1 Age of the Sääksjärvi impact structure, Finland: reconciling the
2 timing of small impacts in crystalline basement with regional
3 basin development
4
5 Gavin G. Kenny1*, Irmeli Mänttäri2, Martin Schmieder3,4, Martin J. Whitehouse1,
6 Alexander A. Nemchin1,5, Jeremy J. Bellucci1, Renaud E. Merle1
7
8 1Department of Geosciences, Swedish Museum of Natural History, SE-104 05 Stockholm,
9 Sweden
10 2Radiation and Nuclear Safety Authority, Laippatie 4, Box 14, FI-00811 Helsinki, Finland
11 3Lunar and Planetary Institute – USRA, Houston TX 77058, USA
12 4HNU – Neu-Ulm University of Applied Sciences, D-89231 Neu-Ulm, Germany
13 5School of Earth and Planetary Sciences (EPS), Curtin University, Perth, WA 6845, Australia
14
16
17 ABSTRACT
18 We report a new age for the Sääksjärvi impact structure, Finland, a 6 km in diameter
19 feature that formed in crystalline rocks of the Precambrian Baltic Shield. Two previous studies
20 have reported 40Ar/39Ar data for Sääksjärvi and suggested conflicting formation ages of ≤330 Ma
21 or approximately 560 Ma, respectively. The former represents a possible complication for
22 models which indicate that the region was covered by sediments of the Caledonian foreland
23 basin throughout much of the Phanerozoic. We conducted a study combining imaging,
24 microstructural analysis and U–Pb dating of shocked zircon from Sääksjärvi. The U–Pb dataset
25 indicates a ca. 600 Ma impact into predominantly ca. 1850 Ma target rocks. A concordia age of
26 608 ± 8 Ma (2σ) confirms Sääksjärvi as the first known Ediacaran impact structure in the Baltic
27 Shield and only the second worldwide. Our data indicate that the Sääksjärvi impact structure
28 formed in exposed crystalline basement rocks of the Baltic Shield prior to the development of the
29 Caledonian foreland basin. Given that most impact structures on Earth are relatively small
30 features, radiometric dating of small impact structures in crystalline basement may place
31 boundaries on the timing and spatial extent of palaeobasins that might be otherwise difficult to
32 constrain.
33
34 1. INTRODUCTION
35 Small impact structures in crystalline basement have a unique ability to shed light on the
36 temporal and spatial extent of palaeobasins. Impact structures sometimes provide, in their impact
37 lithologies or down-faulted segments, a useful record of pre-impact bedrock units that have
38 otherwise been completely removed from around the structure; for example, the Lappajärvi
39 impact structure, Finland, preserves in its structural depression sedimentary units that have
40 facilitated reconstructions of Cambrian palaeogeography in the Baltic region (e.g., Slater and
41 Willman, 2019, and references therein). Similarly, the Siljan impact structure in Sweden, and the
42 Manicouagan impact structure in Canada, also preserve megablocks of early Phanerozoic
43 sedimentary units that are not well preserved in the surrounding regions (e.g., Grieve, 2006;
44 Grieve and Head, 1983; Lehnert et al., 2012, 2014, and references therein). At Paasselkä, and a
45 number of other impact structures in Finland and Sweden, impact melt breccias contain shocked
46 and melted fragments of Mesoproterozoic ‘Jotnian Sandstone’ that has been completely removed
47 by erosion elsewhere in the area (Buchner et al., 2009; Puura and Plado, 2005). Conversely, the
48 presence of small impact structures in crystalline basement can provide evidence that the
49 basement was not deeply covered by sediments at a given time if an accurate and precise age for
50 the structure can be attained. Such structures may therefore offer an opportunity to test, and even
51 place boundaries on, the size and lifetime of sedimentary basins which might otherwise be
52 difficult to constrain. To explore this possibility, we here revisit the age of a small impact
53 structure in crystalline basement (the Sääksjärvi impact structure, Finland) for which conflicting
54 40Ar/39Ar ages have been published (Bottomley et al., 1990; Müller et al., 1990), one of which
55 overlaps a proposed period of significant sedimentary cover in the region.
56
57
58 1.1. The Sääksjärvi impact structure, Finland
59 The Sääksjärvi impact structure is a deeply eroded, approximately 6 km in diameter feature
60 submerged beneath an eponymous lake in southwest Finland (Fig. 1; e.g., Willman et al., 2010).
61 Papunen (1969) suggested an impact origin for Sääksjärvi based on the observation of multiple
62 sets of closely spaced planar structures in quartz within heavily brecciated and vesiculated rocks
63 from the eastern shore of Lake Sääksjärvi and a glacial esker approximately 20 km to the
64 southeast. To the best of our knowledge, there are no known outcrops of in situ impactites at
65 Sääksjärvi (e.g., Papunen 1969, 1973; Pihlaja and Kujala, 2000). The structure has been targeted
66 by six drill holes (totalling 860 m in length), including one into the centre of the gravity low
67 which penetrated 180 m of suevite and other breccias (but not a coherent melt sheet) before
68 terminating in deformed mica gneiss basement at a depth below surface of 242 m (Pihlaja and
69 Kujala, 2000). Platinum group element patterns and Ni/Cr, Ni/Ir, and Cr/Ir ratios of impact melt
70 lithologies suggest the Sääksjärvi impactor was likely a non-magmatic iron meteorite (Schmidt et
71 al., 1997, Tagle et al. 2009).
72 Two studies have reported 40Ar/39Ar data for impact melt rocks and impact melt-bearing
73 breccias from Sääksjärvi. Bottomley et al. (1990) reported data from three samples described as
74 fine-grained melt rocks with quartz and feldspar clasts that were relatively unrecrystallised. The
75 samples displayed two distinct trends in their step-heating spectra: two samples gave
76 progressively older apparent ages from Ar release steps as heating continued (and 39Ar was
77 incrementally released), with ages rising from ca. 330 Ma and forming a plateau between 510
78 and 530 Ma, whereas aliquots from another sample gave a humped spectrum, with steps rising
79 from ca. 400 Ma to a peak at ca. 650 Ma and then forming a plateau at ca. 600 Ma. The dataset
80
81 82 Figure 1. Map of impact structures in the Baltic region. Best estimate ages, structure diameters and target rock
83 compositions from Schmieder and Kring (2020). When the age of an impact structure is given as a range in
84 Schmieder and Kring (2020) we use the midrange value – for example, for the Iso-Naakkima impact structure,
85 which is considered to have formed 1200–900 Ma, we use a value of 1050 Ma. Sääksjärvi is plotted using the new
86 608 ± 8 Ma age reported here. See Table 1 in Supplementary Appendix for full data used to create map.
87 Approximate extent of the Baltic Shield (including Caledonian Orogenic Belt) and Phanerozoic sediments after
88 Murrell and Andriessen (2004). The Caledonian foreland basin lay to the east of the Caledonian Orogenic Belt, over
89 present day Sweden and Finland, but its precise extent is uncertain (see text).
90
91 lacked any statistically relevant plateaus that may have represented an impact age (see discussion
92 of 40Ar/39Ar dating of impact structures by Jourdan [2012]) and Bottomley et al. (1990)
93 interpreted their data to indicate that a ≤330 Ma impact had incompletely degassed target rocks
94 with K-Ar ages of between 500 and 600 Ma. Müller et al. (1990) reported 40Ar/39Ar data for one
95 sample, which was described as a cryptocrystalline melt breccia with >15 % clasts
96 (predominantly quartz and feldspar) and a matrix dominated by plagioclase and pyroxene. The
97 sample gave a humped age spectrum with apparent age domains of ca. 550 and ca. 600 Ma. With
98 no reason to prefer one age over the other, Müller et al. (1990) suggested that the ages should not
99 be taken “too seriously”. Nevertheless, Müller et al. (1990) considered a weighted average age of
100 560 ± 16 Ma to be their best estimate for the age of the impact event. Similar to the dataset of
101 Bottomley et al. (1990), the data presented by Müller et al. (1990) lack statistically robust
102 plateaus and so none of the currently available 40Ar/39Ar data from impact melt at Sääksjärvi are
103 likely to record an accurate impact age.
104 1.2. Geological setting of the Sääksjärvi impact structure
105 The Sääksjärvi impact structure formed in crystalline basement of the Baltic Shield, in 1.9–
106 1.8 Ga rocks related to the Svecofennian orogeny (Fig. 2). During the Phanerozoic these
107 basement rocks were largely covered by sediments as a major foreland basin developed to the
108 east of the Scandinavian Caledonides (Fig. 1) (e.g., Cederbom et al., 2000; Garfunkel and
109 Greiling, 1998; Larson et al., 1999; Middleton et al., 1996; Murrell and Andriessen, 2004).
110 However, the Caledonian foreland basin is not well preserved as rock exposures and estimates
111 for its extent and, particularly, life span vary significantly.
112 Fission-track thermochronology and inverse modelling suggests that most of Finland was
113 buried by approximately 1 km of Upper Paleozoic sediments that thickened to approximately 2.5
114 km towards the Caledonian border zone of western and southern Sweden (Cederbom et al., 2000;
115 Larson et al., 1999). Cederbom et al. (2000) presented models indicating that the basin existed
116 from the Devonian through to the early Mesozoic or even Cenozoic. Murrell and Andriessen
117 (2004) interpreted apatite fission-track data for southern Finland to indicate that late Silurian
118
119 Figure 2. Simplified geological map of the Sääksjärvi impact structure, Finland, after the 1:200 000 bedrock map of
120 the Geological Survey of Finland (GTK, 2017). White shaded area represents a lake (Sääksjärvi) and white lines
121 denote rivers.
122
123 heating, associated with the proposed Caledonian foreland basin, was followed by Cenozoic
124 cooling, representing exhumation and final exposure of the basement. Alternative models
125 suggest that even the deepest portions of the basin, in Sweden, were largely eroded by the mid-
126 Permian, with any Mesozoic sediment cover not exceeding 100 m (Söderlund et al., 2005).
127 1.3. Reconciling the Sääksjärvi impact event and the evolution of the Caledonian
128 foreland basin
129 The suggestion that Sääksjärvi, a small impact structure in crystalline basement with no
130 evidence for incorporation of supracrustal sedimentary rocks, formed at or after 330 Ma
131 (Bottomley et al., 1990) is problematic for models that imply the region was covered with
132 sediment for much of the Phanerozoic.
133 In light of the inconsistency between estimates for the timing of the Sääksjärvi impact event
134 based on 40Ar/39Ar data (Bottomley et al., 1990; Müller et al., 1990) and evidence that the
135 basement, which formed the target rock of the Sääksjärvi impact event, was in fact covered by up
136 to 1.5 km of sediments of the Caledonian foreland basin during a long period comprising much
137 of the Phanerozoic (e.g., Larson et al., 1999), we here revisit the age of the Sääksjärvi impact
138 event with high spatial resolution U–Pb geochronology. Impact ages have long been obtained
139 from U–Pb analysis of shocked accessory phases (e.g., Kamo and Krogh, 1995; Kamo et al.,
140 2011; Krogh et al., 1984, 1993, 1996; Mänttäri and Koivisto, 2001) with recent advances
141 offering the ability to better isolate impact-age domains of heterogeneous grains. These
142 developments include decreasing analytical volumes in in situ geochronology. Another critical
143 methodological advance is the ability to recognise impact-related microstructures that ultimately
144 can be dated (e.g., Cavosie et al., 2015; Erickson et al., 2016, 2017, 2020; Kenny et al., 2017,
145 2019; Timms et al., 2017). Promisingly, microstructures in zircon that are suitable for dating
146 (such as neoblasts in recrystallised grains) are not restricted to the largest impact structures but
147 have been documented at structures as small as the 1.2 km in diameter Meteor Crater, USA
148 (Cavosie et al., 2016).
149 Here we report on imaging, microstructural analysis and U–Pb dating of shocked zircon from
150 the Sääksjärvi impact structure and present a new U–Pb age for the structure that is older than
151 previously suggested ages and consistent with the presence of the Caledonian foreland basin.
152 2. MATERIAL AND METHODS
153 Zircon grains were separated from five samples from the Sääksjärvi impact structure: three
154 samples of impact melt rock (n1047_IMP-2, n1049_IMP-1, and n5947_MS-17-1) and two
155 samples of impact melt-bearing breccia (n1048_IMP-3 and n1050_IMP-11b). Four of the
156 samples were taken from drill cores reported on by Pihlaja and Kujala (2000) and n5947_MS-17-
157 1 was a float sample. Unfortunately, the coordinates of the drill holes, as well as the intervals at
158 which these samples were collected, are not known and the precise origin of the float sample is
159 uncertain (although it was likely collected from the eastern or south-eastern shore of lake
160 Sääksjärvi or in glacial drift southeast of the structure – see Willman et al. [2010] and Fig. 1 in
161 Papunen [1969] for occurrences of impactites in the Sääksjärvi area).
162 The samples were disaggregated in a jaw crusher and ring-and-puck-style mill before their
163 heavy mineral fractions were concentrated by magnetic separation (with a hand magnet and
164 Frantz magnetic separator) and heavy liquid density separation (using methylene iodide with a
165 density of approximately 3.3 g/cm3). Zircon grains with a turbid, or cloudy, appearance were
166 preferentially picked from the heavy mineral separates as it has been shown that these are more
167 likely to be granular and contain age-reset domains than translucent grains (e.g., Schmieder et
168 al., 2015). The selected grains were placed on double-sided sticky tape and cast in 2.5 cm-
169 diameter epoxy mounts. The mounts were polished with a diamond suspension to expose grain
170 interiors and a final polish with colloidal silica prepared the grains for microstructural analysis.
171 This study comprised three separate analytical periods, in 2001 (period #1; the results of
172 which were presented by Mänttäri et al. [2004]), 2018 (period #2), and 2020 (period #3). In 2001
173 zircon grains from samples n1047_IMP-2, n1048_IMP-3, n1049_IMP-1, and n1050_IMP-11b
174 were imaged in scanning electron microscopy (SEM) backscattered electron (BSE) mode and
175 analysed for U–Pb isotopic composition and age by secondary ion mass spectrometry (SIMS) on
176 the Cameca IMS1270 ion microprobe at the NordSIMS Laboratory, Swedish Museum of Natural
177 History, with an analytical pit diameter of approximately 30 μm (generated with a duoplasmatron
178 O ion source). In 2018 zircon grains from sample n5947_MS-17-1 were imaged in BSE,
179 secondary electron (SE), and cathodoluminescence (CL) mode before undergoing U–Pb analysis
180 using the NordSIMS instrument upgraded to Cameca IMS1280, which employed an analytical
181 pit approximately 15 μm in diameter (also generated with a duoplasmatron O ion source). In
182 2020 grains from the 2001 analytical session were revisited. The grains were repolished to
183 remove the original analytical pits before being imaged in BSE, SE, and CL modes. Nine grains
184 with granular and/or porous textures were then selected for microstructural characterisation by
185 electron backscatter diffraction (EBSD) analysis (three of these had undergone U–Pb analysis in
186 2001 whereas the others had not been analysed previously). These grains then underwent U–Pb
187 isotopic analysis in which the analytical pit was approximately 5 μm in diameter (using a
188 Hyperion radio-frequency [RF] plasma ion source). Grain n1048_IMP-3_z5, which was analysed
189 with two 30 μm diameter pits in 2001 and nine 5 μm diameter pits in 2020, was subsequently
190 analysed with two 10 μm diameter analytical pits (also using the Hyperion RF plasma ion
191 source).
192 In analytical period #1, the polished surfaces of the zircon grains were imaged in BSE
193 mode. In periods #2-3, the grains were imaged in BSE, SE, and CL modes on an FEI Quanta
194 FEG 650 SEM equipped with a Gatan ChromaCL2 system at the Swedish Museum of Natural
195 History, Stockholm.
196 Electron backscatter diffraction analysis was performed with an Oxford Instruments
197 Nordlys detector attached to the FEI Quanta FEG 650 SEM. Grains were indexed for zircon,
198 reidite and monoclinic ZrO2 using match units based on crystallographic data from Hazen and
199 Finger (1979), Farnan et al. (2003), and Howard et al. (1988), respectively. Analytical conditions
200 and parameters include: accelerating voltage of 20 kV, working distance ~18 mm, stage tilt of
201 70°, electron backscatter pattern (EBSP) binning of 4x4, EBSP gain set to ‘High’, and
202 background defined with collection of 128 frames, Hough resolution set to 60, band detection
203 min/max of 6/8. Maps were collected with step sizes between 90 and 250 nm. Data collection
204 was performed in Oxford Instruments AZtec software and post-acquisition processing in Oxford
205 Instruments Channel 5 software v. 5.12. Data cleaning in Channel 5 comprised a wildspike
206 correction and a level six nearest neighbor zero solution extrapolation for all grains. Maps of
207 local misorientation represent the average misorientation between pixels in 5x5 (Fig. 3), 7x7
208 (Supplementary Figs 3-4), or 9x9 (Supplementary Fig. 5) grids. Grain boundaries, which are
209 shown in some maps, are here defined by >2° misorientation between adjacent pixels.
210 In the 2001 and 2018 U–Pb sessions, the analytical procedure largely followed the
211 routine of Whitehouse et al. (2001), with the duosplasmatron-generated primary beam yielding
212 analysis pits approximately 30 and 15 μm across, respectively. In the 2020 analytical period the
213 Hyperion-generated primary beam largely generated analysis pits approximately 5 μm across,
214 with two additional analyses performed with a 10 μm pit. A routine correction for common Pb
215 (Pbc) was applied during post-acquisition processing of the U–Pb data. The percentage of non-
206 206 216 radiogenic, i.e. common, Pb (f206%) in the total measured Pb was estimated from measured
204 217 Pb, assuming a modern-day Stacey and Kramers (1975) Pb composition, i.e. that Pbc
218 encountered in the analyses is likely to be modern and related to unavoidable sample
219 contamination (in crevices) during polishing of the grain mount (Zeck and Whitehouse, 1999). If
220 the Pbc is not modern but is actually ancient, U–Pb ages for high-Pbc analyses will be slightly
221 over-corrected, i.e. slightly too young. This is unlikely to be significant and Kenny et al. (2019)
222 reported that their concordia age for relatively Pbc-rich (f206% ranging from 0.15 to 3.08)
223 shocked zircon from the Lappajärvi impact structure would be less than 0.2 % older if the Pbc
224 encountered in those analyses was 1.9 Ga (a likely crystallisation age for those grains) and not
225 modern, as assumed. In all sessions 91500 standard zircon (ID-TIMS 206Pb/238U age = 1062.4 ±
226 0.8 Ma; Wiedenbeck et al., 1995) was used as the calibration reference material. Temora 2 (ID-
227 TIMS 206Pb/238U age = 416.78 ± 0.33 Ma; Black et al., 2004), which has consistently given ages
228 within ± 1% relative to the reference age when analysed at the NordSIMS laboratory (Jeon and
229 Whitehouse, 2015), was analysed as a secondary reference material during the 2018 session and
230 the 10 μm session in 2020, returning weighted average ages of 420.6 ± 3.0 Ma (MSWD = 0.93,
231 probability = 0.46, n = 6) and 420.6 ± 3.3 Ma (MSWD = 0.37, probability = 0.92, n = 8),
232 respectively. We used the decay constant values of Steiger and Jäger (1977) and all uncertainties
233 are presented at the 2σ level.
234 3. RESULTS
235 3.1. Imaging
236 Zircon grains from all five samples have diverse appearances with approximately half of the
237 deliberately picked cloudy grains exhibiting simple to complex internal zonation, with no
238 evidence for impact-related deformation, and half displaying granular and/or porous textures.
239 Of the apparently granular grains, some appear almost entirely recrystallised (e.g., Fig. 3) and
240 others retain conspicuous non-recrystallised domains (e.g., Fig. 4), whereas for others the degree
241 of the recrystallisation is not clear from imaging alone (e.g., Fig. 5). Individual neoblasts within
242 recrystallised grains range in size from less than 1 μm up to approximately 15 μm in dimension
243 and appear to take one of two forms: rounded granules 1-3 μm in size (e.g., Figs 3-4) or
244
245
246 Figure 3. Example of granular zircon from Sääksjärvi which preserves evidence for the former presence of reidite –
247 grain n1048_IMP-3_z13. F-G: Pole figures and plots of high-angle misorientation axes provide microstructural
248 evidence for the former presence of reidite in the form of: (i) domains of neoblasts systematically misoriented by
249 90°, (ii) coincidence among (001) and {110} poles, and (iii) high-angle misorientation axes coincident with poles to
250 {110}. Ellipses show locations of SIMS U–Pb analyses. IPF – Inverse Pole Figure.
251
252 Figure 4. Example of granular zircon from Sääksjärvi which does not preserve evidence for the former presence of
253 reidite – grain n1048_IMP-3_z4 (contrast with Fig. 3). Dashed ellipse A shows the approximate location of a 30 μm
254 analytical pit (session #1, 2001) prior to repolishing for imaging and microstructural analysis shown here.
255
256 Figure 5. A zircon grain (n1049_IMP-1_z11) from Sääksjärvi which appears granular in backscattered electron and
257 cathodoluminescence imaging but microstructural analysis reveals that apparent sub-domains share a single
258 approximate orientation.
259
260 subhedral to euhedral crystallites which are generally larger (e.g., Fig. 6). Some grains display
261 both small, rounded granules and larger, euhedral crystallites (e.g., Fig. 6).
262 Grains with pores generally also display some degree of recrystallisation elsewhere in the
263 grain (e.g., Fig. 6 and Supplementary Figs 1-2). Pores are usually round and approximately 0.2
264 μm in diameter; in some grains they are quite homogeneously distributed (e.g., Fig. 6) whereas in
265 others they occur in bands, suggestive of a possible relationship with pre-impact zoning within
266 the zircon (e.g., Supplementary Figs 1-2).
267 3.2. Microstructural characterisation
268 Six of the nine grains selected for EBSD analysis indexed well, producing whole-grain
269 microstructural maps (Figs. 3-5 and Supplementary Figs 3-5), whereas one grain only indexed on
270 approximately 20% of the exposed surface (Fig. 6) and two did not produce electron backscatter
271 diffraction patterns and so were not mapped (see imaging in Supplementary Figs 1-2).
272 In all grains zircon was the only phase to index. No evidence was encountered for ZrO2 (a
273 product of zircon dissociation at high temperature, which commonly accompanied granular
274 textures in impact-related zircon; e.g., Timms et al., 2017; Cavosie et al., 2018) or the current
275 presence of the high-pressure ZrSiO4 polymorph, reidite (e.g., Glass et al., 2002).
276 Microstructural characterisation reveals that many grains are almost entirely composed of
277 low-strain, newly crystallised domains, neoblasts (Fig. 3), whereas others retain non-
278 recrystallised domains defined by high degrees of strain apparent in maps of local misorientation
279 (e.g., Supplementary Figs 3,5). Of the seven grains which displayed at least some degree of
280 recrystallisation and were analysed by EBSD, four preserve crystallographic evidence for the
281 former presence of reidite in the form of: (i) domains of neoblasts systematically misoriented by
282
283 Figure 6. Shocked zircon that most accurately and precisely records the age of the Sääksjärvi impact event – grain
284 n1048_IMP-3_z5. The grain displays both granular (B) and porous (C) textures. Dashed ellipses A and B show the
285 approximate location of 30 μm analytical pits (session #1, 2001) prior to repolishing for imaging and microstructural
286 analysis shown here. Of the other analysis pits, C to K are approximately 5 μm in diameter and L and M are
287 approximately 10 μm in diameter. The 608 ± 8 Ma concordia age was obtained from data from analytical pits C, D,
288 G, and K, which are shown in red.
289 90°, (ii) coincidence among (001) and {110} poles, and (iii) high-angle misorientation axes
290 coincident with poles to {110} (Fig. 3, Supplementary Figs 3-5), which are indicative of the
291 transformation sequence zircon → reidite → zircon (Cavosie et al., 2016, 2018, Timms et al.,
292 2017). Of the three remaining granular grains, one displays a non-recrystallised core (with up to
293 5° internal misorientation) that is surrounded by granules with apparently random orientations
294 (Fig. 4), one appears granular in BSE and CL images but microstructural analysis shows that
295 sub-domains share a single approximate orientation (and deformation occurs in the form of
296 misorientation up to approximately 10° across the grain; Fig. 5), and the third is partly granular
297 and partly porous, with neoblasts not displaying any systematic relationships (Fig. 6).
298 3.3. U–Pb geochronology
299 The first U–Pb analyses, with analytical pit diameters of approximately 30 μm, were
300 performed in 2001 and target sites for the 43 analyses (on 34 zircon grains from four samples)
301 were based on BSE images. The data define two discordant trends in U–Pb space: one between
302 ca. 1.85 Ga and ca. 600 Ma and the other between ca. 600 Ma and present (Fig. 7A). The ca.
303 1.85 Ga ages are consistent with the Svecofennian ages of the target rocks at Sääksjärvi (Fig. 2)
304 whereas the ca. 600 Ma lower intercept age is not known from any rocks in the area. Most
305 analyses which plot to the right of the ca. 1.85 Ga to ca. 600 Ma discordant array in Tera-
306 Wasserburg (TW) space (Fig. 7A) have 207Pb/206Pb ages of ca. 600 Ma, consistent with later Pb
307 loss superimposed on grains which were fully age reset at ca. 600 Ma. However, a few grains
308 that lie off the ca. 1850–600 Ma trend have 207Pb/206Pb ages >600 Ma, indicating later Pb loss
309 superimposed on grains which were only partially reset at ca. 600 Ma.
310 The second batch of U–Pb analyses, with analytical pit diameters of approximately 15 μm,
311 was undertaken in 2018 and, for this analytical session, target sites for 22 analyses (on 17 zircon
312
313 Figure 7. Concordia diagrams displaying all secondary ion mass spectrometry (SIMS) U–Pb data gathered in this
314 study. A: Data generated with 30 and 15 μm analytical spots (using duoplasmatron ion source during SIMS
315 analyses). B: Data generated with 10 and 5 μm analytical spots (using Hyperion RF plasma ion source during SIMS
316 analyses). C: Data for grain n1048_IMP-3_z5 (see Fig. 5) – colours as in B. All data rendered at 2σ uncertainty.
317 grains from a previously unanalysed sample of impact melt rock, n5947_MS-17-1) were selected
318 based on BSE, SE, and CL images of the grains. Four analyses plot on or close to the concordia
319 curve at ca. 1850 Ma, consistent with the previously recognised Svecofennian ages; one analysis
320 lies on the ca. 1850–600 Ma discordia trend; and the remaining analyses plot to the right of the
321 discordia trend, again consistent with later Pb loss superimposed on the original discordant array.
322 Additionally, a single, almost concordant analysis gave a 207Pb/206Pb age of ca. 2.6 Ga (Fig. 7A),
323 highlighting an age not recognised in the earlier U–Pb dataset.
324 The third set of U–Pb analyses, most with analytical pit diameters of approximately 5 μm and
325 two with pit diameters of 10 μm, were performed in 2020. This time, target locations for 50
326 analyses (on nine zircon grains from three of the samples first analysed in 2001) were based on
327 EBSD maps and complementary BSE, SE, and CL images (except for the two grains which did
328 not index well and were targeted on the basis of traditional images alone; Supplementary Figs 1-
329 2). On the basis of microstructural characterisation, we targeted (i) individual neoblasts and
330 domains of neoblasts (with the aim to date recrystallisation), as well as (ii) non-recrystallised,
331 high-strain domains, and (iii) porous, non-indexing regions in three grains. Similar to earlier
332 analyses, the new dataset defines a trend between ca. 1.85 Ga and ca. 600 Ma with
333 approximately 15 analyses plotting to the right of the array in TW space, indicative of later Pb
334 loss superimposed on the original discordant array (Fig. 7B). However, many isotopic ratios and
335 ages from this session have significantly larger uncertainties than encountered previously – a
336 result of a higher proportion of analyses incorporating significant amounts of Pbc (40% of
337 analyses had f206% >1 compared to 14% of analyses in the 2001 session), likely due to targeting
338 of granular grains. One grain was particularly rich in Pbc, with calculated f206% values ranging
339 between 4.7 and 9.3 (grain n1050_IMP-11b_z18; Supplementary Fig. 4). The application of a
340 Pbc correction to the seven data points from this grain result in over-correction with six of the
341 Pbc-corrected data points lying below the concordia curve in TW space.
342 Plotting a regression through 56 of the 115 analyses from all analytical sessions (excluding
343 analyses which clearly lie off the ca 1.85 Ga to ca. 600 Ma trend, i.e. those with ages >2.0 Ga,
344 those which experienced significant later Pb loss, and those with substantial Pbc components)
345 yields a discordia line with a 1856 ± 15 Ma upper intercept and a 597 ± 7 Ma lower intercept
346 (mean square of weighted deviates [MSWD] = 2.1) (Supplementary Fig. 7). The relatively high
347 MSWD indicates some scatter in the data, which is likely related to subtle post-impact Pb loss
348 even in this screened population, and perhaps also heterogeneous pre-impact ages of the zircon
349 grains. Calculating a weighting mean 207Pb/206Pb age from the 21 analyses defining the
350 horizontal array between ca. 600 Ma and present day gives an age of 679 ± 18 Ma (MSWD =
351 3.4) (Supplementary Fig. 7). The relatively high MSWD on this age again indicates scatter in the
352 data, which, in this instance, can be attributed to variable degrees of Pb loss during the ca. 600
353 Ma event.
354 One grain in the 2020 dataset gave concordant data points at the ca. 600 Ma lower intercept
355 of the discordant array which lack significant Pbc contamination and therefore have small
356 uncertainties (Fig. 7B-C). Four concordant, low Pbc (f206% values of 0.01, 0.01, 0.02, and 0.06)
357 analyses on grain n1048_IMP-3_z5 give a concordia age (Ludwig, 1998) of 608 ± 8 Ma (2σ;
358 MSWD of concordance and equivalence = 1.06; probability of concordance and equivalence =
359 0.39) (Fig. 7C). Given the low Pbc values of these four analyses, whether the Pbc is modern, as
360 assumed in the Pbc correction, or ancient will not affect the age at this level of precision.
361 Significantly, the concordant data points with low Pbc were from the porous portion of this
362 mixed porous-granular grain (Fig. 6A,E and Supplementary Fig. 6; see also Supplementary Fig.
363 8 for a concordia diagram coloured according to zircon texture). Following the nine analyses
364 with 5 μm diameter analytical pits, this grain was targeted with two 10 μm diameter analytical
365 pits. However, neither of these two analyses were concordant, with one (targeting the porous
366 domain) plotting back along the discordia array and the other (targeting a large neoblast) having
367 experienced later Pb loss (Fig. 7B-C).
368 4. DISCUSSION
369 4.1. Age of the Sääksjärvi impact event
370 We interpret the ca. 1.85 Ga to ca. 600 Ma array of U–Pb data presented here to indicate
371 partial Pb loss during an Ediacaran impact into Palaeoproterozoic Svecofennian rocks of the
372 Baltic Shield. This was first suggested by Mänttäri et al. (2004) on the basis of the data acquired
373 in the 2001 analytical period. Within the context of the 115 analyses on 57 zircon grains from
374 five suevite and impact melt rock samples (over three analytical periods) the 608 ± 8 Ma
375 concordia age (Fig. 7C) from a porous zircon grain represents the best age estimate for the
376 Sääksjärvi impact event. We prefer the 608 ± 8 Ma age over the 597 ± 7 Ma lower intercept age
377 calculated from 56 analyses as the high MSWD of the latter indicates some scatter, which is
378 likely related to post-impact Pb loss in even that screened population. The 608 ± 8 Ma age is also
379 favourable to the weighted mean 207Pb/206Pb age calculated from the horizontal array as the older
380 age of the latter can be attributed to variable degrees of incomplete impact-induced Pb loss (as
381 reflected by its high MSWD value).
382 Our new age for the Sääksjärvi impact event is older than previous estimates of ≤330 Ma and
383 ca. 560 Ma, which were based on 40Ar/39Ar data, and makes Sääksjärvi the first confirmed
384 Ediacaran impact structure in Finland and the wider Baltic Shield (given that the 546 ± 5 Ma for
385 the Gardnos impact structure, Norway [Kalleson et al., 2009] is based on ambiguous U–Pb data
386 [e.g., Jourdan et al., 2009]). Furthermore, Sääksjärvi becomes only the second confirmed
387 Ediacaran impact structure worldwide, with the Acraman impact structure, Australia, constrained
388 to this time by the recognition of an associated ejecta layer in the Ediacaran Bunyeroo Formation
389 (Williams and Gostin, 2005; Schmieder et al., 2015, and references therein).
390 4.2. Implications for the Caledonian foreland basin
391 As originally proposed by Mänttäri et al. (2004), the new Ediacaran age for the Sääksjärvi
392 impact event reconciles the presence of this impact structure in crystalline basement with
393 regional sedimentary cover of the Caledonian foreland basin persisting from its Siluro-Devonian
394 onset through much of the Phanerozoic, and eliminates complications raised by a previously
395 proposed age of ≤330 Ma (Fig. 8). Furthermore, the ca. 330 Ma component in the 40Ar/39Ar data
396 of Bottomley et al. (1990) may have resulted from heating caused by burial related to Caledonian
397 sediment loading. The spread of 40Ar/39Ar data between ca. 330 Ma and ca. 650 Ma (Bottomley
398 et al., 1990) may now be reinterpreted as partial resetting of the 40Ar/39Ar system as a result of
399 burial and heating that is recognised in apatite fission-track datasets throughout the Baltic Shield
400 (e.g., Cederbom et al., 2000; Murrell and Andriessen, 2004; Larson et al., 1999). However, it is
401 not possible to ascertain whether the partial resetting related to burial and heating (caused by the
402 overburden of sediment) was superimposed on almost complete resetting of the 40Ar/39Ar system
403 at the 608 ± 8 Ma impact or if it simply represents partial resetting from predominantly
404 Palaeoproterozoic target rock ages. The interpretation that the 40Ar/39Ar data is recording
405 Paleozoic burial is also consistent with late Paleozoic lower-intercept ages of zircon across the
406 Baltic Shield, considered a result of Pb having been leached from the grains by hydrothermal
407 solutions related to the Caledonian Orogeny and foreland basin (Larson and Tullborg, 1998).
408 Similarly, a ca. 385 Ma age obtained from 40Ar/39Ar analysis of impact melt lithologies from the
409
410 Figure 8. Age distribution of impact structures in the Baltic region. A: Previous interpretations of 40Ar/39Ar data for
411 Sääksjärvi (shown in red) overlying the distribution of other impact structures formed in crystalline basement. B:
412 New U–Pb age from this study (shown in green) and distribution which includes the new age for Sääksjärvi. C: Age
413 data for all impact structures of the Baltic region considered in this study (see Fig. 1). Histogram coloured according
414 to scheme of Fig. 1. Five impact structures have poor age constraints and are not represented (see Table 1 in
415 Supplementary Appendix).
416
417 Gardnos impact structure, Norway, has been interpreted to represent thermal overprinting of the
418 Caledonian Orogeny and not the impact event itself (Grier et al., 1999), although this conclusion
419 has been contested (Jaret et al., 2016).
420 4.3. Implications for dating impact events with shocked zircon
421 It is striking that the four precise, concordant data points used to calculate the new age for the
422 Sääksjärvi impact event were from porous, non-indexing (i.e., poorly crystalline) domains of a
423 zircon grain and not from the many recrystallised, or granular, domains analysed.
424 A number of recent studies have, in addition to analysing recrystallised (neoblastic)
425 accessory phases such as zircon and apatite, also obtained impact ages from non-granular, but
426 pore-rich grains (e.g., McGregor et al., 2018, 2019; Schmieder et al., 2019; Schwarz et al., 2020).
427 However, most studies which have incorporated microstructural analysis have focused on
428 targeting recrystallised (neoblastic) domains of shocked minerals (e.g., Erickson et al., 2017,
429 2020; Kenny et al., 2017). An advantage of recrystallised grains is that, in addition to potentially
430 recording an impact age, they may also provide diagnostic evidence of shock if they record
431 crystallographic evidence for the former presence of reidite (Cavosie et al., 2016, 2018; Timms
432 et al., 2017). Porous textures, on the other hand, are not uniquely impact-related and, therefore,
433 an age determined from this texture in ex situ grains (for example, in a detrital sample) cannot be
434 considered a definitive impact age. We suggest that our findings at Sääksjärvi highlight a
435 potential pitfall of restricting analyses to a single textural type and that studies incorporating
436 microstructural analysis should not limit U–Pb analyses to recrystallised domains alone but
437 should also include porous textures if they are present.
438 Our data suggest that the use of an approximately 5 μm diameter analytical spot (enabled by
439 the recent installation of a Hyperion RF ion source at the NordSIMS laboratory) to isolate fully
440 age-reset domains within heterogeneous zircon grains is likely to significantly aid U–Pb dating
441 of impact events. Four of the nine 5 μm analytical spots on grain n1048_IMP-3_z5 gave
442 concordant, low-Pbc data points that enabled calculation of the 608 ± 8 Ma concordia age. In
443 contrast, none of the two 30 μm spots or two 10 μm spots on the same grain gave concordant
444 data. It therefore seems likely that if analyses were restricted to ≥ 10 μm, concordant data may
445 not have been obtained. We note that a number of recent studies have isolated impact-aged
446 domains with a 10 μm analytical spot (e.g., Erickson et al., 2017, 2020; Kenny et al., 2017, 2019;
447 Schwarz et al., 2020) but suggest that the advent of 5 μm spot analyses will aid targeting of
448 specific domains in complex grains, increase the likelihood of fully isolating impact-age regions,
449 and lead to more concordant data points (thereby improving uncertainties on impact ages).
450 Our new findings also reinforce the utility of U–Pb analysis of shocked accessory phases to
451 complement 40Ar/39Ar dating for small impact structures. Interpretation of 40Ar/39Ar spectra is
452 difficult for small impact structures, where the targets rock may not have been heated sufficiently
453 to completely degas radiogenic argon, 40Ar*, previously accumulated in the rock (e.g., Jourdan et
454 al., 2007; Pickersgill et al., 2020). Here, we show that complete resetting of the U–Pb system can
455 be achieved at structures as small as Sääksjärvi and note that is likely possible at even smaller
456 structures too, such as the 1.2 km-diameter Meteor Crater, Arizona, USA where recrystallised
457 zircon were documented (Cavosie et al., 2016). Additionally, 40Ar/39Ar dating is difficult at
458 structures for which there is a large time gap between target rock ages and the cratering event
459 (meaning that even minor degrees of excess 40Ar will significantly affect the calculated age;
460 Jourdan et al., 2007), but this is optimal for the U–Pb system in zircon whereby a large age
461 difference means that the impact-related discordant array intersects the concordia curve at a high
462 angle, allowing one to distinguish impact-related Pb loss from later, e.g., modern, Pb loss (e.g.,
463 Fig. 7A). Conversely, 40Ar/39Ar may be more applicable for craters for which there was a
464 relatively limited time gap between the target rock crystallisation age and the impact event (e.g.,
465 impacts into Phanerozoic target rocks), where U–Pb data will be distributed along the concordia
466 curve (as opposed to intersecting it), making it more complicated to reach unambiguous impact
467 ages (e.g., Erickson et al., 2017; Tohver et al., 2012). Finally, the higher closure temperature of
468 the U–Pb system in zircon relative to the 40Ar/39Ar system means that it may be more likely to
469 record the timing of the impact event and not later cooling of the impact structure (e.g., Kenny et
470 al., 2019; Schmieder and Kring, 2020) and that once the structure has cooled, zircon may be less
471 likely to be compromised by later heating, as appears to have affected the 40Ar/39Ar system at
472 Sääksjärvi.
473 4.4. Radiometric dating of small impact structures
474 Our new dataset highlights the potential of accurate and precise U–Pb dating of small impact
475 structures in crystalline basement to test, and even place boundaries on, models for the timing
476 and extent of palaeobasins that might otherwise be difficult to constrain. Although our new age
477 for the Sääksjärvi impact structure does not place tighter constraints on the timing of the
478 Caledonian foreland basin than were available from, for example, apatite fission-track studies, it
479 demonstrates the possibility that future radiometric dating of impact structures may do so. For
480 example, the Karikkoselkä impact structure, approximately 150 km northeast of Sääksjärvi, is a
481 1.5 km-diameter structure in crystalline basement that could provide new constraints, but its age
482 is still ambiguous: Pesonen et al. (1999) suggested that palaeomagnetic data are consistent with
483 ages of 230-260 Ma, 530-560 Ma, and 1650-1760 Ma, whereas Schmieder et al. (2010)
484 suggested a possible age as young as ca. 9 Ma on the basis of (U–Th)/He analyses of zircon.
485 More broadly, radiometric dating of small impact structures in crystalline basement
486 worldwide may provide constraints on sedimentary basins, the burial histories of
487 palaeolandscapes, and associated tectonics in less well-studied regions. In practise, this potential
488 utility will be hampered by the fact that most impact structures which formed on Earth have been
489 removed by erosion (e.g., Hergarten and Kenkmann, 2015; Johnson and Bowling, 2014), with
490 small impact structures more easily erased from the geological record than larger ones (which
491 produced a deeper record in the lithosphere). Today there are only approximately 110 known
492 impact structures on Earth that have diameters less than 6 km, the approximate size of Sääksjärvi
493 (Schmieder and Kring, 2020). Of these, approximately 60 are between 1.2 and 6 km in diameter,
494 meaning that at least this number are likely amenable to radiometric dating (given that datable
495 microstructures in zircon were identified at the 1.2 km in diameter Meteor Crater, USA [Cavosie
496 et al., 2016]). Encouragingly, most impact structures yet to be discovered on Earth’s surface are
497 likely to be less than 6 km in diameter (with as many as 90 in the 1 to 6 km range; Hergarten and
498 Kenkmann, 2015) and thus their radiometric ages may be of use in regional geological
499 reconstructions.
500 5. CONCLUSIONS
501 New U–Pb data for shocked zircon indicates that the Sääksjärvi impact structure, southwest
502 Finland, formed at 608 ± 8 Ma. This makes Sääksjärvi the first unambiguously dated Ediacaran
503 impact structure in the Baltic Shield, and one of only two confirmed Ediacaran impact structures
504 worldwide.
505 The 608 ± 8 Ma age for Sääksjärvi is older than previous estimates based on 40Ar/39Ar data
506 and confirms that the impact into exposed crystalline basement occurred prior to the
507 development of the Caledonian foreland basin. Apparently younger 40Ar/39Ar ages are likely a
508 result of Caledonian sediment loading and burial.
509 The accurate and precise dating of Sääksjärvi reinforces the utility of shocked accessory
510 phases in radiometric dating of small impact structures. Furthermore, the most accurate and
511 precise age for Sääksjärvi was acquired from porous, poorly crystalline zircon, indicating that
512 future studies should not focus on recrystallisation textures alone.
513 The use of a 5 μm diameter analytical spot (enabled by the Hyperion RF ion source) to
514 isolate fully age-reset domains within heterogeneous zircon grains is likely to significantly aid
515 U–Pb dating of impact events.
516 Dating of small impact structures (both those which are currently known and those yet to be
517 discovered) may help constrain the lifetime and spatial extent of sedimentary basins worldwide.
518 ACKNOWLEDGMENTS
519 We thank Kerstin Lindén for support with sample preparation , Heejin Jeon and Lev Ilyinsky for
520 assistance with ion microprobe analyses, Nick Timms and Timmons Erickson for constructive
521 reviews, and Deta Gasser for efficient editorial handling. This is NordSIMS Contribution No.
522 653 and Lunar and Planetary Institute Contribution No. 2386. The Lunar and Planetary Institute
523 is operated by Universities Space Research Association under a cooperative agreement with the
524 Science Mission Directorate of the National Aeronautics and Space Administration.
525
526 FUNDING
527 This project received funding from the European Union’s Horizon 2020 research and innovation
528 programme under the Marie Skłodowska-Curie Individual Fellowship Grant Agreement No.
529 792030, supporting GGK.
530
531 REFERENCES CITED
532 Black L. P., Kamo S. L., Allen C. M., Davis D. W., Aleinikoff J. N., Valley J. W., Mundil R.,
533 Campbell I. H., Korsch R. J., Williams I. S., and Foudoulis C. (2004) Improved
534 206Pb/238U microprobe geochronology by the monitoring of a trace-element-related matrix
535 effect; SHRIMP, ID–TIMS, ELA–ICP–MS and oxygen isotope documentation for a
536 series of zircon standards. Chemical Geology 205, 115–140.
537 https://doi.org/10.1016/j.chemgeo.2004.01.003
538 Bottomley, R. J., York, D., and Grieve, R. A. F. (1990) 40Argon-39Argon dating of impact
539 craters. Proceedings of the Lunar and Planetary Science Conference 20, 421–431.
540 Buchner, E., Moilanen, J., Öhman, T., and Schmieder, M. (2009) Shock-Molten Sandstone
541 Clasts in Impact Melt Rocks: Age Constraints for the Paasselkä Impact Structure (SE
542 Finland). 40th Lunar and Planetary Science Conference, abstract no. 2169.
543 Cavosie A. J., Erickson T. M., Timms N. E., Reddy S. M., Talavera C., Montalvo S. D., Pincus
544 M. R., Gibbon R. J., and Moser D. (2015) A terrestrial perspective on using ex situ
545 shocked zircons to date lunar impacts. Geology 43, 999–1002.
546 https://doi.org/10.1130/g37059.1
547 Cavosie, A. J., Timms, N. E., Erickson, T. M., Hagerty, J. J., and Hörz, F. (2016)
548 Transformations to granular zircon revealed: Twinning, reidite, and ZrO2 in shocked
549 zircon from Meteor Crater (Arizona, USA). Geology 44, 703–706.
550 https://doi.org/10.1130/g38043.1
551 Cavosie, A. J., Timms, N. E., Ferrière, L., and Rochette, P. (2018) FRIGN zircon—The only
552 terrestrial mineral diagnostic of high-pressure and high-temperature shock deformation.
553 Geology 46, 891–894. https://doi.org/10.1130/G45079.1
554 Cederbom, C., Larson, S. Å., Tullborg, E.-L., and Stiberg, J.-P. (2000) Fission track
555 thermochronology applied to Phanerozoic thermotectonic events in central and southern
556 Sweden. Tectonophysics 316, 153–167. https://doi.org/10.1016/S0040-1951(99)00230-9
557 Erickson, T. M., Cavosie, A. J., Pearce, M. A., Timms, N. E., and Reddy, S. M. (2016) Empirical
558 constraints on shock features in monazite using shocked zircon inclusions: Geology 44,
559 635–638. https://doi.org/10.1130/G37979.1.
560 Erickson, T. M., Kirkland, C. L., Timms, N. E., Cavosie, A. J., and Davison, T. M. (2020)
561 Precise radiometric age establishes Yarrabubba, Western Australia, as Earth’s oldest
562 recognised meteorite impact structure. Nature Communications 11, 300.
563 https://doi.org/10.1038/s41467-019-13985-7
564 Erickson, T. M., Timms, N. E., Kirkland, C. L., Tohver, E., Cavosie, A. J., Pearce, M. A., and
565 Reddy, S.M. (2017) Shocked monazite chronometry: integrating microstructural and in
566 situ isotopic age data for determining precise impact ages. Contributions to Mineralogy
567 and Petrology 172, 11. https://doi.org/10.1007/s00410-017-1328-2
568 Farnan, I., Balan, E., Pickard, C. J., and Mauri, F. (2003) The effect of radiation damage on local
29 569 structure in the crystalline fraction of ZrSiO4: Investigating the Si NMR response to
570 pressure in zircon and reidite. American Mineralogist 88, 1663–1667.
571 https://doi.org/10.2138/am-2003-11-1205
572 Garfunkel, Z., and Greiling, R. O. (1998) A thin orogenic wedge upon thick foreland lithosphere
573 and the missing foreland basin. Geologische Rundschau 87, 314–325.
574 https://doi.org/10.1007/s005310050212
575 Glass, B. P., Liu, S., and Leavens, P. B. (2002) Reidite: An impact-produced high-pressure
576 polymorph of zircon found in marine sediments. American Mineralogist 87, 562–565.
577 https://doi.org/10.2138/am-2002-0420
578 Grier, J. A., Swindle, T. D., Kring, D. A., and Melosh, H. J. (1999) Argon-40/argon-39 analyses
579 of samples from the Gardnos impact structure, Norway. Meteoritics and Planetary
580 Science 34, 803–807. https://doi.org/10.1111/j.1945-5100.1999.tb01393.x
581 Grieve, R. A. F. (2006) Impact structures in Canada. Geological Association of Canada GEOtext
582 Number 5, 210 pp. ISBN: 978-1-897095-11-9
583 Grieve, R. A. F., and Head III, J. W. (1983) The Manicouagan impact structure: An analysis of
584 its original dimensions and form. Journal of Geophysical Research: Solid Earth 88,
585 A807–A818. https://doi.org/10.1029/JB088iS02p0A807
586 GTK (2017) Bedrock of Finland 1:200 000, v. 2.1 https://gtkdata.gtk.fi/Maankamara/index.html
587 accessed: 16th January 2020.
588 Hazen, R. M., and Finger, L. W. (1979) Crystal structure and compressibility of zircon at high
589 pressure. American Mineralogist 64, 196–201.
590 Hergarten, S., and Kenkmann, T. (2015) The number of impact craters on Earth: Any room for
591 further discoveries? Earth and Planetary Science Letters 425, 187–192.
592 http://dx.doi.org/10.1016/j.epsl.2015.06.009
593 Howard, C. J., Hill, R. J., and Reichert, B. E. (1988) Structures of ZrO2 polymorphs at room
594 temperature by high‐resolution neutron powder diffraction. Acta Crystallographica
595 Section B 44, 116–120. https://doi.org/10.1107/S0108768187010279
596 Jaret, S. J., Hemming, S. R., Rasbury, E. T., Ramezani, J., and Jacobsen, S. B. (2016) The case
597 for an impact at the Gardnos impact structure at 385 Ma. 79th Annual Meeting of the
598 Meteoritical Society, abstract no. 6476.
599 Jeon, H., and Whitehouse, M. J. (2015) A critical evaluation of U–Pb calibration schemes used
600 in SIMS zircon geochronology. Geostandards and Geoanalytical Research 39, 443–452.
601 https://doi.org/10.1111/j.1751-908X.2014.00325.x
602 Johnson, B. C., and Bowling, T. J. (2014) Where have all the craters gone? Earth’s bombardment
603 history and the expected terrestrial cratering record. Geology 42, 587–590.
604 https://doi.org/10.1130/G35754.1
605 Jourdan, F. (2012) The 40Ar/39Ar dating technique applied to planetary sciences and terrestrial
606 impacts. Australian Journal of Earth Sciences 59, 199–224.
607 https://doi.org/10.1080/08120099.2012.644404
608 Jourdan, F., Renne, P. R., and Reimold, W. U. (2007) The problem of inherited 40Ar* in dating
609 impact glass by the 40Ar/39Ar method: Evidence from the Tswaing impact crater (South
610 Africa). Geochimica et Cosmochimica Acta 71, 1214–1231.
611 https://doi.org/10.1016/j.gca.2006.11.013
612 Jourdan, F., Renne, P. R., and Reimold, W. U. (2009) An appraisal of the ages of terrestrial
613 impact structures. Earth and Planetary Science Letters 286, 1–13.
614 https://doi.org/10.1016/j.epsl.2009.07.009
615 Kalleson, E., Corfu, F., and Dypvik, H. (2009) U–Pb systematics of zircon and titanite from the
616 Gardnos impact structure, Norway: Evidence for impact at 546 Ma? Geochimica et
617 Cosmochimica Acta 73, 3077–3092. https://doi.org/10.1016/j.gca.2009.02.020
618 Kamo, S. L., and Krogh, T. E. (1995) Chicxulub crater source for shocked zircon crystals from
619 the Cretaceous-Tertiary boundary layer, Saskatchewan: Evidence from new U-Pb data.
620 Geology 23, 281–284. https://doi.org/10.1130/0091-
621 7613(1995)023<0281:CCSFSZ>2.3.CO;2
622 Kamo, S. L., Lana, C., and Morgan, J. V. (2011) U–Pb ages of shocked zircon grains link distal
623 K–Pg boundary sites in Spain and Italy with the Chicxulub impact. Earth and Planetary
624 Science Letters 310, 401–408. https://doi.org/10.1016/j.epsl.2011.08.031
625 Kenny, G. G., Morales, L. F., Whitehouse, M. J., Petrus, J. A., and Kamber, B. S. (2017) The
626 formation of large neoblasts in shocked zircon and their utility in dating impacts.
627 Geology 45, 1003–1006. https://doi.org/10.1130/G39328.1
628 Kenny, G. G., Schmieder, M., Whitehouse, M. J., Nemchin, A. A., Morales, L. F., Buchner, E.,
629 Bellucci, J. J., and Snape, J. F. (2019) A new U-Pb age for shock-recrystallised zircon
630 from the Lappajärvi impact crater, Finland, and implications for the accurate dating of
631 impact events. Geochimica et Cosmochimica Acta 245, 479–494.
632 https://doi.org/10.1016/j.gca.2018.11.021
633 Krogh, T. E., Davis, D. W., and Corfu, F. (1984) Precise U-Pb zircon and baddeleyite ages for
634 the Sudbury area. In: Pye, E. G., Naldrett, A. J., and Giblin, P. E. (eds) The Geology and
635 Ore Deposits of the Sudbury Structure. Ontario Geological Survey, Special Volume 1,
636 431–446.
637 Krogh T. E., Kamo S. L., and Bohor B. F. (1993) Fingerprinting the K/T impact site and
638 determining the time of impact by U-Pb dating of single shocked zircons from distal
639 ejecta. Earth and Planetary Science Letters 119, 425–429. https://doi.org/10.1016/0012-
640 821X(93)90150-8
641 Krogh T. E., Kamo S. L., and Bohor B. F. (1996) Shock metamorphosed zircons with correlated
642 U-Pb discordance and melt rocks with concordant protolith ages indicate an impact origin
643 for the Sudbury structure. In: Basu, A., and Hart, S. (eds) Earth processes: reading the
644 isotopic code. American Geophysical Union 95, 343–353.
645 arson, S. ., and Tullborg, E.-L. (1998) Why Baltic Shield zircons yield late Paleozoic, lower-
646 intercept ages on U-Pb concordia. Geology 26, 919–922. https://doi.org/10.1130/0091-
647 7613(1998)026<0919:WBSZYL>2.3.CO;2
648 Larson, S. Å., Tullborg, E.-L., Cederbom, C., and Stiberg, J.-P. (1999) Sveconorwegian and
649 Caledonian foreland basins in the Baltic Shield revealed by fission-track
650 thermochronology. Terra Nova 11, 210–215. https://doi.org/10.1046/j.1365-
651 3121.1999.00249.x
652 Lehnert, O., Meinhold, G., Bergström, S. M., Calner, M., Ebbestad, J. O. R., Egenhoff, S., Frisk,
653 Å. M., Hannah, J. L., Högström, A. E. S., Huff, W. D., Juhlin, C., Maletz, J., Stein, H. J.,
654 Sturkell, E., and Vandenbroucke, T. R. A. (2012) New Ordovician–Silurian drill cores
655 from the Siljan impact structure in central Sweden: an integral part of the Swedish Deep
656 Drilling Program. GFF 134, 87–98. https://doi.org/10.1080/11035897.2012.692707
657 ehnert, O., Meinhold, G., Wu, R., Calner, M., and Joachimski, M. M. (2014) δ13C
658 chemostratigraphy in the upper Tremadocian through lower Katian (Ordovician)
659 carbonate succession of the Siljan district, central Sweden. Estonian Journal of Earth
660 Sciences 63, 277–286. https://doi.org/10.3176/earth.2014.31
661 Ludwig, K. R. (1998) On the treatment of concordant uranium-lead ages. Geochimica et
662 Cosmochimica Acta 62, 665–676. https://doi.org/10.1016/S0016-7037(98)00059-3
663 Mänttäri, I., Kohonen, J., Kujala, H., and Pihlaja, P. (2004) A revised age for the Sääksjärvi
664 meteorite impact, southwestern Finland: The connexion with a Caledonian foreland
665 basin. 32nd International Geological Congress (Florence, Italy), abstract no. 1434.
666 Mänttäri, I., and Koivisto, M. (2001) Ion microprobe uranium-lead dating of zircons from the
667 Lappajärvi impact crater, western Finland. Meteoritics and Planetary Science 36, 1087–
668 1095. https://doi.org/10.1111/j.1945-5100.2001.tb01946.x
669 McGregor, M., McFarlane, C. R. M., and Spray, J. G. (2018) In situ LA-ICP-MS apatite and
670 zircon U–Pb geochronology of the Nicholson Lake impact structure, Canada: Shock and
671 related thermal effects. Earth and Planetary Science Letters 504, 185–197.
672 https://doi.org/10.1016/j.epsl.2018.10.006
673 McGregor, M., McFarlane, C. R. M., and Spray, J. G. (2019) In situ multiphase U–Pb
674 geochronology and shock analysis of apatite, titanite and zircon from the Lac La
675 Moinerie impact structure, Canada. Contribution to Mineralogy and Petrology 174, 62.
676 https://doi.org/10.1007/s00410-019-1598-y
677 Middleton, M. F., Tullborg, E.-L., Larson, S. Å., and Björklund, L. (1996) Modelling of a
678 Caledonian foreland basin in Sweden: petrophysical constraints. Marine and Petroleum
679 Geology 13, 407–413. https://doi.org/10.1016/0264-8172(95)00087-9
680 Müller, N., Hartung, J. B., Jessberger, E. K., and Reimold, W. U. (1990) 40Ar‐39Ar ages of
681 Dellen, Jänisjärvi, and Sääksjärvi impact craters. Meteoritics 25, 1–10.
682 https://doi.org/10.1111/j.1945-5100.1990.tb00964.x
683 Murrell, G. R., and Andriessen, P. A. M. (2004) Unravelling a long-term multi-event thermal
684 record in the cratonic interior of southern Finland through apatite fission track
685 thermochronology Physics and Chemistry of the Earth 29, 695–706.
686 https://doi.org/10.1016/j.pce.2004.03.007
687 Papunen, H. (1969) Possible impact metamorphic textures in the erratics of the Lake Sääksjärvi
688 area in southern Finland. Bulletin of the Geological Society of Finland 41, 151-155.
689 Papunen, H. (1973) Chemical composition and origin of the shock metamorphic rocks of the
690 Sääksjärvi area, Finland. Bulletin of the Geological Society of Finland 45, 29–34.
691 Pesonen, L. J., Elo, S., Lehtinen, M., Jokinen, T., Puranen, R., and Kivekäs, L. (1999) Lake
692 Karikkoselkä impact structure, central Finland: new geophysical and petrographic results.
693 In: Dressler, B. O., and Sharpton, V. L. (eds) Large Meteorite Impacts and Planetary
694 Evolution II, GSA Special Paper 339, Geological Society of America, Boulder, CO, pp
695 131–148.
696 Pickersgill, A. E., Mark, D. F., Lee, M. R., and Osinski, G. R. (2020) 40Ar/39Ar systematics of
697 melt lithologies and target rocks from the Gow Lake impact structure, Canada.
698 Geochimica et Cosmochimica Acta 274, 317–332.
699 https://doi.org/10.1016/j.gca.2020.01.025
700 Pihlaja, P. J., and Kujala, H. (2000) Some geochemical features of impact melts, suevites and
701 target rocks at Sääksjärvi impact structure, Finland. In: Plado, J., and Pesonen, L. J. (eds)
702 Programme and Abstracts of ESF workshop: Meteorite Impacts in Precambrian Shields.
703 Geological Survey of Finland and University of Helsinki, Espoo: p 50
704 Puura, V., and Plado, J. (2005). Settings of meteorite impact structures in the Svecofennian
705 crustal domain. In: Koeberl, C., and Henkel, H. (eds) Impact tectonics. Springer, Berlin,
706 Heidelberg.: pp. 211–245.
707 Schmidt, G., Palme, H., and Kratz, K. L. (1997) Highly siderophile elements (Re, Os, Ir, Ru, Rh,
708 Pd, Au) in impact melts from three European impact craters (Sääksjärvi, Mien, and
709 Dellen): Clues to the nature of the impacting bodies. Geochimica et Cosmochimica Acta,
710 61, 2977–2987. https://doi.org/10.1016/S0016-7037(97)00129-4
711 Schmieder, M., and Kring, D. A. (2020) Earth's Impact Events Through Geologic Time: A List
712 of Recommended Ages for Terrestrial Impact Structures and Deposits. Astrobiology 20,
713 91–141. https://doi.org/10.1089/ast.2019.2085
714 Schmieder, M., Buchner, E., Jourdan, F., Schwarz, W. H., Trieloff, M., van Soest, M. C.,
715 Wartho, J. A., Hodges, K. V., Moilanen, J., Hietala, S., and Öhman, T. (2010) Updating
716 the Finnish Impact Cratering Record. 41st Lunar and Planetary Science Conference,
717 abstract no. 2036.
718 Schmieder, M., Tohver, E., Jourdan, F., Denyszyn, S. W., and Haines, P. W. (2015) Zircons
719 from the Acraman impact melt rock (South Australia): Shock metamorphism, U–Pb and
720 40Ar/39Ar systematics, and implications for the isotopic dating of impact events.
721 Geochimica et Cosmochimica Acta 161, 71–100.
722 https://doi.org/10.1016/j.gca.2015.04.021
723 Schmieder, M., Shaulis, B. J., Lapen, T. J., Buchner, E., and Kring, D. A. (2019) In situ U–Pb
724 analysis of shocked zircon from the Charlevoix impact structure, Québec, Canada.
725 Meteoritics and Planetary Science 54, 1808–1827.
726 https://doi.org/10.1111/maps.13315
727 Schwarz, W. H., Hanel, M., and Trieloff, M. (2020) U‐Pb dating of zircons from an impact melt
728 of the Nördlinger Ries crater. Meteoritics and Planetary Science 13437.
729 https://doi.org/10.1111/maps.13437
730 Slater, B. J., and Willman, S., (2019) Early Cambrian small carbonaceous fossils (SCFs) from an
731 impact crater in western Finland. Lethaia 52, 570–582. https://doi.org/10.1111/let.12331
732 Stacey, J. S., and Kramers, J. D. (1975) Approximation of terrestrial lead isotope evolution by a
733 two-stage model. Earth and Planetary Science Letters 26, 207–221.
734 https://doi.org/10.1016/0012-821X(75)90088-6
735 Steiger, R. H., and Jäger, E. (1977) Subcommission on geochronology: convention on the use of
736 decay constants in geo-and cosmochronology. Earth and Planetary Science Letters 36,
737 359–362. https://doi.org/10.1016/0012-821X(77)90060-7
738 Söderlund, P., Juez-Larré, J., Page, L. M., and Dunai, T. J. (2005) Extending the time range of
739 apatite (U–Th)/He thermochronometry in slowly cooled terranes: Palaeozoic to Cenozoic
740 exhumation history of southeast Sweden. Earth and Planetary Science Letters 239, 266–
741 275. https://doi.org/10.1016/j.epsl.2005.09.009
742 Tagle, R., Schmitt, R. T., and Erzinger, J. (2009) Identification of the projectile component in the
743 impact structures Rochechouart, France and Sääksjärvi, Finland: implications for the
744 impactor population for the earth. Geochimica et Cosmochimica Acta 73, 4891–4906.
745 https://doi.org/10.1016/j.gca.2009.05.044
746 Timms, N. E., Erickson, T. M., Pearce, M. A., Cavosie, A. J., Schmieder, M., Tohver, E., Reddy,
747 S. M., Zanetti, M. R., Nemchin, A. A., and Wittmann, A. (2017) A pressure-temperature
748 phase diagram for zircon at extreme conditions. Earth-Science Reviews 165, 185–202.
749 https://doi.org/10.1016/j.earscirev.2016.12.008
750 Tohver E., Lana C., Cawood P., Fletcher I., Jourdan F., Sherlock S., Rasmussen B., Trindade R.,
751 Yokoyama E., Souza Filho C., and Marangoni, Y. (2012) Geochronological constraints
752 on the age of a Permo–Triassic impact event: U–Pb and 40Ar/39Ar results for the 40 km
753 Araguainha structure of central Brazil. Geochimica et Cosmochimica Acta 86, 214–227.
754 https://doi.org/10.1016/10.1016/j.gca.2012.03.005
755 Wiedenbeck M., Allé P., Corfu F., Griffin W. L., Meier M., Oberli F., Quadt A. V., Roddick J.
756 C., and Spiegel W. (1995) Three natural zircon standards for U-Th-Pb, Lu-Hf, trace
757 element and REE analyses. Geostandards Newsletter 19, 1–23.
758 https://doi.org/10.1111/j.1751-908X.1995.tb00147.x
759 Williams, G. E., and Gostin, V. A. (2005) Acraman – Bunyeroo impact event (Ediacaran), South
760 Australia, and environmental consequences: twenty-five years on. Australian Journal of
761 Earth Sciences 52, 607–620. https://doi.org/10.1080/08120090500181036
762 Willman, S., Plado, J., Raukas, A., and Bauert, H. (2010) Meteorite Impact Structures –
763 Geotourism in the Central Baltic. NGO Geoguide Baltoscandia, Tallinn, Estonia, 80 pp.
764 ISBN: 978-9985-9973-2-1
765 Zeck, H. P., and Whitehouse, M. J. (1999) Hercynian, Pan-African, Proterozoic and Archean
766 ion-microprobe zircon ages for a Betic-Rif core complex, Alpine belt, W Mediterranean –
767 consequences for its P-T-t path. Contributions to Mineralogy and Petrology 134, 134–
768 149. https://doi.org/10.1007/s004100050474
769
770 Supplementary data
771 Supplementary Table 1 (Data used to produce map in Fig. 1 in main text)
772 Supplementary Figure 1 (Imaging of grain n1049_IMP-1_z4)
773 Supplementary Figure 2 (Imaging of grain n1049_IMP-1_z7)
774 Supplementary Figure 3 (Imaging and EBSD analysis of grain n1050_IMP-11b_z16)
775 Supplementary Figure 4 (Imaging and EBSD analysis of grain n1050_IMP-11b_z18)
776 Supplementary Figure 5 (Imaging and EBSD analysis of grain n1050_IMP-11b_z14)
777 Supplementary Figure 6 (Post-analysis imaging of grain n1048_IMP-3_z5)
778 Supplementary Figure 7 (Concordia diagram with regression through sub-population)
779 Supplementary Figure 8 (Concordia diagram coloured according to zircon texture)
780
781
Name Longitude Latitude Country Size (km) Age (Ma) Target rock Notes
Dellen 16.70 61.86 Sweden 19 141 Crystalline Gardnos 09.00 60.65 Norway 5 546 Crystalline The U–Pb data for Gardnos are considered equivocal. Granby 14.93 58.42 Sweden 3 466 Sedimentary Hummeln 16.25 57.37 Sweden 1.2 465 Mixed Iso-Naakkima 27.15 62.18 Finland 3 1050 Sedimentary Age range: 1200-900 Ma Jänisjärvi 30.92 61.97 Russia 14 687 Crystalline Kaali 22.67 58.40 Estonia 0.11 0.00324 Sedimentary Impact into unconsolidated material Kärdla 22.77 59.02 Estonia 4 455 Mixed Karikkoselkä 25.25 62.22 Finland 1.5 Crystalline Age range 260-230 Ma, or 9 Ma Keurusselkä 24.60 62.13 Finland 30 1151 Crystalline Lappajärvi 23.67 63.15 Finland 23 78 Mixed Lockne 14.82 63.00 Sweden 10.75 455 Mixed Diameter range: 7.5-14km Lumparn 20.13 60.15 Finland 9 Mixed Age < 458 Ma Målingen 14.55 62.92 Sweden 0.7 455 Mixed Mien 14.86 56.42 Sweden 9 122 Crystalline Neugrund 23.67 59.33 Estonia 20 535 Crystalline Age range: 540-530 Ma Paasselkä 29.08 62.03 Finland 10 231 Mixed Ritland 06.44 59.23 Norway 2.7 520 Mixed Age range: 540-500 Ma Sääksjärvi 22.40 61.40 Finland 6 608 Crystalline Saarijärvi 28.38 65.28 Finland 1.5 560 Crystalline Age range: 600-520 Ma Siljan 14.87 61.03 Sweden 52 381 Mixed Söderfjärden 21.58 63.03 Finland 6.6 Crystalline Age range: 1880-640 Ma Summanen 25.38 62.65 Finland 2.6 Crystalline Age < 1880 Ma Suvasvesi N 28.17 62.70 Finland 3.5 85 Crystalline Suvasvesi S 28.22 62.60 Finland 3.8 Crystalline Age range: 1880-710 Ma Tvären 17.42 58.77 Sweden 2 458 Mixed
Within bounds of map but not considered
Dobele 23.25 56.58 Latvia Ilumetsä 27.42 57.97 Estonia Mishina Gora 28.05 58.72 Russia Vepriai 24.58 55.08 Lithuania 782
783 Supplementary Table 1. Data used to produce map in Fig. 1 in main text. Data from Schmieder and Kring (2020),
784 except age for the Sääksjärvi impact structure which is from this study.
785 786 Supplementary Figure 1. Imaging of grain n1049_IMP-1_z4.
787 788 Supplementary Figure 2. Imaging of grain n1049_IMP-1_z7.
789 790
791 792 Supplementary Figure 3. Imaging and EBSD analysis of grain n1050_IMP-11b_z16. 793
794 795 Supplementary Figure 4. Imaging and EBSD analysis of grain n1050_IMP-11b_z18.
796 797 Supplementary Figure 5. Imaging and EBSD analysis of grain n1050_IMP-11b_z14. 798
799 800 Supplementary Figure 6. Post-analysis backscattered electron imaging of grain n1048_IMP-3_z5. The bright
801 material surrounding the grain and in some analytical pits and vesicles is gold remnant from the coating for ion
802 microprobe analysis. Pits would have been free from gold after analysis but some was trapped in a number of the
803 pits during brief polishing to remove the coating from the rest of the grain. No evidence for the presence of ZrO2
804 was encountered. The 608 ± 8 Ma concordia age was calculated from data from analytical pits C, D, G, and K.
805 806 Supplementary Figure 7. Tera-Wasserburg concordia diagram with a regression through the sub-population
807 defining the ca. 1850 to ca. 600 Ma trend and a weighted mean 207Pb/206Pb age of analyses which record later Pb
808 loss. MSWD – mean square of weighted deviates. 809
810 811 Supplementary Figure 8. Tera-Wasserburg concordia diagram coloured according to zircon texture. Note that
812 information on texture is not available for analytical period #1 (undertaken in 2001) and so these data are not
813 presented.
814