Meteorite impact craters and possibly impact-related structures in Estonia




Abstract– Three structures (Neugrund, Kärdla, and Kaali) of proven impact origin make Estonia the most cratered country in the world by area. In addition, several candidate impact structures exist, waiting for future studies to determine their origin. This article is an overview of these proven and possible impact structures, including some breccia layers. It summarizes the information and descriptions of the morphology; geological characteristics; and mineralogical, chemical, and geophysical data available in the literature. The overview was prepared to make information in many earlier publications in local journals (many of which had been published in Estonian or Russian) accessible to the international community. This review summarizes the facts and observations in a historical fashion, summarizing the current state of knowledge with some additional comments, and providing the references.


The first successful link between a terrestrial geological feature and a meteorite impact was made in 1905 when D. M. Barringer convincingly argued that Meteor Crater in Arizona had been formed by the impact of a high-speed iron meteorite. Soon after, the idea reached Europe. In 1910, A. G. Högbom, while reporting on Meteor Crater at a meeting of the Swedish Geological Society, suggested that the Swedish Mien and Dellen lakes could represent meteorite impact structures due to the presence of unusual melt rocks in these lake basins. Unfortunately, the idea remained ridiculed until the beginning of the 1960s when Swedish geologists F. E. Wickman and N.-B. Svensson finally proved the impact origin of these two structures (Svensson and Wickman 1965; Svensson 1968).

Reference to impact cratering as a geological process has finally appeared in textbooks during the last decades, as the popularity of the subject increased dramatically. The increased exposition and publicity have produced a boom in identifying new impact structures (e.g., Earth Impact Database,; Dypvik et al. 2008), but also some threats. Several Web-distributed databases have appeared, some listing impact structures without giving any proof of origin (for indicative criteria and verification of an impact origin see, e.g., Henkel and Pesonen 1992; Koeberl 2002; French and Koeberl 2010). If cases without generally accepted impact features such as shatter cones, petrographic shock effects, or geochemical signatures are repeatedly referred to, this may lead to misconceptions and misinformation of the public. This also applies to the Estonian record (Table 1) where the attribute “possible impact origin” is sometimes missing. In Estonia, the impact origin of the Kaali structures that had been known already for centuries (described already in 1827 by J.W.L. von Luce) was confirmed only in 1938 after extensive field excavation (Reinvaldt 1933). A “meteoritic origin” had, however, been suggested already much earlier in a textbook on general geology by the Estonian schoolteacher J. Kalkun (1922) who was interested in Meteor Crater.

Table 1.   General characteristics of proven and possible impact structures of Estonia.
NameCoordinatesaKnown dimensionsImpactitesTarget affected by crater/structureAgeMorphology
Lat. NLong. E
  1. aIn case of crater fields, center coordinates of the largest structure are given;

  2. D = diameter; DR = rim (rim-to-rim) diameter; DER = external ring diameter; DCP = central peak diameter; dT = true depth; dA = apparent depth; dR = rim height. For locations of crater structures see Fig. 1.

Proven impact craters
Neugrund59°20′23°32′DR = 7 km
DER = 20 km
Lithic breccia,
breccia dikes
Water (50 m),
Ediacaran to Cambrian sediments (100 m),
crystalline basement
Approximately 535 MaComplex submarine crater, buried under sedimentary rocks, rim exposed
Kärdla58º59′22º48′DR = 4 km
DER = 12…15 km
DCP≈ 0.3 km
dT = 0.4 km
dA = 0.2 km
hR = 0.1 km
Lithic breccia, ejecta, breccia dikesWater (<100 m),
Ediacaran to Ordovician sediments (140 m),
crystalline basement
Approximately 455 MaComplex crater, buried under sedimentary rocks
Kaalijärv58º22′22″22º40′10″Main crater:
DR = 105…110 m
dT = approximately 30 m
dA = 22 m
hR = 4…8 m
Fragments of projectile, lithic brecciasGlacial till (<10 m), Silurian dolostonesHoloceneCrater field, 8 simple craters, exposed, the main crater is filled with a water-body and some sediments; others are dry, partly ruined by human activity
Ilumetsa57º57′36″27º24′11″Main structure:
DR = 76 m
dA = 9 m
hR = 1…4.5 m
Not observedGlacial till, Devonian sandstonesHoloceneTwo simple structures, exposed, the bigger structure hosts bog
Tsõõrikmäe58º4′50″27º27′51″DR = 38…40 m
dA = 5.8 m
hR = 1…2.5 m
Not observedGlacial till (approximately 6 m), Devonian clays and sandstonesHoloceneSimple structure, exposed, filled with water and peat, partly ruined by draining system
Simuna59º3′50″26º22′39″DR = 8.9 m
dT = 2.1 m
dA = 1.4 m
hR = 0…0.3 m
Not observedSand (1.1 m),
glacial till
1 June 1937; not directly observedSimple structure, exposed
Lasnamäe59º26′29″24º51′26″Fractures in an area of 3 × 5 mNot observedOrdovician limestone, possibly also soft sediments20,000–25,000 yr bpSet of radial and tangential fractures, no crater form, covered by soil

Proven Impact Craters of Estonia


The Neugrund crater, located in the southwestern part of the Gulf of Finland (Fig. 1; Table 1), is the latest proven (Suuroja and Suuroja 2000) meteorite impact structure in Estonia. It is also the oldest one, with a stratigraphic age of approximately 535 Ma (Suuroja and Suuroja 2000), i.e., Early Cambrian. A shallow, circular central plateau of postimpact sediments fills this submarine structure and had been long known to sailors, but only became geologically studied as late as in the 1980s (Malkov et al. 1986; Talpas et al. 1993) in the course of regional geological–geophysical mapping of the sea around the former USSR and Estonia. However, on Osmussaar Island (Fig. 1), large (i.e., up to tens of meters in size), brecciated, erratic boulders (Fig. 2) that nowadays are known to represent pieces from the crystalline rim of Neugrund were observed and photographed already in the 1920s (Öpik 1927). It took almost 70 yr before Suuroja and Saadre (1995) suggested that these boulders might be associated with an impact structure.

Figure 1.

 Geological sketch map of the Estonian sedimentary basement and a cross-section showing the locations of proven and suggested impact structures. MG = Mishina Gora meteorite impact crater in Russia. The dotted line shows approximately the northern distributional boundary of the Middle Devonian Narva Breccias. Distribution of erratic boulders (area between the two dashed lines) derived by glacial action from the exposed rim of the Neugrund structure is from Suuroja and Suuroja (2004). Distribution of the Osmussaar Breccias is redrawn from Alwmark et al. (2010).

Figure 2.

 Impact breccia boulder composed of clasts of gneiss as found on the coastline in NW Estonia. The boulder originates from the exposed part of the underwater crystalline rim of the Neugrund impact structure. Pen for scale is approximately 13 cm long.

The suggestion was further supported by earlier findings (Kala and Eltermann 1969; Suuroja et al. 1987) of an up to 19 m thick, strongly deformed and/or brecciated layer in cores drilled in the course of small-scale (1 : 200,000) geological mapping of the basement in the onshore area and on Osmussaar Island. This layer is overlain by so-called Osmussaar Breccias (see below), which are significantly younger and not related to the Neugrund impact. The distal breccia layer (ejecta and tsunami deposits) produced by the Neugrund impact lies within the Early Cambrian Lontova (regional) Stage of greenish-gray claystone (Suuroja and Suuroja 2000; Suuroja et al. 2002a). The stratigraphic age of the distal breccia layer and occurrences of the postimpact succession down to Early Cambrian silt and sandstones (observed by diving into the erosional canyon-like feature (Ring Canyon; Suuroja and Suuroja 2004) between postimpact crater fill and crystalline rim serve as the basis for dating the structure and suggest that it originated by impact into a shallow marine environment. The target for the Neugrund bolide was composed of water, 100 m of Lower Cambrian clay-, silt-, and sandstones, underlain by Svecofennian (1.8–1.9 Ga; Claesson et al. 2001) crystalline basement.

Shock metamorphism in the Neugrund crater is evidenced by findings of planar deformation features with up to four orientations and spacings of 2.5–5 μm (Fig. 3) in quartz from the rim and distal breccias (for details and illustrations see Suuroja and Suuroja 2000, 2004).

Figure 3.

 Shock-metamorphised quartz grain with three sets of partially decorated PDF. Sample from a vein of matrix-supported impact breccia from the brecciated crystalline rim of the Neugrund crater. Cross-polarized light.

The Neugrund structure has a rim diameter of about 7 km; its depth is unknown (Fig. 4). Impact structures of such size most certainly have a central uplift, but it could not be directly observed in Neugrund because the depression is filled by postimpact sediments and allochthonous breccias of unknown thickness. The topmost layers of the postimpact infill are composed of Middle and Upper Ordovician limestones and dolostones that form an approximately 4.5 km wide central plateau over the basin. These sedimentary rocks are relatively resistant, especially when compared with fractured lithologies, and have protected the structure from severe erosion. Nowadays, only a few meters of water cover the plateau making it difficult to apply traditional marine seismic methods. The plateau is surrounded by a 200–500 m wide and 20–70 m deep erosional canyon, which is most pronounced on the northern side of the structure. The bottom of the canyon is filled with Quaternary sediments. At the outer side of the canyon, the eroded rim of crystalline rocks still stands out. It extends 60–120 m above the target surface of crystalline bedrock and is 2.5–3 km wide, whereas on the NW side, three separate ridges are outlined (Suuroja and Suuroja 2004). Based on reinterpretations of decades-old data as well as new results (10 marine expeditions between 1996 and 2006 [Suuroja and Suuroja 2010] including seismic reflection profiling, side scan sonar profiling, observing submarine outcrops by a video robot, diving, and sampling), Suuroja and Suuroja (2004) outlined a ring fault 20 km in diameter (related to a 10–60 m high terrace) and a 3–5 km wide area (circular depression) between the rim wall and the fault (Fig. 4). Within the depression, the crystalline basement is covered by a 10–50 m thick pile of Quaternary sediments. However, the existence, exact position, and structure of the ring fault and of the circular depression need further studies (the published data represent copies of old analog prints, whereas no data from the 2000 and 2001 seismic profiling have been published so far). Also, relations of the observed faults to impact need further analysis, as displacements of basement rocks due to nonimpact processes are common at the bottom of the Baltic Sea (e.g., Tuuling and Flodén 2001) and known also on the mainland.

Figure 4.

 Cross-section of the Neugrund impact structure (redrawn after Suuroja et al. 2002a).

Aeromagnetic data (terrain clearance 300 m, profiles every 500 m, readings every 50 m; Metlitckaya and Papko 1992) over the Neugrund structure show a local negative anomaly. However, the anomaly is about the same wavelength and amplitude (a few hundreds of nT) as the regional NEE-trending elongated magnetic features, and no clear circular magnetic anomaly can be seen (Suuroja et al. 2002a). An arc of positive anomalies that spatially corresponds to the rim wall surrounds the central minimum from the north, east, and southeast. Neither a central uplift nor possible impact-melt-related features of Neugrund are evident in the aeromagnetic field, but measurements from sea/ice may reveal new details. To adjust the age and the inner structure of Neugrund, drilling from the central plateau is needed.


The first hints that Kärdla could be an impact crater on Hiiumaa Island, western Estonia (Fig. 1; Table 1) were given by Eichwald (1843) and Schrenk (1854), who described abnormally tilted (up to 20° whereas the normal tilt of Estonian sedimentary rocks is <1°) limestones in Paluküla (a village at the NE edge of the Kärdla crater) quarry. In the 1920s, the Paluküla area received attention because of the discovery of hydrocarbons (Scupin 1924), such as diffused patches of oil in limestones, liquid oil in pores, and asphalt (Suuroja 2002). In the beginning of 1967, while drilling a well for groundwater, crystalline rocks in Paluküla were discovered occurring at a very shallow depth of 22 m, whereas the typical depth for the region is 250 m. At that time, the geological feature was connected to tectonic movements related to the Caledonian Orogeny (Viiding et al. 1969). The find initiated search for crystalline building stone, including an extensive drilling program, as well as gravity and ground magnetic mapping (Barankina and Gromov 1973; Suuroja et al. 1974; Kala et al. 1976) that revealed a crater structure of 4 km diameter.

The impact origin of Kärdla was proven by the discovery of quartz with planar deformation features (PDF; Masaitis et al. 1980) in allochthonous breccia. The most frequent PDF orientations were found along the {10inline image3} and {10inline image2} directions (Kirsimäe et al. 2002; Puura et al. 2004), corresponding to moderate shock pressures up to 35 GPa (e.g., Stöffler and Langenhorst 1994). Planar deformation features in quartz of the distal ejecta blanket were found as well (Fig. 5) (Suuroja and Suuroja 2006; Kleesment et al. 2006). Quartz in the ejected material has often been found coated by carbon-rich amorphous material or intergrown with apatite. The carbon isotope composition is reminiscent of the Early Ordovician Dictyonema shale (uppermost part of the target; Kleesment et al. 2006), but also coincides with the general extraterrestrial carbon isotope composition (Deines 1989). Apatite is of postimpact origin, formed prior to or simultaneous with the formation of diagenetic carbonate cementation (Kleesment et al. 2006).

Figure 5.

 Quartz grain with three sets of partially decorated PDF. Sample of matrix-supported impact breccia from the Kärdla crater. Drill core K-18 (center of the crater), depth 325.5 m. Cross-polarized light.

The Kärdla impact took place in an epicontinental sea of 20 (Puura and Suuroja 1992; Ormö and Lindström 2000) to 50 m (Suuroja et al. 2002b), or even 100 m (Suuroja and Suuroja 2006), depth, at approximately 455 Ma (dated biostratigraphically: the lowermost Diplograptus multidens graptolite zone, early Idavere age, lowermost Caradoc, Upper Ordovician; Grahn et al. 1996). Thus, the crater belongs to a family of structures that formed during the Middle to Late Ordovician phase of likely enhanced flux (see Schmitz et al. 2001) of L chondrites after the “chondrite” event at 470 ± 6 Ma (Korochantseva et al. 2007). Under the seawater, the target was composed of Ordovician and Cambrian sedimentary cover (140 m of limestones, sandstones, siltstones, and clays) overlying the Precambrian crystalline basement (Puura and Suuroja 1992).

The crater depression is filled with allochthonous fragmental breccias, covered by resurge conglomerates, conglomeratic turbidites, and sands that were eroded from uplifted walls prior to burial (Puura and Suuroja 1992) (Fig. 6). Seawater backsurge likely created (Suuroja et al. 2002b) at least two gullies through the crystalline rim; however, spatially varied degrees of rim uplift and collapse could have been the initial cause (Jõeleht and Plado 2010). Sedimentation resumed immediately after the impact and the structure was infilled and rims covered after about 7 Myr (Ainsaar et al. 2002). During its predominantly continental period from Carboniferous to Paleogene (Puura et al. 1999), Kärdla was never reopened and its inner structure is perfectly preserved. Also, the ejecta blanket is well preserved (as assessed with 42 drill holes; Suuroja and Suuroja 2006), except at the crater rim where it was eroded prior to the formation of the sedimentary cover. The ejecta blanket lies in a succession of Upper Ordovician carbonate rocks as a 0.01–9.6 m thick bed of fragmented and disintegrated Cambrian and Lower Ordovician sand, silt, and clay, with minor admixture of fragments from all the other lithological units of the target. The deposit is well cemented, mostly by calcite with dolomite admixture, and contains sparse, disseminated authigenic pyrite crystals (Kleesment et al. 1987). Thickness and grain size of the bed decrease with distance; the most distal occurrence was discovered in a drill core at 42 km from the crater center (Suuroja and Suuroja 2006).

Figure 6.

 A) Simplified cross-section of the Kärdla impact structure. B) Ground-penetrating radar section over the NE rim of Kärdla. The parallel reflections highlight layering of the post-impact Ordovician limestones, which are tilted inward and outward at the inner and outer slopes, respectively. Dielectric permittivity value 9 was used to convert the time-scale to depth-scale in B. Please note that the vertical scale is exaggerated approximately 20 times. Details of the ground-penetrating radar study can be found in Jõeleht and Plado (2010).

Suuroja et al. (2002b) suggested an iron meteorite as the projectile for the Kärdla impact. This suggestion received criticism (Puura et al. 2004) due to relatively low contents of Ni, Co, and Ir in impact breccias, which they thought favored some chondritic addition. To confirm an extraterrestrial component, full analyses of the platinum group elements or isotopic studies are required. Puura et al. (2001, 2002) reported magnetic metalliferous particles, rich in litho- and chalcophile elements, from suevite of Kärdla and some other impact structures. Further analyses (unpublished data by the author) of non–impact-related material have shown that the method used by Puura et al. was inadequate for separation of uncontaminated magnetic particles.

The thermal pulse caused by the impact produced hydrothermal activity that was active till about 9000–10,000 yr after the impact (Jõeleht et al. 2005). The presence of a hydrothermal system is proven by mineralogical, chemical, and stable C and O isotope studies (Kirsimäe et al. 2002; Versh et al. 2005), whereby three evolutionary stages in the development of the system have been outlined: (i) an early vapor-dominated stage (precipitation of adularia type K-feldspar); (ii) the main two-phase (vapor to liquid) stage (chlorite/corrensite, euhedral K-feldspar, quartz); and (iii) a late liquid-dominated stage (calcite, dolomite, quartz, chalcopyrite/pyrite, Fe-oxyhydrate). Precipitation of different hydrothermal parageneses occurred simultaneously at different locations as cooling was site-specific: the fastest cooling occurred in the groundwater recharge area in the crater depression, whereas the central uplift retained higher temperatures for longer times (Jõeleht et al. 2005). Versh et al. (2006) conducted a search for possible fossil remains of the impact-induced hydrothermal microbial activity, but with no success. The chemical anomaly (K-enrichment and Ca-Na-depletion of feldspar- and hornblende-bearing rocks in the allochthonous breccia units and subcrater basement [parautochthonous breccias]) that gradually decreases with depth is also interpreted (Puura et al. 2000, 2004) to have been caused by early-stage alteration by an impact-induced hydrothermal system.

Nowadays, due to extensive drilling programs during the Soviet era (until 1991), Kärdla has been investigated with more than 100 drill holes (Kala et al. 1984). Most of the holes are short, as they are located on the rim (most of them on the NE rim), but three centrally located holes open up the postimpact sediments and allochthonous infill (the latter is divided from bottom to top, into allogenic, slumped, fallback, and resurge breccias; see Suuroja et al. 2002b) for research. Two of the deeper interior holes, K-1 (Põldvere 2002) and K-18, proved the existence of a small (height approximately 100 m, diameter approximately 300 m) central uplift. The uplift is not recognizable from ground gravity or magnetic maps (Puura and Suuroja 1992) because it is located at considerable depth and there is insufficient petrophysical contrast between the central uplift and its surroundings (Plado et al. 1996).

Based on space and aerial photographs (Puura and Suuroja 1992), and also on marine seismic reflection profiles (Suuroja et al. 2002b), Kärdla is surrounded by an elliptical (SW–NE elongated) feature (ring fault). The 2–8 km wide zone between the rim and the ring feature is reported (Suuroja et al. 2002b) as being strongly disturbed, including the presence of fractured and folded blocks. However, the existence and the exact nature of the ring feature require further studies.

The Kärdla structure causes an about −3 mGal local gravity anomaly that is surrounded by an up to 2 mGal positive anomaly. The negative anomaly is partly due to the low-density postimpact sediments infilling the structure, and the low density of impact breccias, whereas the gravity high reflects the uplifted and denser crystalline rim. The local magnetic anomaly (the amplitude of the vertical component is −100 nT above the depression and up to +400 nT above the crystalline rim) looks almost the same as the gravity anomaly. It highlights the locations of relatively higher sections of the crystalline rim. Also, the recent reflection seismic and ground-penetrating radar (GPR) investigations (Jõeleht and Plado 2010) show that the rim of Kärdla did not form with uniform height, thus, leading to differential erosion by the resurging sea. In the NE segment, the crystalline rim formed and remained high, whereas the topmost part was levelled due to erosion before it was covered by late Ordovician layered carbonate sediments (Fig. 6). The variations in rim height can be attributed either to oblique impact from the NE or differences in the physical properties of the crystalline target.


The Kaali meteorite crater field (Fig. 7) on Saaremaa Island, western Estonia (Fig. 1, Table 1), consists of nine structures, including the main crater 110 m in diameter (Fig. 8), “Kaalijärv” (järv = lake in the Estonian language), that formed in a layered target of Quaternary till overlying Silurian (Paadla Stage, Ludlow) dolostones. With the discovery of meteoritic iron in the summer of 1937, Reinwald (1938) finalized a long epoch of search for the origin of the Kaali structures (Reinvaldt 1933) that was started with a description of circular topographic features plus uplifted and fractured dolomite blocks at the Kaali site in 1827 by the naturalist J.W.L. von Luce. Thus, several earlier suggestions for their origin (gas explosions, oozing out of a bed of clay, karst weathering, solution of rock-salt from salt domes, or expansion of anhydrite by hydration, excavated by human; overviews are given by Spencer 1938; Aaloe 1963a; and Raukas et al. 2005) were rejected. Meri (1976) tied the Kaali phenomenon to the descriptions in a book by Pytheas from Massalia (Marseilles), who visited Saaremaa between 350 and 325 bc (for an overview and discussion, see, e.g., Veski et al. 2001, 2007; and references therein). However, if the glassy spherules found in the mires of Saaremaa and Hiiumaa, and dated to 7500–7600 yr bp (Raukas et al. 1995; Raukas 2000a; see discussion below), originate from the Kaali event, this connection would be wrong.

Figure 7.

 Aerial photograph (Estonian Land Board) of the Kaali crater field with all the identified circular structures and their diameters indicated. Location of the main crater in respect to the field favors a SSE azimuth of the projectile.

Figure 8.

 Cross-section of the main Kaali crater structure (after Tiirmaa 1994).

The Kaali craters were taken under heritage protection in November 1937; however, the continuous destructive excavations combined with farming and road-building have partly ruined the original shapes and structures, especially those of the satellite structures. During the Soviet era, Saaremaa and other islands of Estonia belonged to a border zone with restricted access, meaning practically no tourism. After the collapse of the USSR, Kaalijärv became a popular and frequently visited natural monument, because of its spectacularly exposed impact features of perfectly round shape and prominent rim with outcrops of outward tilted limestone layers at the inner slope. In 2005, a local nonprofit company established a museum featuring local geology and introducing meteoritics as exemplified by the Kaali craters.

After World War II research was concentrated on (i) collecting and studying the remnant pieces of the meteorite and micrometeorites all around the crater field, (ii) characterizing the structure of the craters by extensive excavations and some geophysical methods, (iii) finding links between the impact and archeological finds, and (iv) dating the event.

Altogether 2.5 kg of coarse octahedrite of class IAB (Spencer 1938; Bronsten 1962; Aaloe 1968) pieces have (officially) been collected at Kaali; most are stored in the meteoritical collection of the Institute of Geology, Tallinn Technical University. In addition to iron, the material contains 7.25 wt% of Ni, but is also rich in rare elements like Ir, Ga, Ge, Re, Pt, and Au (Yavnel 1976; Kracher et al. 1980). Mineralogical studies (Yudin 1968) of fragments from the Kaali crater field revealed typical iron meteorite minerals like kamacite (mean abundance = 96.8 vol%;), taenite (1.8 vol%), and schreibersite (1.7 vol%) (Yudin and Smyshlyayev 1963). Plastic deformation structures for kamasite and taenite, and brittle deformation of kamasite, schreibersite, and olivine grains in kamacite were found (Yudin et al. 1982). Based on the sizes of the Kaali structures, compositions of the target, and the projectile, Bronsten and Stanyukovich (1963) estimated the initial mass (400–10,000 t) and velocity (15–45 km s−1) of the projectile, which were reduced to 20–80 t and 10–20 km s−1 at the moment of impact, respectively.

In the 1970s, the hunt for pulverized meteoritic matter started. Aaloe and Tiirmaa (1981) studied finds of microscopic magnetite- and/or silicate-rich material (already reported by Aaloe 1958; Yudin and Smyshlyayev 1963) in soils within the Kaali craters and in their surroundings (5–10 g m−3). Much higher concentrations (up to 60 g m−3) of such magnetic material, collected from 15 to 45 cm depth in the cultivated fields, were occasionally found in Saaremaa and even on the mainland of Estonia, approximately 80 km from Kaalijärv (Tiirmaa 1988). All these finds were interpreted and referred to as resulting from the flight of the Kaali meteor and its explosion (Tiirmaa 1992; Tiirmaa and Raukas 1996) without considering other possible sources, e.g., anthropogenic pollution. A map of the distribution of such material by Tiirmaa (1988), further modified by Shymanovich et al. (1993), highlights a road from the mainland to Kuressaare, the largest town in Saaremaa, and matches the areas of pronounced agricultural activity.

As the site was known to be rich in magnetic fragments of the meteorite, detailed (1 × 1 m grid) magnetic mapping of the Kaalijärv main crater and crater #3 (Fig. 7) was carried out in 1955 (Pobul 1958). No considerable magnetic anomalies were observed, except weak (up to 10 nT) local microanomalies in crater #3. In 1957, the location of the most prominent anomaly was excavated and 201 g of meteorite fragments (per 4.2 m3 of soil) of 1–5 mm size were collected (most from a depth of 0.5–1 m; Pobul 1963). Application of ground seismic (refraction) methods and vertical electrical sounding (VES) around the main crater (Aaloe et al. 1976) revealed a zone of shattered rocks at twice the width of the crater. Later (Aaloe et al. 1982), apparent resistivity measurements (profiling with the Schlumberger electrode array) were carried out inside and around the main crater and craters #3, #1, and #6 (Fig. 7). Slight undulations (100–500 Ωm) were recorded and, unconvincingly, tied in with the direction (east–west) of the projectile. The ESE direction was suggested by Reinvaldt (1933) while describing the triangled funnel at the bottom of fractured dolostone of crater #4, opened in 1927. However, distribution of the structures in the field favors a SSE direction (also noted by Krinow in 1960), as the largest crater is located at, or near, the downrange boundary of the crater strewn field (e.g., Passey and Melosh 1980). However, while tracing an ellipse of distribution with free flight of imagination, a wide range of directions from east to south may be considered.

The Kaalijärv structures suffered extensive scientific excavation during the entire 20th century and even earlier (see Reinwald 1938). The excavations were performed (i) to prove the origin of the crater field (Reinwald 1933 [main crater, craters #1 and #4]; Reinwald 1939 [craters #2/8, #5, and #6]), (ii) search for meteoritic iron and investigation of the internal structure (Aaloe 1958 [#5]; Aaloe 1963a [main crater]; Aaloe and Tiirmaa 1982 [#4]; Koval’ 1974 [#2/8]; Pobul 1963 [#3]; Tiirmaa 1994 [#3 and #7]), and (iii) to study a settlement / ceremonial site to the NE of Kaalijärv’s rim by archeological means (Lõugas 1978, 1980).

The age of the Kaali impact is not well constrained (except that it is of postglacial origin), and during the last decade has been the subject of extensive discussion. Materials from three different locations (and different ages) have been regarded to reflect the event and have been studied by radiocarbon dating, AMS, and palynological methods. First, organic matter (pieces of charcoal, wood, peat, gyttja, seeds, needles), pollen, and diatom evidence within the postimpact sediments (at different depths) of Kaalijärv (Kessel 1981; Saarse et al. 1991; Veski et al. 2004) and satellite structures #2/8 and #4 (Aaloe et al. 1963; Saarse et al. 1991; Veski et al. 2004; for locations of structures see Fig. 7) revealed a wide range of ages, but the deepest (and presumably oldest) material was accumulated from about 1750 cal. yr bc. The Kaali site has been inhabited at least since the late Bronze age, which was confirmed by radiocarbon dating of the remains of burnt logs (760–260 cal. yr bc; Lõugas 1978; Veski et al. 2004). Second, glassy spherules in three mires in Saaremaa and one in Hiiumaa (Fig. 1; at a site located approximately 60 km from Kaalijärv) were found in the mid-1990s (Raukas et al. 1995). The peat layer containing these Si-, Ca-, Fe-, and Ni-rich particles was dated to 7500–7600 yr bp (6270–6500 bc cal. yr; see Veski et al. 2004) by the 14C and palynological methods (Raukas 2000b). Reintam et al. (2008) reported microimpactites with a similar age within gleysol approximately 7.5 km south of Kaali and hinted at their possible connection to the Kaali event. No isotopic data for such an association exist to date, and content of PGEs of the spherules or the host layer has not been studied yet either. Consequently, these phases have not been compared with meteorite fragments found at the Kaali crater field or with magnetic material described by Marini et al. (2004). So, their possible link to the Kaalijärv impact still remains a hypothesis. Third, Rasmussen et al. (2000) reported the presence of a distinct Ir-enriched layer in the Piila mire, approximately 8 km from Kaalijärv. The peat containing the enrichment was radiocarbon-dated to 400–370 cal. yr bc. Veski et al. (2001) found this iridium-marker-horizon rich in charcoal, charred wood, wood stumps, and terrigenous material (calcite, anhydrite, and quartz with traces of K-feldspar and plagioclase) and poor in pollen over an area of 5 km2 across the Piila bog. They readjusted the earlier date by Rasmussen et al. (2000) to 800–400 cal. yr bc. Criticism by Raukas et al. (2005) points out that the Ir-enriched layer is close to a geochemical barrier due to replacement of fen peat with bog peat. Also, this date is clearly younger than the date obtained from crater-filling postimpact sediments.

Possible and Suggested Structures


The Ilumetsa crater field in SE Estonia (Figs. 1 and 9, Table 1) consists of two simple structures with rim-to-rim diameters of 75–80 m (Põrguhaud/Hell’s Grave) and about 50 m (Sügavhaud/The Deep Grave) with true depths of about 8 and 3.5 m, respectively. Both structures are surrounded by up to a few meter high rims that are higher in their eastern parts (maximum rim heights are 4.5 and 1.5 m, respectively). The structures were first reported after large-scale regional geological mapping in 1938. Immediately, some excavations were made and an impact origin was proposed by A. Luha, head of the Estonian Geological Committee (Raukas et al. 2001). Unfortunately, results of these early studies remained unpublished. In 1958, two more small peat-filled features were reported (Aaloe 1961) by local people, but nowadays, these are unidentifiable. However, maps and discussions referring to three, four, or even five structures can be found in different publications.

Figure 9.

 Cross-section of the larger Ilumetsa structure (Hell’s Grave, after Aaloe 1979).

Middle Devonian brittle, light-yellow silt- and sandstones, overlain by an up to approximately 2 m thick layer of brown basal till and approximately 0.5 m of glaciofluvial sand, host the Ilumetsa structures. The excavations by Aaloe (Aaloe 1963b; cross-sections were republished by Raukas et al. 2001) showed that the high eastern rims consist of Devonian and Quaternary sands mixed with moraine matter, which contains numerous lenses and nests of boulder clay. Thus, the descriptions by Aaloe (1963b) and a reduced thickness of glacial deposits favor an explosive origin of the Ilumetsa structures and rule out possible glacial explanations (kettle- or sink-holes).

A small lens of gyttja and peat, at maximum approximately 3 m thick, in the center of Hell’s Grave has been dated with 14C and palynology (Liiva et al. 1979). The results from the lowermost silty sand bed with gyttja suggested a 14C age of 6030 ± 100 yr bp (4790–5060 cal. yr bc) and the gyttja, at slightly higher level, yielded an age of 5910 ± 100 14C yr bp (4690–4940 cal. yr bc). This age excludes an anthropological origin, e.g., formation by explosives, for the crater. In the summer of 1996, Raukas (2000b) discovered glassy spherules in a 5.7 m deep peat layer from the Meenikunno bog, approximately 6 km WSW from the Ilumetsa structures. The spherules were reported by Raukas et al. (2001) to be some millimeters in size and interpreted as representing dissipated melt or condensed vapor. The chemical composition of the particles/peat layer containing the spherules has never been studied; thus, a cosmic origin is not proven yet. The peat with spherules yielded 14C ages (see Raukas et al. 2001) of 6542 ± 50 yr bp (7420–7500 cal. yr bp; depth 5.6–5.7 m) and 6697 ± 50 yr bp (7560–7610 cal. yr bp; depth 5.7–5.8 m). Thus, it is approximately 600 yr older than the gyttja from Hell’s Grave, which has been explained by delay of gyttja formation in the crater-like structures (Raukas et al. 2001).

Pobul (1963) reported results of magnetic profiling across the two Ilumetsa structures. Local positive and negative anomalies up to 20 nT were identified, corresponding to the rim and inner parts of the structures, respectively.

Even though a meteorite impact origin of Ilumetsa is likely, as there are no indications for other possible processes, its cosmic origin has not been proven yet. High-precision geochemical (PGE) studies of the soil around the structures, compared with the geochemical signature of the target, may give further hints.


The Tsõõrikmäe structure (Fig. 1; Table 1) was first described by Pirrus and Tiirmaa (1984) after field studies (including leveling, excavations of the rim feature, and drillings) in the late 1970s and early 1980s. They found a relatively wide (5–10 m) but low (up to 2.5 m) rim of 38–40 m diameter surrounding a circular depression filled with a lens of peat up to 4.5 m thick (Fig. 10). Pollen and 14C analyses were carried out on the peat and organic inclusions within the underlying silt. The deepest layers revealed an age of 9320 ± 100 14C yr bp (10,673–10,303 cal. yr bp; Pirrus and Tiirmaa 1984b), thus excluding a human origin of the structure. The structure is located within an approximately 6 m thick layer of reddish-brownish glacial till, which lies on variegated Devonian clays and reddish sandstones. Along with a meteorite impact also a glacial origin (kettle lake) has been proposed (unpublished data referred by Pirrus and Tiirmaa 1984b); however, the local geomorphological situation (low thickness of glacial deposits [base moraine] and no sandur surfaces nearby) does not support this idea.

Figure 10.

 Cross-section of the Tsõõrikmäe structure (after Pirrus and Tiirmaa 1984a).

The area around the structure and the structure’s rim are covered by a continuous, 0.3–1.2 m thick layer of sand that is (glacio)lacustrine or eolian in origin (Pirrus 1995). The sand is reported (Pirrus and Tiirmaa 1984a) to fill small vertical fractures within the top part of the till. Compared to the above-described 14C data for organic inclusions, the presence of a possibly lacustrine sand layer—in comparison to the above 10,673–10,303 cal. yr bp age—hints at a slightly older age for the structure, because the area has been under water during the proglacial lake stage only. The lacustrine conditions took place after the retreat of the Scandinavian Ice Sheet between approximately 14,700 and 14,500 cal. yr bp (Otepää Stage; Kalm 2006) and the period between Baltic Ice Lake stages A2 (approximately 12,800 cal. yr bp) and BI (approximately 12,300–12,100 cal. yr bp; for calibrations see Rosentau et al. 2009). If the structure is of impact origin, it suggests (i) formation in a shallow above just-subsided glacial till near the retreating ice margin, and (ii) that organic sedimentation within the depression started with some delay.


The Simuna structure (Fig. 1; Table 1), 8.5 m wide from rim to rim (whereas the rim is highest [0.3 m] in the west) and 1.9 m deep, was suggested (Pirrus and Tiirmaa 1991) to be the result of a documented (Kipper 1937) meteorite (Virumaa Bolide) fall on 1 June 1937. The bolide was reported (Kipper 1937) to move from ENE to WSW under an azimuth of 259° and to fall to surface at an angle of approximately 60°. Having reached denser air, it exploded at a height of 28 km somewhat to the east from the Viru Roela settlement and Simuna structure. In 1937, the impact was not noticed by the local community, but a fall of a small piece of meteorite was observed approximately 15 km north of the structure. In 1984, a local resident (H. Ross) informed the Meteoritic Commission of the Estonian Academy of Sciences about a depression in the forest close to Simuna village. In 1986, the structure, which has been formed in Quaternary sediments composed of sand and glacial till, was excavated (Fig. 11) and determined not to result from karst or a bomb explosion. A layer of soil under the rim was found (Pirrus and Tiirmaa 1991). Bronsten (1991) analyzed the dimensions of the structure and argued that the initial mass of the bolide was ≥50 tons, and Simuna was formed by the impact of a body with a mass of ≤67 kg. Contrary to this, no traces of the projectile were found during the excavations. No chemical analyses of the soil have been reported to date.

Figure 11.

 Cross-section of the Simuna structure (redrawn after Pirrus and Tiirmaa 1991). No orientation of the section has been indicated in the original publication.


The Lasnamäe site (Fig. 1; Table 1) was discovered by T. Kiipli and A. Põldvere in 1983 (Kiipli and Põldvere 1984; Pirrus 2008). It is a small, approximately 20 × 30 cm wide and approximately 5 cm deep, weathered depression that occurs on a smooth horizontal surface of Middle Ordovician limestones. The cavity is surrounded by a cobweb-like set of fractures, 1.3–2.0 m long, mostly radial and a few tangentially oriented (Fig. 12). The site was interpreted (Kiipli and Põldvere 1984) to have formed by a meteorite impact rather than by human-initiated explosion, because it was originally buried under approximately 4 m of peat and 0.5 m of lacustrine lime and only opened in 1983 for building purposes. No extraterrestrial material was found at the site by visual inspection. The interpreted age of 20,000–25,000 yr bp (Pirrus 2008) is based on the evidence that the last glaciation eroded the structure. Presently, the site is again covered by soil and cannot be directly observed.

Figure 12.

 A sketch of suggested impact-induced fractures on the smooth surface of Ordovician limestone, redrawn from Pirrus (2008). The central feature has been formed due to weathering and does not represent a shock feature.


A meteorite impact origin for the Vaidasoo circular structure 325 m in diameter (59º16′5″N; 25º2′34″E; Fig. 1) was suspected by amateur archeologist P. Böckler in 2004. The structure is marked by a circular raised bog 325 m in diameter surrounded by a wider elliptical transitional mire. In 2006, the structure was drilled and found to be filled with an approximately 10 m thick layer of Pleistocene and Holocene silt, limy lake deposits, and peat. This complex covers fissured and crushed limestones (Suuroja et al. 2006). Based on the drilling results and geophysical data (gravity anomaly of −0.2 mGal), the structure was interpreted (Suuroja et al. 2006) as a possible deeply eroded remnant of a meteorite impact crater. However, ground-penetrating radar studies (Fig. 13) suggest that no circular structure exists in the mineral base below the cover of peat and silt. Thus, Vaidasoo is circular in plan view just because of the raised bog, which often grows in circular patterns; an impact origin is not favored by the author.

Figure 13.

 Ground-penetrating radar image across the Vaidasoo bog. A relative dielectric permittivity value of 70 has been considered while calculating the time-scale to depth-scale. Please note the vertical scale is exaggerated. Shape of the reflection from the top surface of the mineral base (highlighted by black arrows) does not support an impact origin of Vaidasoo. The white arrows locate reflections from high-voltage power lines.

Breccia Layers

Two breccia layers, Narva and Osmussaar (for distributions see Fig. 1), within the subhorizontal Paleozoic sedimentary succession have attracted researchers and have been linked to impact events. The Narva Breccias belong to a mixed carbonate-siliciclastic sequence of Eiphelian age (Middle Devonian). The breccias are distributed across the ancient shallow epeiric Baltic Basin, occupying an area of about 160,000 km2 in the Baltic countries and adjacent areas of NW Russia and Belarus (Tänavsuu-Milkeviciene et al. 2008). Among other explanations for their origin (a slump caused by paleoseismic activity [Karajajute-Talimaa and Narbutas 1964]; brecciation by early diagenetic sulfate crystallization [Paškevičius 1997]), a probable impact origin was proposed by Masaitis (2002). He suggested that the Narva Breccias may represent a possible tsunami-caused subaqueous slumped and brecciated deposit from the Kaluga impact, which struck a shallow (300–500 m deep) sea of the East European platform at 380 Ma. Also, Masaitis (2002) hinted at approximately 0.1 mm sized angular quartz grains with two sets of PDF found in a drill-core sample from NE Lithuania, but details of this find have not been published so far. A sedimentological study by Tänavsuu-Milkeviciene et al. (2008) states that the Narva Breccias occur as individual 0.2–2.7 m thick beds. The brecciated successions were found, 0.6–20 m thick, at 11 stratigraphic positions with a maximum thickness of 40.8 m. Also, it was noted that the relatively small 15 km sized Kaluga event was evidently too small for direct or tsunami-caused brecciation of the entire basin and was approximately 10 Ma younger than the Narva Breccias. Tänavsuu-Milkeviciene et al. (2008) concluded that the breccias were formed in a supratidal environment because of evaporite dissolution and/or wave action. No detailed mineralogical or geochemical data on the Narva Breccias have been published so far.

Brecciated in situ lime- and sandstones, together with 2–3 m thick carbonate-cemented sandstone clastic dikes, in NW Estonia are called Osmussaar Breccias. The brecciated layer belongs to the Volkhov and Kunda regional stages (Dapingian—early Darriwilian) and is unconformably overlain by normal horizontally bedded and undisturbed bioclastic Darriwilian limestones (Aseri, Lasnamägi, and Uhaku regional stages). Laterally, the carbonate lithoclasts in the brecciated layer vary in size from tens of meters to tens of centimeters (Suuroja et al. 1999). Matrix of the breccia, i.e., of the clastic dikes, is composed of fine- to medium-grained well-rounded quartz grains with some feldspar and clay minerals, all embedded in a fine-grained matrix of recrystallized calcite (Alwmark et al. 2010). The thickest (1.0–1.5 m) occurrences of the brecciated layer are found on Osmussaar Island, but a thinner (few tens of cm, approximately 70 km east of Osmussaar) and less fractured layer occurs in drill-core sections in an area of more than 1000 km2 along the NW coast of Estonia (Fig. 1) (Suuroja et al. 2003; Alwmark et al. 2010).

The origin of these approximately 466 Ma Osmussaar Breccias was first interpreted (Öpik 1927) as a result of earthquake-triggered seafloor destruction. Puura and Tuuling (1988) suggested that they represented karst-modified desiccation and/or earthquake cracks formed during several episodes. After discovering the nearby Neugrund crater, Suuroja and Saadre (1995) linked the brecciation to an impact event at approximately 475 Ma. Later, after revealing the early Cambrian age for Neugrund (Suuroja and Suuroja 2000), this link was rejected (Suuroja et al. 2003). However, after finding quartz grains with PDF in the matrix of the breccias, the same authors proposed a possibility that the material is derived from the Neugrund crater, but was reworked and sedimented into the earthquake-fractured limestone layer by hydraulic injection of fluidized material.

In 2010, Alwmark et al. reported that the Osmussaar Breccias contain abundant (>13 grains kg−1; 63–130 μm in diameter) angular chromite grains of extraterrestrial origin along with well-rounded chromium-rich spinels of the same size. Chemical composition of chromite was found almost identical to that found at the Thorsberg quarry at Kinnekulle in Sweden (Schmitz et al. 2001; Schmitz and Häggström 2006), in various mid-Ordovician limestones of the region, and resurge deposits from the Lockne impact (Alwmark and Schmitz 2007). Thus, it was concluded that the impactor responsible for the Osmussaar Breccias was an ordinary chondrite of L-type, which, considering the age of the breccias, links them with the L-chondrite breakup at 470 ± 6 Ma (Korochantseva et al. 2007). There is no source crater positively tied to the Osmussaar Breccias, although there are unpublished suggestions that the breccias represent tsunamiites due to impact in a SW direction from Osmussaar Island. The tsunami origin was further supported by Tinn et al. (2010), who studied ostracod assemblages in the sandstone dikes. They found the composition varied and was unusual, as originated from different microhabitats. They explained their results by rapid mixing and displacement of bottom sediments.


At present, there are three confirmed impact structures in Estonia making it the most cratered country per area (including aquatic area, Estonia amounts to approximately 0.015% of Earth) in the world. The number of impact structures is especially striking when comparing it with wide almost “crater-free” areas in Asia, South America, or central Africa. There are five principal reasons for this bias. First, shallow epicontinental seas covered the area from the Ediacaran to the Devonian, causing immediate burial and good preservation of the freshly formed craters. This period was succeeded by a prolonged (1.6–0.6 Ga) epoch of erosion and peneplanization of the Svecofennian (1.9–1.6 Ga) orogenic crust. Since the Devonian, erosion has dominated, but it has not been extensive, as the present sedimentary section has never been buried deeply (Kirsimäe et al. 1999) nor has it been tectonized. Second, the low thickness of the Quaternary sediments in northern and western Estonia allowed even small-sized explosions to penetrate the layers in early Paleozoic sedimentary beds (predominantly lime- and dolostones, but also silt- and sandstones), thus, inducing their better preservation. Third, there was an increased terrestrial meteorite influx in the Middle and Late Ordovician following the L-chondrite asteroid breakup at 470 ± 6 Ma (Korochantseva et al. 2007). Fourth, a relatively simple geological build-up (Fig. 1) of the territory and wide deposits of Early Paleozoic mineral resources (kukersite oil shale, Dictyonema argillite, and phosphorite) together with economic interest and intensive drilling programs produced a high level of geological knowledge of this region. Finally, the existence of the well-exposed Kaalijärv crater field and early (Reinwald 1938) success in proving its impact origin helped increase people’s awareness and attract researchers to the subject.

Acknowledgments—  This study was supported by the Estonian Target Financing project SF0180069s08. I thank Wolf Uwe Reimold for initiating the overview and editorial handling. Reviews provided by Henning Dypvik and Lauri J. Pesonen were most helpful in pointing out areas where the original manuscript needed revision. Help by Sten Suuroja and Agnieszka P. Baier to improve the manuscript is acknowledged.

Editorial Handling—  Dr. Uwe Reimold