BMus/2; t = 328.5 ± 1.1 Ma100.00199.41.2 (1.4)
Table 4.   Isotopic argon data set and ages for secondary K-feldspar separated from impact-metamorphosed gneiss, Videix, Rochechouart impact structure; errors in parentheses include error on the standard; bold notations indicate data included in the plateau calculation.
Step #Temp (°C)36Ar (fA)±1σ37ArCa (fA)±1σ38Ar (fA)±1σ39ArK (fA)±1σ40Ar (fA)±1σ40Ar* (fA)±1σ36Ar/40Ar (×10−6)±1σ39Ar/40Ar (×10−5)±1σ39Ar (%)Age (Ma)±1σ
 15008.170.1012.420.251.830.241.280.192459.33.044.929.433224051.87.90.08170109 (109)
 26003.510.114.300.091.640.2349.920.052672.16.31634.933.0131442186853.15158.33.1 (3.1)
 36652.540.1010.080.213.380.18166.330.297342.18.36591.129.8346132265510.49189.90.9 (1.1)
46901.690.1716.060.363.130.26163.760.137296.07.86797.649.4231232244310.33198.41.4 (1.5)
57002.950.1413.650.274.470.07210.240.269700.116.68828.443.8304152167513.26200.61.0 (1.2)
67151.030.1620.770.462.240.4099.700.274489.316.04184.249.5230362221106.29200.52.3 (2.4)
77400.980.0816.560.332.000.1489.330.164036.112.43745.022.124419221395.63200.31.2 (1.4)
88001.390. (1.6)
98801.130.143.600.093.480.16183.560.398059.730.87725.339.51401822771011.58201.01.1 (1.3)
109500.490.265.420.143.060.19168.470.247228.437.97082.965.6683523311310.63200.81.8 (1.9)
111030n.d.0.224.670.162.110.17128.660.215401.514.85401.514.8n.d.n.d.2382298.12200.50.6 (0.9)
121170n.d.0.565.670.252.750.38162.310.286834.260.06834.260.0n.d.n.d.23756010.24201.11.7 (1.8)
1313306.691.464.150.111.930.2936.160.2435034311526.536.0190947910321272.28201.64.7 (4.7)
14145003.1512.040.25−0.870.622.290.0770.825.470.825.4n.d.n.d.323411600.1414952 (52)
Weighted plateau: 200.5 ± 1.1 (1.3) Ma (1σ), n = 10 steps, MSWD = 0.21, P = 0.99J = 0.0028005 ± 0.0000011Std: BMus/2; t = 328.5 ± 1.1 Ma100.00198.01.0 (1.2)

Dating Results and Interpretation

40Ar/39Ar step-heating analyses yielded two plateau ages of 201.4 ± 2.4 Ma (2σ; ±2.8 Ma including standard error; MSWD = 1.07, P = 0.38) for primary spherulitic-blocky K-feldspar (∼19–99% of 39Ar released; steps 4–11 of 12 extractions shown in Table 3) and 200.5 ± 2.2 Ma (2σ; ±2.6 Ma including standard error; MSWD = 0.21, P = 0.99) for secondary idiomorphic K-feldspar (∼14–100% of 39Ar; steps 4–13 of 14 extractions shown in Table 4), respectively (Fig. 7), giving a combined age of 201 ± 2 Ma (2σ). Except for some younger apparent ages in the low-temperature extractions, both plateaux show concordant step ages that overlap at the 95% confidence level. Ages calculated from the total gas release, including apparent younger ages obtained during the low-temperature heating steps from sample sites of lower argon retentivity, are just slightly (∼2 Ma) younger than the plateau ages. A summary of isotopic data for the K-feldspars is given in Tables 3 and 4. Inverse isochron plots show a 36Ar/40Ar intercept at approximately 0.0034 for both K-feldspar types (Fig. 8). This ratio is indistinguishable from atmospheric argon composition and suggests that the samples are not disturbed by inherited 40Ar or secondary 40Ar loss (e.g., McDougall and Harrison 1999; Jourdan et al. 2008). Given the largely Ca-free nature of the essentially pure K-feldspar phases, the K/Ca (39Ar/37Ar) ratio is variable (partially not defined, as Ca-derived 37Ar was below detection) over the age plateaux. A relatively high amount of 40Ar degassed during the highest-temperature extractions (1350–1450 °C) and a comparatively low K content in handpicked aggregates suggest that spherulitic-blocky K-feldspar was intergrown with some microcrystalline quartz that contained additional atmospheric argon released at high extraction temperatures. Our new combined K-feldspar age is in contradiction with some of the earlier dating results, yet within the general range of previously determined isotopic and paleomagnetic ages (Fig. 2) (compare summary by Kelley and Spray 1997) and suggests a Rhaetian impact age (see international stratigraphic chart by Ogg et al. 2008). The new age is, within error, indistinguishable from the isotopic age for the Triassic/Jurassic boundary and the slightly earlier phase of Central Atlantic Magmatic Province (CAMP) volcanism (e.g., Marzoli et al. 2008), recently dated at 201.7 ± 0.6 Ma (Pálfy et al. 2008) and 201.58 ± 0.17 Ma (Schaltegger et al. 2008) using the 206Pb/238U technique (keeping in mind the miscalibration between U- and K-decay series), as well as 201.7 ± 2.4 to 197.8 ± 0.7 (Vérati et al. 2007) and 201.0 ± 1.4 to 198.1 ± 1.1 Ma (Jourdan et al. 2009a) by the 40Ar/39Ar method, respectively. Furthermore, the new age is also coeval with one previously reported K–Ar age (yet among other older and younger published ages for this still poorly dated impact structure) obtained for the approximately 80 km Puchezh-Katunki impact structure, Russia (200 ± 3 Ma; Masaitis and Pevzner 1999; see also Pálfy 2004; Schmieder and Buchner 2008; Jourdan et al. 2009b for discussion of the Puchezh-Katunki impact age and Thompson et al. 2010 for recent crater size estimates).

Figure 7.

40Ar/39Ar age spectra for primary K-feldspar (top) and secondary K-feldspar (bottom) separated from the impact-metamorphosed Videix gneiss; uncertainties are at the 1σ level (parentheses include errors on the standard; compare isotopic data given in Tables 3 and 4).

Figure 8.

 Inverse isochron plots for primary K-feldspar (circles; top) and secondary K-feldspar (triangles; bottom). Gray symbols represent argon extraction steps outside the plateau fraction not included in isochron calculation; small numbers denote individual extraction steps; cross symbols mark uncertainties at the 1σ level (parentheses include errors on the standard; compare isotopic data given in Tables 3 and 4).


Petrology of the Videix Gneiss

The paragenesis of the impact-metamorphosed Videix gneiss (primary spherulitic-blocky K-feldspar–α-quartz–hematite–secondary idiomorphic K-feldspar [adularia] as the main phases; Figs. 3A and 3B) suggests that spherulitic-blocky K-feldspar (Figs. 3C and 3D) formed as an early recrystallization product of normal K-feldspar glass, most likely as sanidine at high post-shock temperatures of approximately 850–1000 °C (Bischoff and Stöffler 1984). The partially spherulitic character of primary K-feldspar might be related to early post-shock devitrification processes at fast cooling rates (e.g., Lofgren 1971). Microcrystalline quartz aggregates that have commonly retained the original angular shape and display relict shock features (Figs. 3E and 3F) probably represent domains of strongly recrystallized diaplectic quartz glass or maybe quartz clasts partially molten but not assimilated by the surrounding (feldspar) melt, suggesting peak shock pressures of >35 GPa and post-shock temperatures of up to approximately 1500 °C (Stöffler and Langenhorst 1994; French 1998; Schmieder et al. 2008; Lambert 2010). The presence of ballen α-quartz (Fig. 3F) (Ferrière et al. 2009a) in relict toasted silica domains is compatible with temperatures exceeding 1000 °C and indicates the transition of protolith quartz to β-cristobalite during post-shock heating and back to quartz (via α-cristobalite) upon cooling; this is also in agreement with the incorporation of resolvable amounts of Al2O3 (≤1 wt%) into silica at high temperatures (e.g., Osinski 2004). Due to pervasive annealing and recrystallization, no primary shock features (such as optically sharp planar deformation features) are preserved in the microcrystalline silica domains. The vesicular microfabric of recrystallized quartz clasts and the distinct toasting effect, furthermore, underline the influence of a strong post-shock thermal overprint (Ferrière et al. 2009b; cf. Whitehead et al. 2002). Very similar thermal effects in silica have also been observed in melt rocks (and, particularly, in lithic clasts within the melt rocks) from the Rochechouart impact structure (e.g., at Montoume or Babaudus; Lambert 2010). The crystallization of hematite may be related to the breakdown of precursor protolith biotite (e.g., Schneider 1974). The same idiomorphic K-secondary feldspars as in the Videix gneiss, are widely represented in Rochechouart impactites such as in vesicles of impact melt rocks (e.g., near Babaudus, Valette, and La Chauffie in the central part of the Rochechouart impact structure), and in the Chassenon suevite. Similar observations have been reported from some other impact structures on Earth, e.g., Chicxulub (Mexico; Tuchscherer et al. 2006) and Kärdla (Estonia; Versh et al. 2005). Hydrothermal systems associated with larger terrestrial impact structures, responsible for penetrative alteration of impact breccias and melt rocks, are known from a long list of impact structures (see summary by Naumov 2005). Potassium metasomatism in impactites, indicated by a notable enrichment in K with respect to the unshocked target rocks, is also known from a number of terrestrial impact structures, such as Ilyinets (Ukraine; Gurov et al. 1998), the Ries crater (Germany; Osinski 2005), or Siljan (Sweden; Reimold et al. 2005). A review of the selective melting of K-feldspar and superimposed K-metasomatism (together with the relative depletion of Na and Ca) in the Rochechouart impactites is given by Lambert (1981, 2010) and Reimold et al. (1984). In particular, specific potassium enrichment in the Videix gneiss, together with high K2O/Na2O (∼100 within the K-feldspars) and K2O/CaO (∼1000 within the K-feldspars) ratios compared to the unshocked target rock gneisses (K2O/Na2O ∼1.2; K2O/CaO ∼1.8), appears to be in analogy to shocked and selectively K-enriched lithic target rock clasts in the Rochechouart impact melt rocks (Lambert 2010; his Table 2); the same phenomenon was also observed at other terrestrial impact structures (e.g., strongly K-enriched gneiss clasts in impact melt rocks from the Brent crater, Canada; Grieve 1978). Cavity-filling adularia in the Videix gneiss most likely crystallized at moderate to low temperatures in a typical range of approximately 350–100 °C (e.g., Naumov 2005; Versh et al. 2005). Thus, at least at the hand specimen scale, the association of spherulitic-blocky primary K-feldspar (most likely originally “impure” sanidine recrystallized from monomineralic K-feldspar melt and possibly depleted in Na and Ca during the cooling process; compare Buchner et al. 2010) intergrown with microcrystalline and locally ballen-textured quartz and idiomorphic secondary K-feldspar (adularia) reflects the local cooling history of this “impact metasomatite” from high post-shock temperatures of up to approximately 1500 °C down to approximately 200 °C or even lower. This suggests that the Videix gneiss was subjected to the pervasive K-metasomatism and hydrothermal overprint caused by (and typical for) the Rochechouart impact, possibly in close contact to the (overlying) hot impact melt sheet.

Timing of K-Feldspar Crystallization and K-Metasomatic–Hydrothermal Overprint

Apart from the 15 km diameter impact deposit area, the youngest crystallization age known for basement rocks in this part of the Massif Central is about 300 Ma. No endogenic magmatic-thermal events are known to have occurred later in the region (e.g., Alexandrov et al. 2000; Lambert 2010 and references therein). Some post-Variscan Pb-Zn, U, F-Ba, F, and Pb-Zn-Ba metallogenic phases have been retraced at various places in the French Massif Central (mainly in its southern part; Marignac and Cuney 1999); however, these are ascribed to the onset of Tethys opening and Early Mesozoic distension leading to the formation of low-enthalpy geothermal systems (circulation of fluids at rather low temperatures of ≤200 °C) but also to the continental leaching of Variscan feldspars (Marignac and Cuney 1999), which therefore does not explain the high-temperature (≥1000 °C) nature of the Videix gneiss (formation of ballen silica as an internal impact-metamorphic geothermometer) and subsequent K-metasomatic-hydrothermal overprint. Younger Cenozoic volcanism in the Massif Central, responsible for the subrecent evolution of the volcanic province to what it appears today (with the Cantal, Mont Dore, Aubrac, Devès, Velay, and Chaîne des Puys volcanic fields), occurred between ∼65 Ma and ∼3450 yr before present (Downes 1987; Wilson and Downes 1991). Thus, we can exclude a major endogenic thermal event at approximately 200 Ma that could have disturbed the isotopic system in the Rochechouart impactites. For the Videix gneiss, this provides evidence that thermal metamorphism was caused by the Rochechouart impact event.

Impact-induced hydrothermal systems that may cause penetrative alteration and recrystallization of crater-filling and subcrater rocks have been recognized at a number of terrestrial impact structures (e.g., Abramov and Kring 2004, 2007; Naumov 2005). As inferred from the similarly sized ∼24 km Nördlinger Ries (Germany) and ∼24 km Haughton (Canada) impact structures, assuming a similar diameter for Rochechouart (23 km; Kelley and Spray 1997), the cooling of the hot proximal impact ejecta from ∼600 °C down to ∼100 °C probably took several thousands of years, with slower rates of further cooling down to ambient conditions (Pohl et al. 1977; Osinski et al. 2001). In addition, the structural uplift of crustal rocks at complex impact structures (especially within the central uplifts of impact structures >20 km in diameter) causes elevated transient, local geotherms with thermal gradients estimated at approximately 100 °C km−1 (e.g., Masaitis and Naumov 1993; Naumov 2005 and references therein). Assuming a larger original diameter for the Rochechouart impact structure (∼40–50 km; Lambert 2010; see Fig. 1), hydrothermal-metasomatic activity might have taken some tens of thousands of years. In either case, this time interval—i.e., the temporal difference between the moment of impact and post-impact K-metasomatic alteration of impactites and the hydrothermal crystallization of K-feldspar in the Videix gneiss—is too short to be reliably resolved by the 40Ar/39Ar dating technique within uncertainties (in contrast to the mineralization at the considerably larger approximately 250 km Sudbury impact structure, Ontario, Canada, where hydrothermal activity lasted up to approximately 2–4 Myr; Ames et al. 1998; Naumov 2005). It is, hence, suggested that the isotopic age of K-feldspars in the Videix gneiss reflects the recrystallization of K-feldspar from local (monomineralic) feldspar melts and the superimposed early post-impact K-metasomatic-hydrothermal recrystallization, in turn dating the Rochechouart impact age within error.

40Ar/39Ar Dating Results

Since the establishment of the 40Ar/39Ar dating method (see original study by Merrihue and Turner 1966), this technique has been applied to impact melt rocks and glasses from a number of impact structures on Earth (e.g., Bottomley et al. 1990; Müller et al. 1990; Swisher et al. 1992; Deutsch and Schärer 1994; Trieloff et al. 1994, 1998; Schwarz and Lippolt 2002; Buchner et al. 2003; Reimold et al. 2005; Jourdan et al. 2008; Schmieder et al. 2009). K-feldspar is considered to be preferentially suitable for 40Ar/39Ar dating and also used as standard in 40Ar/39Ar geochronology (e.g., the Fish Canyon sanidine; McIntosh et al. 1990; McDougall and Harrison 1999; Jourdan and Renne 2007). In impact geochronological studies, K-feldspar (sanidine) has been used as dating material for the approximately 143 Ma Gosses Bluff impact structure (Australia; Milton and Sutter 1987), the approximately 74 Ma Manson impact structure (Iowa, USA; Izett et al. 1998), and the 14.6 ± 0.2 Ma Nördlinger Ries impact structure (Germany; Buchner et al. 2010), all associated with fairly robust ages (see also Jourdan et al. 2009b for a recent appraisal of impact ages). The use of adularia as geochronological proxy is largely restricted to thermochronologic studies of hydrothermal ore deposits; however, similar to sanidine, adularia has yielded high-precision 40Ar/39Ar ages in many cases (e.g., Mertz et al. 1991; Lang et al. 1994; Ward et al. 2005; Arancibia et al. 2006).

The characteristics of the 40Ar/39Ar age plateaux for K-feldspars in the Videix gneiss (Fig. 7), the consistency of ages between the high- and low-temperature phases of K-feldspar (i.e., primary and secondary K-feldspar), as well as robust data set statistics (e.g., a representative number of heating steps with concordant step ages and a high probability of 99% for the adularia age; Fig. 7; isotopic Tables 3 and 4), provide a valuable internal geochronological age countercheck. The virtually fresh state of the impact-metamorphosed gneiss (i.e., there is no indication of intense rock alteration in thin section), comparatively large and optically homogenous domains and crystals of K-feldspar, and the lack of visible fluid inclusions suggest that the internal K–Ar isotopic system is not disturbed by argon loss, recoil effects, or excess argon (e.g., McDougall and Harrison 1999; Kelley 2002; Jourdan et al. 2007a); this is underlined by a good fit of data in inverse isochron plots (Fig. 8). Particular 39Ar recoil features—occasionally disturbing age spectra of fine-grained impact melt rocks (e.g., Bottomley et al. 1990; Müller et al. 1990; Trieloff et al. 1994, 1998)—seem to be absent, similar to age spectra of tektites (Schwarz and Lippolt 2002).

Previous dating of impact melt rocks and glasses from the Rochechouart impact structure was hampered by either poor geochronometers (e.g., the Rb–Sr or K–Ar method associated with large dating errors) or poorly suitable dating material. For example, the fine-grained aphanitic impact melt rocks from the Babaudus area are notably rich in alteration phases, predominantly clay minerals and iron hydroxides. These rock properties may be responsible for the loss of radiogenic argon upon alteration and/or argon recoil redistribution within the sample during irradiation (e.g., McDougall and Harrison 1999; Jourdan et al. 2007b), causing inaccurate and imprecise (younger) apparent ages (Fig. 2) (e.g., results by Kraut and French 1971; Lambert 1974). 40Ar/39Ar dating of the Champagnac pseudotachylites by Kelley and Spray (1997) was carried out using laser spot dating on a thin rock tile and age determination by inverse isochron calculations. Ages of single spots yielded comparatively old ages between 231 Ma and approximately 1.67 Ga, indicative of a distinct excess argon component (Kelley and Spray 1997; McDougall and Harrison 1999; Kelley 2002) in parts of the sample material that was probably not completely degassed upon friction melting. Furthermore, the MSWD value calculated in the inverse isochron plot given by Kelley and Spray (1997) was rather high (up to 4.0), defining a less robust 214 ± 8 Ma “preferred” age for the Rochechouart impact. The new Rhaetian age breaks the link between the Rochechouart impact structure and the postulated approximately 214 Ma Late Triassic “impact crater chain” (Rochechouart–Manicouagan–Lake Saint Martin–Red Wing Creek–Obolon; Spray et al. 1998; Walkden et al. 2002). Similarly, the impact structures of Lake Saint Martin (Canada) and Obolon (Ukraine) have recently been shown to be probably older and younger than Late Triassic (although still poorly defined by isotopic and stratigraphic means), respectively (Valter et al. 2000; Schmieder and Buchner 2008; Wartho et al. 2009; Jourdan et al. 2010). The concordant plateau ages for primary spherulitic-blocky and secondary idiomorphic K-feldspar also demonstrate that, in addition to impact melt rocks and glasses, K-metasomatic/hydrothermal phases crystallized during the post-shock cooling process are suitable for the isotopic dating of impact events (see also Ames et al. 1998).

Seismites and Tsunami Deposits Associated with the Rochechouart Impact?

Although no structural, lithological, or sedimentological criteria for a (shallow) marine impact scenario have been found to date, paleoenvironmental considerations suggest that the Rochechouart impact occurred very close to (or even beyond) the latest Triassic shoreline of the continental Massif Central, with the marine Aquitaine Basin and Biscay Rift (westernmost Tethys) to the West (e.g., Curnelle et al. 1982; Ziegler 1990; Smith et al. 1994). This is in accord with earlier suggestions that the Rochechouart impact occurred within a shallow marine continental basin and the detection of elevated amounts of sulfur, phosphorous, and chlorine in the glassy mesostasis of the Champagnac pseudotachylites (Kelley and Spray 1997), which might indicate the influence of seawater.

Recently, Simms (2003, 2007) described a still puzzling approximately 2–4 m thick “seismite” of large extent (>250,000 km2) partially overlain by “tsunamite” in the uppermost Triassic of the British Isles (Cotham Member of the Lilstock Formation, Penarth Group, upper Rhaetian) a few meters below the biostratigraphic Triassic/Jurassic boundary. These peculiar deposits indicative of extensive soft-rock deformation were suggested to be incompatible with endogenic terrestrial mechanisms but consistent with a hitherto unknown high-energy end-Triassic impact event: “The seismite certainly suggests an event that is unique in the Phanerozoic history of Britain and Ireland, and that is consistent with bolide impact,” and “impact of a km-scale asteroid here potentially could produce the observed sedimentological effects across the UK” (Simms 2007; see also Hesselbo et al. 2007 for discussion). Similarly, Mader (1992, p. 306) noted potential tsunami deposits in the Lodève Basin of Southern France (Pégairolles de l’Escalette section, Languedoc-Roussillon; see Courtinat et al. 2003, for stratigraphic column), where an intercalation of “exceptional and quasi-instantaneous character” in the Upper Member of the Rhaetian contains “allochthonous material” and “could be attributed to a tsunami having been triggered by seismic instability in the basin” (Mader 1992). In the Northern Apennines of Italy, Rhaetoliassic sediments of the Grotta Arpaia section (basal Grotta Arpaia beds, Portovenere Limestone Member of the La Spezia Formation) interpreted as basin deposits distal to a carbonate ramp, contain slump deposits with storm layers, the composition of which “implies that they were not due to tectonic activity at the platform margin, but that they originated inside the basin and affected the middle and outer part of the ramp” (Ciarapica 2007). Upper Rhaetian exotic high-energy deposits were also reported from the High Tatra Mountains of the Slovak Western Carpathians (“slumped beds” of the upper Fatra Formation; Michalík et al. 2007).

In agreement with the new 40Ar/39Ar age data and paleogeographic studies, the near-coastal to shallow marine impact of the Rochechouart meteorite (see Fig. 9 for impact site at ∼200 Ma) can be regarded as a potential trigger mechanism for a large tsunamigenic earthquake and the formation of end-Triassic tsunami deposits in the westernmost Tethys domain. The Aquitaine Basin, Biscay Rift, Western Approaches Trough, Bristol Channel, and Burgundy Gate (Fig. 9) (compare paleogeographic maps by Ziegler 1988; Ziegler and Kent 1982; Ziegler et al. 1983; Blakey 2006) represented channel-like sea passages that linked the Rochechouart impact site and the British Isles at the time of impact, maintaining high wave energy. The western Tethyan rift basins represented domains of pronounced subsidence during the Late Triassic (Ziegler and Kent 1982), suggesting a steep relief of the basins and deeper marine conditions. The distance between the impact site and the area of tsunami deposition was about 700–1300 km for the British Isles and notably shorter, about 300 km, for southern France (Pégairolles de l’Escalette). Just as well, the Northern Apennine and Western Carpathian domains were at paleodistances of ≤2000 km in the direction of the Neotethys Sea. Tsunamis, large sea waves caused by tectonic activity (earthquakes and faulting), submarine landslides, and volcanic eruptions, are known to produce distinct sedimentary features of high-energetic and complex erosion and deposition in the coastal run-up areas, commonly characterized by erosive fining-upward sequences (e.g., Bondevik et al. 1997; Dawson and Shi 2000; van den Bergh et al. 2003; Cantalamessa and Di Celma 2005); however, the destructive properties and specific sedimentological features produced by tsunamis strongly depend on the location of the tsunami source, the regional geologic and topographic conditions, as well as on water depth and sea floor geometry (e.g., Dypvik and Jansa 2003; Satake 2007), and are still not fully understood (e.g., Bondevik et al. 1997; Gersonde et al. 2002). During the last decades, tsunami deposits formed by impact-triggered tsunamis have been recognized, e.g., those caused by the Chicxulub impact at the Cretaceous/Paleogene boundary (e.g., tsunamites in Texas and Mexico; Bourgeois et al. 1988; Smit et al. 1996), or in the Archean Hamersley Basin of Western Australia (Hassler et al. 2000). Just as well, tsunamites are associated with the Devonian Alamo impact, Nevada and Utah, USA (e.g., Warme and Kuehner 1998), the Devonian Kaluga impact, Russia (Dypvik et al. 2004), and the Cretaceous Tookoonooka impact, central Australia (Bron 2009). Numerical modeling suggested that the Latest Jurassic marine impact that created the approximately 40 km Mjølnir impact structure in the Barents Sea off Norway also triggered a large tsunami (Shuvalov et al. 2002; Glimsdal et al. 2007). A summary of impact-generated tsunamis is given by Ward and Asphaug (1999), Dypvik and Jansa (2003), Dypvik et al. (2004), and Goto (2008). The Rochechouart impact energy exceeded the energy of the largest man-witnessed terrestrial earthquake (Valdivia, Chile, May 22, 1960; Richter scale magnitude 9.5 and a released energy of ∼1.12 × 1019 J), capable of producing major seismic waves in the Earth’s crust and tsunami waves in the sea. An impact energy of approximately 1.3 × 1021 J (equivalent to an earthquake of Richter scale magnitude 10.8–10.9) would have been associated with the impact of an iron meteorite approximately 1 km in diameter into largely crystalline rocks, at a typical impact velocity of 25 km s−1 and an impact angle of 45° (Marcus et al. 2004; Collins et al. 2005); these values are conservative estimates for the current (minimum) 20–25 km diameter of the eroded Rochechouart impact structure. Assuming a larger, approximately 2 km iron projectile and an original diameter of the Rochechouart impact structure of approximately 40–50 km (Lambert 2010), given the same impact angle and velocity, an impact energy of approximately 1.05 × 1022 J (corresponding to an earthquake of Richter scale magnitude 11.4–11.5) would have been released. In both cases, an impact-induced tsunamigenic earthquake could have caused the collapse of the local continental margin of the westernmost Massif Central and submarine landslides in the Aquitaine Basin and Biscay Rift as possible supporting mechanisms, maybe similar to the Eocene marine Montagnais impact scenario on the Scotian shelf of Eastern Canada (Jansa 1993). Comparable to tsunamis caused by endogenic earthquakes, an impact tsunami wave is thought to contain kinetic and gravitational energy in the range of about 7–9% of the total impact energy (Ahrens and O’Keefe 1983; Dypvik and Jansa 2003). According to Jansa (1993), a marine impact producing an approximately 50 km crater would result in a 200 m high wave within a radius of 300 km and still a 40 m wave within a 3000 km radius. The height of the tsunami wave is greatly amplified when reaching the shallow sea and coastal areas, especially when the coastal geometry (for example, V-shaped bays, channels, or fjords, etc.) supports effects of tsunami wave focusing and resonance (Satake 2007), as might have been the case on the Rhaetian British Isles.

Figure 9.

 Paleogeographic map of the western Tethys domain in the latest Triassic. Rochechouart impact site is indicated by black star (after Ziegler et al. 1983; Ziegler 1988; Blakey 2006).

Possible Distal Rochechouart Impact Ejecta?

In contrast to the ∼215 Ma and ∼100 km in diameter Manicouagan impact structure (Hodych and Dunning 1992; Ramezani et al. 2005) associated with seismites in the Upper Triassic of the Fundy Basin in eastern Canada and a distal ejecta layer in Britain (Tanner 2002; Walkden et al. 2002; Kirkham 2003; Thackrey et al. 2008), no distal impact ejecta have so far been recognized to originate from the smaller Rochechouart impact. According to Walkden et al. (2002), a possible Rochechouart ejecta layer on the British Isles would have a primary thickness of about 1–1.6 mm; one could expect that ejecta might have been deposited on top of the seismites and prior to the formation of the tsunamites described by Simms (2003, 2007), which would warrant further investigations. In the more distal areas, Lintnerová et al. (2004) and Michalík et al. (2007) reported layers of enigmatic spherules in the Upper Rhaetian Fatra Formation of the High Tatra Mountains of Slovakia, some meters below the biostratigraphic Triassic/Jurassic boundary. A nonimpact (volcanic or sedimentary) origin is discussed for the spherules, however, Michalík et al. (2007) resume that the spherules “bear resemblance to small vitreous particles found in Cretaceous–Paleogene beds of the Caribbean region” and might have been formed by “calcitization of glass droplets originating from a bolide impact, such as that postulated to occur immediately before the Rhaetian–Hettangian boundary.” Moreover, allegedly shocked quartz grains were reported in the Upper Rhaetian of the Northern Apennines, Italy (Bice et al. 1992) and the Northern Calcareous Alps, Austria (Badjukov et al. 1987) but unequivocal signs of shock could not be confirmed (Hallam 1990, 2002; Mossman et al. 1998; Simms 2007). It also remains speculative whether a secondary peak within the complex iridium anomaly closely below the actual Triassic/Jurassic boundary in the Blomidon Formation (eastern Canada; Tanner and Kyte 2005) could be related to the impact of the Rochechouart iron meteorite (Tagle et al. 2009). The comparatively small crater size, however, suggests no direct relationship between the Rochechouart impact and the global end-Triassic mass extinction (e.g., Hallam 1990, 2002; Hallam and Wignall 1997; Pálfy et al. 2001; Ward et al. 2001; Hesselbo et al. 2002; Tanner et al. 2004). Possible distal impact ejecta produced by the Rochechouart impact in the European sedimentary record will require further study (see also Lambert 2010).


40Ar/39Ar dating of “primary” spherulitic-blocky K-feldspar (201.4 ± 2.4 Ma; 2σ) and “secondary” idiomorphic K-feldspar of the adularia morphological type (200.5 ± 2.2 Ma; 2σ) separated from impact-metamorphosed gneiss found near Videix in the western central part of the Rochechouart impact structure combined yielded a Rhaetian age of 201 ± 2 Ma (2σ), coeval within uncertainty with the Triassic/Jurassic boundary. Primary (spherulitic-blocky; most likely sanidine crystallized from monomineralic K-feldspar melt) and secondary K-feldspars in the gneiss reflect the early post-shock crystallization of K-feldspar intergrown with locally ballen-textured silica at high temperatures of ≥1000 °C and subsequent cooling and concomitant K-metasomatic and hydrothermal overprint. Considering a larger original crater diameter of approximately 40–50 km for Rochechouart, K-metasomatic-hydrothermal activity at Rochechouart probably lasted not longer than some tens of thousands of years, which is not resolved by the 40Ar/39Ar age within uncertainties. The new K-feldspar age is substantiated by robust internal statistics (concordant well-defined plateau ages for both types of K-feldspar; inverse isochron plots that indicate that there is no notable excess argon component). Furthermore, the new age is by some million years younger than the 214 ± 8 Ma (Norian) age obtained by Kelley and Spray (1997; 40Ar/39Ar laser dating of pseudotachylites) and, therefore, contradicts the Late Triassic multiple impact theory postulated by Spray et al. (1998). In turn, a Rhaetian impact age is compatible with a near-coastal (or shallow marine) impact scenario making the Rochechouart impact event a potential trigger mechanism for the formation of seismites and tsunami deposits in the upper Rhaetian of the British Isles recently described by Simms (2003, 2007). Furthermore, upper Rhaetian tsunami and storm deposits in southern France and northern Italy, as well as assumedly shocked quartz in upper Rhaetian sediments of the Northern Apennines (Italy) and an enigmatic spherule layer in the Tatra Mountains of northern Slovakia, might be related to the Rochechouart event. The confirmation of an impact origin of these deposits is still outstanding.

Acknowledgments— We are grateful to Elmar Jessberger (Westfälische Wilhelms-Universität Münster, Germany), Simon Kelley (Open University, Milton Keynes, United Kingdom), Timothy Swindle (University of Arizona, Tucson, Arizona, USA), Kalle Kirsimäe (University of Tartu, Estonia), and Gordon Osinski (University of Western Ontario, London, Ontario, Canada) for their very helpful reviews and suggestions. We, furthermore, acknowledge Odile Dupuy and Claude Marchat (Association Pièrre de Lune), as well as the “Espace Météorite Paul Pellas” museum, Rochechouart, for their very kind support in the field. We also want to thank Mike Simms (Ulster Museum, Cultra, Northern Ireland), Fred Jourdan (Western Australian Argon Isotope Facility-JdL, Perth, Australia), and Jozef Michalík (Slovak Academy of Science, Bratislava, Slovakia) for additional comments, as well as Jens Hopp (Universität Heidelberg), Thomas Theye, and Christoph Wimmer-Pfeil (Universität Stuttgart) for their technical support. We thank Klaus Tschira Stiftung gGmbH for support to perform isotopic dating reported in this study. M. S. and E. B. acknowledge the Landesgraduiertenförderung Baden-Württemberg and the Stifterverband für die Deutsche Wissenschaft for their grants, respectively.

On the behalf of the authors, this study is dedicated to the outstanding natural and cultural heritage preserved in the Astroblème de Rochechouart-Chassenon. Furthermore, we want to support the efforts by involved geoscientists and officials to promote future scientific studies and public outreach focused on this unique site.

Editorial Handling— Dr. Gordon Osinski