Scientific drilling conducted at the inner slope of the Miocene central Ries impact crater recovered a partial section of crater lake sediments. Four sequences were recovered, composed of suevite-derived sandstones, thin lignite seams, bituminous shales, and marlstones to claystones. These flooding-evaporation sequences reflect the impact of short-term climatic fluctuations on a hydrologically closed basin. The superimposed trend from sequences rich in bituminous shales in the lower parts of the section to sequences dominated by organic-poor claystones and intercalated lignites in the upper parts of the section resembles that of the 300-m-thick central crater basin succession, which has previously been considered to reflect a climate-controlled development from an alkaline saline lake to a freshwater lake with temporary coal swamps. In the sediment core of Enkingen, however, the change from bituminous shales to organic-poor claystones with intercalated lignites is associated with a general increase in salinity, as indicated by (1) palynomorphs, (2) increase in δ13C of the lipid biomarker archaeol (bis-O-phytanylglycerol), and (3) the occurrence of 13C-enriched C20/C25-archaeol (O-phytanyl-O-sesterterpanylglycerol) specific to halophilic Archaea. In addition, the unidirectional trend in 87Sr/86Sr of carbonates, declining from ratios of Variscan basement rocks toward marine ratios, indicates a change from (1) weathering of crystalline rocks and suevite to (2) ejected Jurassic sediments (Bunte Breccia) in the catchment area as the major source of ion influx to the lake. From that trend, a change in lake water composition and a general increase in ion concentrations are inferred. These new results can be applied to a reassessment of major parts of the lacustrine succession of the Ries crater. We use these data to propose a new hypothetical model for the chemical and ecological evolution of the Ries crater lake: (1) After the establishment of a stratified brackish eutrophic soda lake due to silicate weathering and evaporation, the increasing influx of waters from the Bunte Breccia carbonate and authigenic silicate precipitation led to a mesotrophic halite lake with marine-like ion ratios and concentrations. (2) Further increase in ions, among them Mg2+ and Sr2+, resulted in hypersaline conditions with gypsum precipitation, low primary production, and phreatic Sr-rich dolomitization in marginal carbonates. (3) The final, sudden change to oligotrophic freshwater conditions is explained by the formation of an outlet late in the lake history. We conclude that the chemical and ecological evolution of the Ries lake therefore appears to have been mainly controlled by the weathering history of the catchment area, with climate fluctuations causing superimposed cycles. Similarly, changes in terrestrial palynomorph associations may at least partly reflect a change in soil types in the catchment area, from fertile, moist soils on suevite to dry karst soils and soils on Bunte Breccia. These interpretations imply that the initial suevite blanket of the Ries crater was much more continuous and widespread than previously assumed.
Impact crater lakes potentially form valuable climatologic archives, because these suddenly formed “sediment traps” accurately record climatic variations during postimpact times, given a sufficient and continuous sedimentation rate (e.g., Koeberl et al., 2007; Shanahan et al., 2009; Melles et al., 2011). Moreover, since the discovery of a number of possible former crater lakes on Mars (Forsythe, 1990; Cabrol and Grin, 1999; Fassett and Head, 2008a, 2008b), lakes in terrestrial impact craters are of special interest as potential analogs with respect to sedimentary, hydrochemical, and paleobiological evolution.
The 15-m.y.-old Ries crater is one of the most intensively studied impact craters on Earth (see, e.g., Pohl et al., 1977; von Engelhardt, 1990; von Engelhardt et al., 1995; Kring, 2005; Wünnemann et al., 2005). The sediment core of the Nördlingen 1973 research drilling project provided an excellent basis for deciphering the crater lake history, represented by a succession of basal siliciclastics, bituminous laminites, marlstones, and finally claystones with lignite seams. This succession was interpreted as reflecting a change from semiarid to humid climate (see summary in Füchtbauer et al., 1977). Nonetheless, several open questions persist:
(1) Due to a lack of marker beds, the correlation with marginal carbonates remains controversial (Füchtbauer et al., 1977; Jankowski, 1981; Schauderna, 1983; Arp, 1995). Basinal and marginal successions are both reported to show a trend from supposed moderately saline and alkaline conditions to freshwater conditions, possibly reflecting increasing humidity (Jankowski, 1981; Arp, 1995). It is as yet unclear whether both trends are time-correlated, or if they represent two successive climatic fluctuations.
(2) In a hydrologically closed basin, the climate change inferred to explain decreasing salinities implied a significant lake-level rise, for which, however, no evidence exists. Thus, the suggested decrease in salinity of the Ries crater lake is difficult to explain without assuming subbottom leakage or the formation of an outlet.
(3) The relation of intrinsic (e.g., weathering history) versus extrinsic (climate) factors controlling sedimentary history remains ambiguous, although some attempts have been made to address this question (e.g., Jankowski, 1981).
(4) The chemical evolution of the lake water within this hydrologically closed crater basin, a potential analog for extraterrestrial crater sedimentary systems, remains not well understood. Even the precise time span of the lake’s existence (0.3–2.0 Ma; Pohl, 1977; Jankowski, 1981) is poorly constrained.
In 2006, the Geological Survey of the Bavarian Environment Agency conducted a research drilling project at Enkingen to target a magnetic anomaly in the central crater near the southeastern sector of the Inner Ring (Pohl et al., 2008, 2010; Reimold et al., 2010, 2012). Prior to penetrating through suevite (i.e., a polymict impact breccia with a particulate matrix containing lithic and mineral clasts in all stages of shock metamorphism including cogenetic melt particles that are in a glassy or crystallized state; Stöffler and Grieve, 2007) and impact melt agglomerates, the drilling recovered a partial section of the Ries crater lake sediments, providing an excellent opportunity to reinvestigate previous interpretations of the lake history in view of these outstanding questions. Together with supplementary data from marginal carbonates, the Enkingen core data suggest a new model for the chemical and ecological evolution of major parts of the Ries crater lake.
STUDY AREA AND GEOLOGICAL CONTEXT
The Ries crater is located in southern Germany, 120 km NW of Münich, and it separates the Upper Jurassic limestone karst plateaus of the Franconian and Swabian Alb (Fig. 1). The crater was formed by an asteroid impact during the middle Miocene ∼15 m.y. ago (Gentner and Wagner, 1969; Staudacher et al., 1982; Abdul Aziz et al., 2008; for differing age determinations, see Buchner et al., 2010). The impactor, an ∼1-km-sized asteroid of unknown composition (Schmidt and Pernicka, 1994), penetrated ∼580 m of Triassic–Jurassic and Tertiary sedimentary cover (Bolten and Müller, 1969; Haunschild, 1969; Schmidt-Kaler, 1969), before melting and exploding in crystalline rocks of the Variscan basement (Stöffler, 1977; von Engelhardt, 1990). Gravitational collapse of the initially 2.0–2.8-km-deep transient crater cavity (Stöffler, 1977; von Engelhardt et al., 1995) resulted in a complex basin, 24 km in diameter and 600 m deep, including an inner “crystalline” ring and a megablock zone (Pohl et al., 1977; Stöffler, 1977) (Figs. 1 and 2). The ejecta blanket, the so-called “Bunte Breccia,” extends up to 26 km outward from the crater rim (Hüttner, 1969; von Engelhardt, 1990), with the “fall-out suevite” as patchy remnants upon it (von Engelhardt et al., 1969; Gall et al., 1975). Within the central crater, a “crater suevite” formed by collapse of the vapor plume and ground surging (Pohl et al., 1977; Stöffler et al., 1977, 2013; Artemieva et al., 2013). As a consequence of the gravitational collapse, a closed basin with high diameter-depth ratio formed, resting on a highly brecciated crystalline basement and surrounded by a suevite- and ejecta-covered Upper Jurassic limestone plateau of the Franconian and Swabian Alb.
Lake deposits comprise three lithofacies units: (1) a basinal argillaceous series, (2) mar-ginal carbonates, and (3) marginal fluvio-deltaic siliciclastics. The basinal argillaceous series was transected by the Nördlingen 1973 research drilling (427 m above sea level [a.s.l.]) located in the western part of the central crater basin (Figs. 1 and 2). The 314-m-long sediment core covers large parts of the lake history and reveals a sedimentary succession consisting of a basal member, a laminite member, a marlstone member, and a claystone member with lignite seams (Füchtbauer et al., 1977; Jankowski, 1981). This succession has been interpreted as the transitional development from a playa lake (unit A: basal member) to an alkali-saline lake (unit B: laminite member), which finally decreased in salinity (unit C: marlstone member) until reaching freshwater conditions (unit D: claystone member) (Füchtbauer et al., 1977; Jankowski, 1977, 1981; Rothe and Hoefs, 1977; Dehm et al., 1977).
The youngest lake sediments, which probably filled the crater basin completely, were removed in the central crater by Pleistocene fluvial erosion after the re-establishment of river systems crosscutting the Ries area (Schröder and Dehm, 1950; Bolten, 1977; Bader and Schmidt-Kaler, 1990). A potential equivalent of this member, however, is preserved in marginal carbonates and siliciclastics now exposed at the crater rim (430–530 m a.s.l.). Here, calcitic spring mounds, dolomitic algal bioherms, carbonate sands, palustrine limestones, siliciclastic sands, conglomerates, and—at the very top—charophyte-bearing limestones occur. The latter are considered the final freshwater deposits of the lake, before silting up or outlet formation (Bolten, 1977).
The Enkingen (SUBO 18) drilling site is located at the southern periphery of the central crater, at the inner slope of the inner ring (Figs. 1 and 2), where the basinal argillaceous series forms major parts of the Ries lake deposits. The drilling was carried out from 29 November to 8 December 2006, with a core diameter of 10 cm and almost 100% core recovery (Pohl et al., 2008, 2010; Reimold et al., 2010, 2012).
MATERIAL AND METHODS
Cores were described using sedimentologic and lithologic parameters. Petrographic thin sections 28 × 48 mm in size were prepared from 10 Enkingen samples embedded in Araldite 2020 resin. In addition, 14 large thin sections (7.5 × 10 cm) were prepared from samples of basin margin carbonates of Nähermemmingen and Erbisberg (Fig. 1). These carbonates belong to the topographically lowest (i.e., oldest) accessible surface occurrences of lake-margin sediments, represent potential time equivalents of the Enkingen sediments, and have been analyzed for 87Sr/86Sr ratios for comparison purposes.
The mineralogical composition of 12 Enkingen core samples was analyzed by powder X-ray diffraction using a Philips PW 1800 diffractometer operating at 45 kV and 40 mA with monochromated Cu Kα radiation. The range 4–70 °2θ was scanned with a step width of 0.02 °2θ. The counting time was 3 s per step. Mineral identification was carried out using the X’Pert HighScore Plus software (PANalytical).
The elemental composition of the samples was determined by X-ray fluorescence analysis (PANalytical Axios Advanced XRF spectrometer fitted with a 4 kW Rh anode SSTmAx X-ray tube) on glass fusion disks. Samples with high S and Corg contents were previously heated at 650 °C for 3 h in a furnace.
Total carbon (Ctot), organic carbon (Corg), and carbonate carbon (Ccarb) were determined using a Euro EA CNS analyzer (HEKAtech GmbH, Germany). Corg was measured after decarbonatization with 2 N HCl. Ccarb was calculated as the difference of Ctot and Corg.
Samples for carbon and oxygen stable isotope measurements were obtained under a binocular microscope from cut core slabs and hand specimens using a steel needle to sample separate textures. Carbonate powders were reacted with 100% phosphoric acid (density >1.95) at 70 °C using a Thermo Kiel IV carbonate preparation line connected to a Finnigan Delta plus mass spectrometer. All values are reported in per mil relative to Vienna Peedee belemnite (VPDB) by assigning a δ13C value of +1.95‰ and a δ18O value of −2.20‰ to NBS 19 standard. Oxygen isotopic compositions of dolomite were corrected using the fractionation factor given by Rosenbaum and Sheppard (1986), i.e., 1.00993 at 70 °C. Reproducibility was checked by replicate analysis of laboratory standards and is better than ±0.05‰ (1σ).
Twenty-five carbonate fractions of carbonate and calcareous siliciclastic rock samples were analyzed for 87Sr/86Sr isotopic composition. All analyses were carried out on a Thermo-Finnigan Triton© thermal ionization mass spectrometer (TIMS). Prior to digestion, all samples were mixed with a tracer solution enriched in 87Rb-84Sr. Concentrations were calculated using the isotope dilution technique. Rb and Sr were separated from one single rock digest using standard cation-exchange procedures. Reproducibility for NBS SRM 987 (n = 5) was 0.71025 ± 0.00008 and 0.05650 ± 0.00004 for 87Sr/86Sr and 84Sr/86Sr, respectively. The analytical mass bias was corrected with 88Sr/86Sr of 0.1194 using an exponential law. Analytical mass bias correction for Rb measurements was achieved via repeated analyses of NBS SRM 984 yielding an 85Rb/87Rbraw of 2.6013 ± 0.0002 (n = 5), resulting in an exponential mass bias of 2.45‰/u. All sample measurements were performed under the same conditions and corrected with an exponential mass bias derived from the standard measurements. Model calculations to test possible hydrochemical developments of the Ries crater lake were performed using the program PHREEQC (Parkhurst and Appelo, 1999). For this purpose, analysis data of present-day surface and groundwaters of the Ries basin (Winkler, 1972) were equilibrated to atmospheric pCO2, modified by subtraction of water, mixing, and limit values for mineral saturation states, to test if and how the chemical composition of potential Ries crater lake analogs can be achieved. Details of the calculations are provided in the Appendix.
Biomarker and compound-specific C isotope methods were as follows. Samples (3–5 g) were crushed and powdered using a pebble mill (Retsch MM 301). Ground samples were subsequently extracted with distilled dichloromethane (DCM)/methanol (3/1), DCM, and n-hexane using ultrasonication. Extracts were combined and transesterified using trimethylchlorosilane (TMCS)/methanol (1/9; v/v) for 1 h at 80 °C. Biomarkers were then extracted using n-hexane and DCM and subsequently separated by column chromatography (silica gel 60). Hydrocarbons (F1) were eluted with 3 dead volumes (dv) of n-hexane, fatty acid methyl esters (FAME) with 5 dv DCM, and alcohols with 5 dv DCM/acetic acid ethyl ester (5/1; v/v). Before analyses by gas chromatography–mass spectrometry (GC-MS; see following), elemental sulfur in F1 was removed by addition of activated elemental copper. Alcohols in F3 were silylated with BSTFA (N,O-bis[trimethylsilyl]trifluoroacetamide) for 1 h at 80 °C before GC-MS analysis.
All fractions were analyzed by combined GC-MS using a Varian CP-3800 gas chromatograph coupled to a Varian 1200L mass spectrometer. The GC was equipped with a fused silica capillary column (Phenomenex Zebron ZB-5MS, 30 m, 0.25 µm film thickness, 0.32 mm inner diameter) and an on-column injector. Helium was used as the carrier gas at 1.5 mL/min. The temperature program was 80 °C (for 5 min) and 6 °C/min to 310 °C (held 30 min). The MS source was operated at 200 °C at 70 eV ionization energy. The δ13C values of biomarker were analyzed using a Thermo Scientific Trace GC coupled to a Delta Plus isotope-ratio MS. Details can be found elsewhere (Blumenberg et al., 2012).
All samples for palynomorph analysis were treated with HCL, HF, and KOH following the standard preparation method described by Kaiser and Ashraf (1974). To remove flocculating organic matter and to improve transparency of individual palynomorphs, the residues were slightly oxidized by applying H2O2. In addition, residues were sieved with a mesh size of 10 µm. In total, 7 samples were analyzed, with 250–515 palynomorph specimens counted per sample. The taxonomy was based on Thomson and Pflug (1953), Thiele-Pfeiffer (1980), and Stuchlik (1994).
Lithofacies Types and Sedimentary Succession
The drilling penetrated 4.5 m of Quaternary gravel, sands, and silts of Eger Creek. Below that, i.e., from 4.5 m to ∼23 m, largely siliciclastic sediments of the Ries crater lake were recovered (Fig. 3). These are described in detail in Table DR1.1 From 23 m to 100 m depth, gray suevite and reddish-brown impact melt agglomerates were penetrated; these have been described and investigated in detail by Pohl et al. (2010) and Reimold et al. (2012). The recovered Ries crater lake sediments can be subdivided into three general lithologic units: (1) a basal facies association characterized by sandstones (23.0–17.83 m depth in core), (2) a bituminous shale facies association (17.83–10.65 m), and (3) a claystone facies association (10.65–4.5 m) (Fig. 3).
Seven principal lithofacies types can be distinguished (Table 1). Their mineralogical and chemical compositions are shown in Table 2. Images and micrographs are provided in Figure DR1 (see footnote 1). These lithofacies types are arranged in sequences (Fig. 3), which start with coarse siliciclastic sediments with intercalated carbonates or lignites (facies 1, 2, 7), grade into laminated fine-grained sediments (facies 3–5), and end with claystones with poorly visible stratification (facies 6).
The first sequence (23.0–18.5 m) starts with a clay-rich conglomeratic sandstone (facies 1), distinguished from the suevite below by the inclusion of gastropod shell fragments. The following argillaceous Hydrobia limestone (facies 2) is overlain by a conglomeratic sandstone (facies 1), again derived from suevite. Then, olive-gray laminated claystones (facies 5) follow, with thin intercalated beds of coarse quartz grains and white-gray micrites.
The second sequence (18.5–15.0 m) starts with marly conglomeratic sandstones (facies 1) with thin carbonaceous clay intercalations (facies 7), followed by a thin arenaceous limestone bed with Hydrobia (facies 2), overlain by bituminous shales (facies 4), which grade into laminated claystones (facies 5).
The third sequence (15.0–9.0 m) begins with a thin carbonate-cemented sandstone bed (facies 2). Above that, bituminous shales (facies 4) with rare layers of laminated dolomite (facies 3) form major parts of this sequence. At the top, laminated (facies 5) and stratified claystones (facies 6), the latter with slumping structures, complete the sequence.
The fourth sequence (9.0–4.5 m) begins with carbonaceous clay beds (facies 7) and bituminous shales (facies 4), which grade into laminated (facies 5) and stratified claystones (facies 6). A coarse siliciclastic basal layer is not developed, and the top parts of the sequence are cut off by Quaternary gravels.
Stable Carbon and Oxygen Isotopes of Carbonates
Stable carbon (δ13C) and oxygen (δ18O) isotope ratios were measured for microcrystalline carbonates. Calcite cements (spar) and two gastropod shell fragments were analyzed for comparison (Fig. 4; Table 3). The values obtained show five distinct groups corresponding to their lithostratigraphic position:
(1) The micrite matrix of argillaceous Hydrobia limestones (facies 2) shows values between −3.4‰ and −2.7‰ for δ13C and −5.2‰ and −3.0‰ for δ18O. These values are much higher than those of freshwater limestones (e.g., Platt, 1992), but still below marine values (e.g., Veizer et al., 1999).
(2) The late meteoric calcite spar in gastropod shells of the argillaceous Hydrobia limestones (facies 2) exhibits the most negative δ18O values (around −8.0‰) at δ13C values between −6.2‰ and −3.9‰. These values are interpreted to reflect meteoric freshwater conditions.
(3) The isotope data of calcitic micrite laminae of laminated and stratified claystones (facies 5 and 6) vary between moderately negative (δ13C: −4.0‰; δ18O: −2.7‰) and moderately positive ratios (δ13C: +3.3‰; δ18O: +2.1‰), exhibiting a clear positive covariation for δ13C and δ18O (r = 0.95). This covariation reflects closed basin conditions (Talbot, 1990). Strong variations within single samples point to sample mixing with late meteoric spar.
(4) The dolomicrite laminae of bituminous shales (facies 4) plot in a field distant from all other values: δ18O shows positive values between +2.9‰ and +4.2‰, while δ13C is characterized by negative values between −2.7‰ and −5.3‰. A weak negative covariation for δ13C and δ18O (r = −0.66) is recognizable.
(5) The two analyzed gastropod shell fragments with moderately negative values for δ13C and δ18O plot between matrix values of argillaceous Hydrobia limestones and late meteoric spar cement, and thus may be diagenetically altered.
Carbonate lithofacies types (facies 2, 3) as well as the carbonate fraction of siliciclastic lithofacies types (facies 1, 4–7) were analyzed for 87Sr/86Sr ratios (Fig. 5; Table 4). Generally, the values cluster between marine ratios of Upper Jurassic limestones (0.70771–0.70905; Pache et al., 2001) and “crystalline” ratios of the Variscan basement rocks of the Ries area (mean: 0.7134; Schnetzler et al., 1969; Horn et al., 1985).
Within this range, argillaceous Hydrobia limestones (facies 2) as well as laminated dolomites (facies 3) show comparatively low values between 0.71125 and 0.71185. Among the siliciclastic lithofacies types, the calcitic matrix of clayey polymictic conglomeratic sandstone (facies 1) reveals an intermediate ratio of 0.71303, while ratios of the micritic portion of stratified and laminated claystones (facies 5, 6) have higher values (0.71277–0.71463). Highest 87Sr/86Sr ratios up to 0.71576 are from the dolomicritic portion of the bituminous shales.
In general, the 87Sr/86Sr ratios show a positive correlation with the Rb concentrations of the carbonate fraction, and with whole-rock Al2O3 + SiO2 (Figs. 5A and 5B). In turn, 87Sr/86Sr ratios show a negative correlation with carbonate contents and Sr concentrations (correlation coefficients: −0.79 and −0.91, respectively; graphs not shown). There is no depth correlation, except for the carbonate beds, which show a successive decrease in values from 0.71185 at 21.10 m depth to 0.71125 at 12.01 m depth (Fig. 3).
Biomarkers and Compound-Specific C Isotopes
Selected samples from different lithofacies types were studied for relative abundances and δ13C values of biomarkers (Fig. 6; Table DR2 [see footnote 1]). All samples contain sequence-specific as well as unspecific biomarkers.
Long-chain alcohols and fatty acids with a strong even-over-odd predominance were found in all samples. These compounds have previously been reported from comparable sulfur-rich shales from the Ries area (Barakat and Rullkötter, 1994), and most likely originate from the leaf wax constituents of higher plants. Stable isotope signatures of n-octacosanoic acid (−27‰ to −29‰) suggest an origin from C3 plants (Eglinton and Eglinton, 2008; Table DR2 [see footnote 1]).
Here, 17β,21β-bishomo-hopanoic acid, a compound most likely derived from heterotrophic soil bacteria (Pancost and Sinninghe Damsté, 2003), was detected in stratified claystones (at 5.47 m; facies 6), carbonaceous clay (at 8.46 m; facies 7), and bituminous shales (at 14.32 and 17.60 m; facies 4). The predominance of the biological (17β,21β) versus the geological (17α,21β) isomer indicated a low thermal maturity for the organic matter and was also observed in previously studied shales from the Ries area (Barakat and Rullkötter, 1994, 1995). Similarly, iso-branched dialkyl glycerol diethers with 15 carbon atoms (not shown), which are common in soil bacteria (Oppermann et al., 2010; Sinninghe Damsté et al., 2011), also occurred in these samples.
Occurrences of thiophenes and tocopherols (mainly α-tocopherol) were restricted to bituminous shale samples (facies 4). Both compound classes have been reported from shales of the Ries crater lake and suggested as paleosalinity indicators (Barakat and Rullkötter, 1997). Tocopherols are widespread in photoautotrophic organisms such as cyanobacteria, algae, and higher plants (Peters et al., 2004). Thiophenes were diagenetically formed at high sulfide concentrations causing sulfurization of unsaturated organic compounds. The observed structural similarities with the tocopherols indicate an origin from photoautotrophic organisms. Similar to thiophenes and tocopherols, dinosterol was found only in bituminous shales.
Long-chain unsaturated methyl and ethyl ketones (alkenones; C37 and C38 plus minor C39) are abundant in the bituminous shales (facies 4), whereas minor concentrations occurred in Hydrobia-bearing limestone (at 18.22 m; facies 2), and traces in carbonaceous clay (at 8.46 m; facies 7). These alkenones are indicative of haptophytic algae. The observed C37 to C39 homologs are abundant in marine environments but have also been reported from freshwater settings (Liu et al., 2011). In lakes, the presence of alkenones appears to be associated with brackish to saline conditions (3–50 g/L) (Liu et al., 2011, and references therein). Long-chain alkenones have as yet not been reported from the Ries area, but high abundances of plausible diagenetic alkenone derivatives, C37 and C38 midchain thiolanes, have been observed in comparable Ries shales (Barakat and Rullkötter, 1999) and may be taken as additional support for the widespread occurrence of haptophyte algae in this environment.
Archaeol, indicative of Archaea, occurs in all samples investigated, with the highest concentrations in a bituminous shale at 14.32 m (facies 4) and in the topmost samples (facies 6). The δ13C values of archaeol (Fig. 6) revealed a significant increase from −27‰ in geologically older sections (Hydrobia-bearing limestone facies 2, bituminous shale facies 4) to −23‰ in younger samples (stratified claystones facies 6, carbonaceous clay facies 7). In addition, a C20/C25-archaeol with a particularly high δ13C value of −21.1‰ was found in the carbonaceous clay sample at 8.46 m (facies 7; Fig. DR2 [see footnote 1]).
Seven samples representing different lithofacies types in the lacustrine succession were studied for palynomorphs (Table 5). All samples contained very well-preserved terrestrial and aquatic palynomorphs. In total, 77 palynomorph taxa representing 52 botanical families or genera were identified (Table 5). Only two taxa could not be assigned to present-day organisms. The phytoplankton included acritarchs like Sigmopollis, coccoid green algae such as Botryococcus, and mass occurrences of other coccal green algae, possibly belonging to the orders of Volvocales or Chlorococcales (Figs. 7A and 7B).
The terrestrial palynomorphs were assigned to “arcto-tertiary” and “paleotropic” elements (Kolcon and Sachsenhofer, 1999). Based on that, five terrestrial plant associations (modified from Kohlman-Adamska, 1993; Stuchlik, 1994; Kolcon and Sachsenhofer, 1999; Larsson et al., 2006) and one algal association were distinguished (Fig. 8):
(1) The swamp forest association (SFA) was composed of 6 taxa, and with 4%–19% abundance, it formed a significant part of the flora. Taxodiaceae (bald cypress, Fig. 7C) and Cupressaceae (cypress), the fossil pollens of which are not distinguishable, dominated the association, while accessory elements were small trees and shrubs like Alnus (alder), Betula (birch), Salix (willow), Myricaceae (bayberry, Fig. 7D) as well as Cercidiphyllum (Katsura). Alnus, Salix, and the Myricaceae showed a maximum between 17.60 and 12.01 m, i.e., in laminated dolomites (facies 3) and bituminous shales (facies 4).
(2) The temperate “mixed mesophytic” forest canopy association (TFA) consisted of 13 different taxa, in total with abundances between 14% and 37%. The major part of the pollen was derived from conifers, in particular from Pinaceae (pines, Fig. 7E). Solely for Pinus (pine), four palynomorph species were found, while Picea (spruce) and Abies (fir) were of subordinate importance. Angiosperm trees and shrubs were represented by Carpinus (hornbeam), Carya (hickory), Pterocarya (wingnut), Celtis (hackberry), Fagus (beech), Ostrya (hop-hornbeam), Quercus (oak), Ulmus (elm, Fig. 7F), or Tilia (lime tree). These arcto-tertiary elements, however, were represented in low numbers, and the diverse Pinus and Picea pollen clearly dominated this association. The TFA showed a comparatively high abundance of 31% in laminated claystones (facies 5) at 19.50 m, followed by a decrease to 14% in the bituminous shales (facies 4) and their dolomite intercalations (facies 3). Finally, the abundance increased again in stratified claystones (facies 6), with a maximum of 37% in a carbonaceous clay (facies 7) at 8.46 m depth.
(3) The warm temperate plant association (WTPA) included 18 taxa with preference for warm and humid climatic conditions (“paleotropic flora”), among them the Fagaceae (beech family) with their subfamily Castaneoideae, and the Juglandaceae (walnut family) with Engelhardia (Fig. 7H) and Platycarya. The WTPA forms 30%–50% of the terrestrial palynomorphs in the samples, with maximum values in bituminous shales (facies 4) and minimum values in the stratified claystones (facies 6) at the top of the section.
(4) The halophilic-xerophilic association was composed of two species. The halophilic Chenopodiaceae/Amaranthaceae (goosefoot family; Fig. 7I) consistently occurred, with maxima at 18.22 m depth (facies 2) and in upper parts of the section (8.56 m depth; facies 6). Similarly, Ephedra (Fig. 7J), a xerophilic gymnosperm shrub, occurred at 18.22 m (facies 2) and 5.47 m (facies 6), while bituminous shales (facies 4) were almost devoid of them.
(5) The mesophytic understory plant association (MUPA) consisted of 13 taxonomic groups, among them the Pteridophyta (ferns), Poaceae (true grasses), Asteraceae, Ericaceae, Cyperaceae, Artemisia, Labiatae, and rare Bryophyta. In addition, Sparganiaceae (cattail), widespread in present-day freshwater marshlands, occurred in significant numbers in the bituminous shale (facies 4) sample at 17.60 m depth. A characteristic of the MUPA was a dominance of the Pteridophyta, with 11 different spore genera, and the Poaceae. In total, the MUPA formed only a minor component (2.5% to 9.1%) of the terrestrial palynomorphs in the samples.
(6) The algal association was represented by coccoid and coccal green algae and acritarchs with a very distinct distribution pattern. Samples from the lower parts of the section, i.e., between 19.50 (facies 5) and 14.25 m (facies 3), contained the coccoid green alga Botryococcus and Sigmopollis sp., and small (<20 µm) spherical nonpollen palynomorphs (Fig. 7K), which may be planktonic algal spores (Batten, 1996). Sigmopollis sp. was common in marine, brackish, and freshwater environments throughout the Neogene and Quaternary (e.g., Head, 1993; Hannah et al., 2000; Mudie et al., 2010). In contrast, samples from the upper part of the core, i.e., carbonaceous claystone (facies 7) and stratified claystones (facies 6) at 8.46 m and 5.47 m, respectively, exhibited a mass occurrence of small green algae (Figs. 7A, 7B, and 8) that were possibly remnants of unicellular flagellate green algae of the orders Volvocales or Chlorococcales. A similar algal genus of the family of Chlamydomonadaceae, the rarely preserved genus Phacotus, has already been reported from the Ries by Müller and Oti (1981) who also noted increased abundances of Botryococcus, as observed in this study.
INTERPRETATION: SEQUENCES, SALINITY INCREASE, AND CATCHMENT AREA
The sedimentary succession encountered in the Enkingen drill core, with its four sequences (Fig. 3), represents only a fraction of the total Ries crater lake history. Due to a lack of marker beds, a precise correlation with the 314-m-thick succession of the Nördlingen 1973 drill core could not be established. Nonetheless, the alternation of bituminous shales and laminated olive-gray claystones with analcime, but only traces of clinoptilolite, pointed to a stratigraphic position at the transition of the laminite (unit B; Füchtbauer et al., 1977) to marlstone member (unit C) of the Nördlingen 1973 drill core (Fig. 9). There, the unit B–unit C transition sediments (280–300 m a.s.l.) lie 90–110 m deeper than the likely equivalents from the Enkingen drill core (390–405 m a.s.l.), which implies a steep dipping of the strata toward the lake center (Schauderna, 1983).
With respect to the general hydrological setting, the covariation of δ13C and δ18O of the carbonates (Fig. 4) indicated a hydrologically closed basin, thus confirming previous interpretations (Füchtbauer et al., 1977; Rothe and Hoefs, 1977; Jankowski, 1977, 1981; Talbot, 1990). Only the low-slope linear regressions found in single samples (e.g., at 21.10 m depth) may point to an artifact of sampling, specifically mixed values of lake carbonate and late diagenetic meteoric microspar. The positive δ13C and δ18O values, observed at 8.56 m depth, may correspond to highest salinities, while low, negative δ13C and δ18O ratios suggest lower salinities for the lower part of the section (18.22 m, 19.50 m, 21.10 m depth). Strongly deviating values for dolomites intercalated in the bituminous shales may reflect their precipitation within anoxic pore water at the lake bottom, as suggested, e.g., for the present-day Lake Qinghai, Tibet (Yu, 2005). Incorporation of carbon derived from organic substances, e.g., at the sulfate-methane interface (Meister et al., 2007) during dolomite formation, could explain the 13C-depleted carbon isotope ratios.
Depositional Environment and Sequences
The arrangement of lithofacies types shows meter-scale sequences, which are interpreted here as reflecting flooding-evaporation cycles. Flooding-evaporation cycles have previously been described by Jankowski (1981) from the basal member (unit A; Füchtbauer et al., 1977) of the Ries lake sedimentary succession in the Nördlingen 1973 drill core, but are not explicitly described from the younger units B–D, i.e., the laminite, marlstone, and claystone members (Jankowski, 1981). Nonetheless, flooding-evaporation cycles are also well known from lake-margin carbonates with blooms of filamentous green algae (Cladophora) during highstand conditions (Jankowski, 1981; Arp, 1995).
(1) In the Enkingen drill core, the first and second sequence of the Enkingen section start with conglomeratic sandstones (facies 1) interpreted as suevite-derived debrites (Table 1), showing unsorted clasts floating in a fine-grained matrix. These sediments point to rapid flooding at the inner slope of the inner ring and subaquatic redeposition. The third sequence starts with a laminated calcareous sandstone to arenaceous limestone and, thus, sediments more distant to the source of clastics (suevite). The fourth sequence finally starts with intercalations of carbonaceous clays (facies 7) at its base, reflecting strong influx of terrestrial plant debris. Influx from terrestrial habitats into all sequences is indicated by 17β,21β-bishomo-hopanoic acid from soil bacteria and n-octacosanoic acid from higher plant waxes.
(2) Following or alternating with conglomeratic sandstones, argillaceous Hydrobia limestones (facies 2) were deposited. Mud support and root voids (see Arp, 1995, pl. 8/2 and 9/9) in conjunction with a lack of vadose features in these carbonates indicate a subaquatic deposition below seasonal lowstands (i.e., the infralittoral; see Arp, 1995). The lack of eulittoral to supralittoral carbonates and the immediate deposition of infralittoral sediments on the debrites suggest a rapid lake-level rise.
(3) The continuing lake-level rise then caused drowning of the site below the chemocline of the lake. Bituminous shale (facies 4) deposition in sulfidic bottom waters of a stratified lake is indicated by high Corg contents, high sulfur contents, and isoprenoidal thiophenes (see also Rullkötter et al., 1990; Barakat and Rullkötter, 1997). The abundance of tocopherol and related thiophenes indicates high primary productivity coupled with good preservation under sulfidic conditions. Indeed, nutrients introduced from the catchment area during the flooding may have fostered high primary production.
C37 to C39 alkenones, as well as structurally related thiolanes in comparable shales (Barakat and Rullkötter, 1999), most likely derived from haptophytes, point to less saline conditions if compared to following facies association, at least for the water column.
The terrestrial palynomorph spectrum of the bituminous shales is characterized by the “warm temperate plant association” (WTPA) with Castaneoideae and Juglandaceae (e.g., Engelhardia, Platycarya), both preferring warm and humid conditions. At this time, the lake margins were covered by a diverse and dense forest vegetation. Likewise, swamp forest associations had their maximum extent during bituminous shale deposition. Aquatic palynomorphs are represented by Sigmopollis and Botryococcus, the latter known to be widespread in poorly oxygenated and stratified water bodies (e.g., Traverse, 1955; Guy-Ohlson, 1992; Lenz et al., 2011). In addition, dinoflagellates were most likely also important primary producers since dinosterol, a biomarker specific to dinoflagellates (Boon et al., 1979; Volkman et al., 1998), is abundant in the bituminous shales (Fig. 6; see also Barakat and Rullkötter, 1999).
(4) Stagnant to decreasing lake levels are assumed for the deposition of laminated greenish-gray claystones (facies 5). The continuous lamination indicates the persistence of the lake water stratification (cf. Jankowski, 1981). The decrease of Corg in the sediments can therefore be explained by a collapse of primary production due to substantial nutrient consumption and/or salinity increase, and not by the loss of water stratification and onset of holomixis (cf. Jankowski, 1981). The terrestrial palynomorph spectrum is then characterized by an increase in the halophilic-xerophilic association, while the aquatic association, with Botryococcus and mass occurrences of small coccal green algae, points to a change in lake water composition. Some biomarkers (17β,21ββ-bishomo-hopanoic acid, n-octacosanoic acid) indicate that terrestrial influx was still important, but the increase in δ13C values of archaeol, from about −28‰ in the lower part of the section to about −23‰ in the laminated claystones, points to more saline, possibly even hypersaline, conditions. Furthermore, the carbonaceous clay reveals 13C-enriched C20/C25-archaeol (Fig. 6). This biomarker has so far only been reported from halophilic Archaea and is therefore considered a robust marker for hypersaline conditions (De Rosa et al., 1982; Teixidor et al., 1993). The masses of coccal green algae (Volvocales, Chlorococcales?), on the other hand, point to freshwater conditions.
These observations may be explained by the following model:
(1) Episodic periods of heavy rain causing freshwater influx and superposition upon hypersaline lake water.
(2) Due to the nutrient influx and reduced salinities in the mixolimnion, a temporary bloom of planktonic green algae followed.
(3) Evaporation and disappearance of the fresher mixolimnion forced the green algal cells to transform to cysts due to unfavorable conditions. High sulfate reduction rates in the monimolimnion caused high sulfur contents in the sediments.
In conclusion, we interpret the allochthonous carbonaceous clay at the base of the fourth sequence (facies 7, “lignites”) of Enkingen to represent short-term intercalations resulting from heavy rain and storm events (tempestites) that superficially affected a hypersaline water body. Consequently, these intercalations do not indicate a general salinity decrease for the entire lake water body. On the other hand, the observed meter-scale sequences are interpreted as short-term climatic fluctuations between warm-humid and warm-arid conditions. Time estimates for such sequences, based on laminae couple thicknesses in bituminous shales (0.5–1.5 mm/couple for Enkingen shales at 17.60 m depth) are 4000–12,000 yr.
Although the four sedimentary sequences represent only a fraction of the whole lake history, some general trends can be derived. The 87Sr/86Sr ratios in lacustrine carbonates are a valuable tool to trace paleohydrology and different tributaries (e.g., Hart et al., 2004; Gierlowski-Kordesch et al., 2008; Davis et al., 2009; Jin et al., 2009). In the Ries Lake basin, 87Sr/86Sr values in carbonates are considered to reflect the relative fraction of dissolved ions derived from Variscan basement rocks versus that derived from marine Jurassic sediments (Pache et al., 2001).
At first glance, 87Sr/86Sr values show a wide scatter between marine Jurassic limestone and Variscan basement rock values, without any significant trend. However, samples with low carbonate contents (facies 1, 5, 6) exhibit a clear positive covariation of 87Sr/86Sr with aluminosilicate contents (Fig. 5), suggesting an isotopic exchange with silicates from the crystalline basement. This exchange is explained by the former presence of unstable silicate glasses in the suevite-derived siliciclastic influx, causing altered values in the carbonate fraction during glass decomposition and diagenesis. Excluding these samples, the remaining carbonates show a clear trend with values decreasing from 0.71185 to 0.71125, which is interpreted to reflect the lake water signal: Initial lake waters had a higher proportion of Variscan rock- and suevite-derived ions, while the influx from karst waters percolating through the Upper Jurassic carbonates of the catchment area increased with time and led to more “marine” ratios. Indeed, bioherm carbonates from younger lake stages finally show the lowest 87Sr/86Sr ratios (Fig. 10).
The Sr isotope trends within the carbonates have two implications. First, the ionic composition of the lake water changed with time, from solutes derived from weathering crystalline rocks to waters with an increasing proportion of waters from the Bunte Breccia. Without a subbottom leakage or significant increase in lake volume, this change in ionic composition was associated with an increase in salinity.
Second, the weathering substrates and soil types in the catchment area changed with time, from soils on suevite and crystalline rocks to soils on Bunte Breccia, and finally to dry karst soils on Upper Jurassic limestones. Soils on suevite could have been more fertile and moist, if compared to karst soils on ejected Upper Jurassic limestone blocks and autochthonous carbonates. Hence, a long-term change in the floral composition (i.e., the decrease of paleotropic and increase of arcto-tertiary elements) in the catchment area could partly be explained by a change in soil types. This also implies a much more widespread initial distribution of a suevite blanket, in contrast to previous assumptions of a discontinuous, patchy initial distribution (von Engelhardt et al., 1969; Gall et al., 1975). According to Salger (1977), a substantial erosion of suevite is also indicated by montmorillonite-rich strata in the lake sediments encountered by the Nördlingen 1073 drill core (von Engelhardt et al., 1995).
DISCUSSION: A NEW HYPOTHETICAL MODEL FOR RIES CRATER LAKE EVOLUTION
The data of the Enkingen drill core—in conjunction with data of reference localities—provide new arguments for a general reinterpretation of the Ries basin sedimentary succession (Fig. 9). In addition to climatic changes, chemical and geomicrobiological processes in the lake as well as the weathering history of the catchment area have to be taken into account.
Earlier investigations based on the Nördlingen 1973 drill core suggest that, after initial alluvial and playa sedimentation (with temporary freshwater conditions; Jankowski, 1981), a salinity maximum was reached during deposition of the analcime-rich and bituminous shales of the laminite member (unit B) in a permanent and stratified, highly alkali-saline lake. The following transition from bituminous shales (laminite member, unit B) to olive-gray marlstones and claystones (marlstone member, unit C) was interpreted as a change from semiarid to more humid climatic conditions, causing a salinity decrease, disappearance of salinity stratification, and finally holomixis (Füchtbauer et al., 1977; Rothe and Hoefs, 1977; Jankowski, 1981). In addition, Jankowski (1981) speculated that the salinity decrease might have been initiated by the hydrological change from a starved central basin shielded by the inner ring to a central basin with direct fluvial influx. Finally, the formation of an outlet during the onset of the claystone member deposition (unit D; with “lignite” seams) should have resulted in a progressive decrease in salinity and, finally, to swamp and freshwater conditions (Füchtbauer et al., 1977; Dehm et al., 1977), possibly with several shallow lakes within the basin (Jankowski, 1981).
The general trend from bituminous shales to claystones in the new Enkingen drill core is here interpreted to be linked to an increase, not a decrease, in salinity. Biomarkers and compound-specific carbon isotope data, in accordance with increasing δ13C and δ18O values and the change in aquatic palynomorph association, demonstrate a general trend to higher total ion concentrations toward the top of the investigated succession. Indeed, C20/C25-archaeol specific to halophilic Archaea was detected in the lignite (carbonaceous clay). Extrapolating our results to the whole lacustrine succession implies that the marlstone and claystone member with lignite seams encountered by the Nördlingen 1973 drill core does not necessarily reflect brackish to freshwater conditions, but instead represents salinities ranging from marine-like to hypersaline. Indeed, Füchtbauer et al. (1977, p. 18) pointed out that “as long as no outlet existed, it would be difficult to understand, how the lake became less saline during shallowing, even in a more humid climate.” In addition, von der Brelie (1977) mentioned that the increase of Chenopodiaceae pollen in the upper third of the Ries lake succession (upper part unit C, and unit D) could be interpreted as reflecting an increase in salinity. Actually, the co-occurrence of supposed freshwater indicators (lignite seams, freshwater gastropods; Dehm et al., 1977) and salinity indicators (traces of evaporite minerals; Füchtbauer et al., 1977) in unit D of the Nördlingen 1973 drill core does not require short-term salinity fluctuations of the entire lake water body, because there is no indication of autochthonous coal formation. In fact, episodic freshwater influx on top of the saline water body, introducing plant detritus and freshwater organisms into an increasingly hypersaline lake, and subsequent evaporation of the water layer provide a more straightforward explanation.
The view that salinities in the Ries Crater Lake increased rather than decreased also helps to provide a solution for previous discrepancies with respect to the time correlation of basinal series with marginal carbonates (Fig. 2) in the Ries crater (see Wolff and Füchtbauer  and Schauderna  versus Jankowski ). Accepting higher salinities for the claystone member with lignite seams (unit D) and intercalated ostracode beds (drillings of Deffner and Fraas, 1877), these deposits are not time equivalents of late relict freshwater sediments at marginal heights, but instead are more likely to form the equivalent of the Hainsfarth member green-algal bioherms (Fig. 2) with their monospecific ostracode fauna—a tentative correlation that needs to be tested by 87Sr/86Sr chemostratigraphy.
In any case, the suggested increase in salinity of the Ries crater lake was associated with a change in dissolved ion sources from suevite to Bunte Breccia/Upper Jurassic carbonate. The corresponding alteration of the chemical composition is reflected by the unidirectional trend in 87Sr/86Sr ratios. While authigenic analcime and clinoptilolite within the bituminous shales (laminite member, unit B) demonstrate highly alkali-saline conditions (Jankowski, 1981; “soda lake”; Table 6), foraminifera associated with Hainsfarth member algal bioherms (Arp, 1995) may indicate marine-like ion ratios (“marine lake”; Table 6). The youngest bioherms, i.e., bioherms of the Staudigberg member (Fig. 2), are devoid of ostracodes and aquatic gastropods such as Hydrobia but reveal abundant insect larval tubes as well as fecal pellets of the brine shrimp Artemia. Therefore, these bioherms, as well as the following oolitic carbonate sands, may reflect further increasing salinities (Table 6). A hydrochemical trend is also reflected by microbialite types and fabrics, i.e., the change from nonskeletal stromatolites in the Adlersberg member to skeletal stromatolites at the top of the Hainsfarth member (Table 6), which indicates a decrease in dissolved inorganic carbon (DIC) and an increase in Ca2+ (Arp et al., 2001). In fact, cyanobacterial stromatolites with lacustrine-vadose laminated organomicrite veneers are very similar to those of Lake Thetis, a saline lake with marine-like major ion composition but increased alkalinities (Reitner et al., 1996).
Based on these data and interpretations, a new hypothesis for the chemical and ecological evolution of the Ries crater lake is proposed (key arguments summarized in Table 6). Model calculations are provided to demonstrate that this hypothetical evolution of the Ries lake can be achieved by using present-day waters from the inner ring, Bunte Breccia, and Malmkarst (i.e., Upper Jurassic limestone aquifer) (Table 7; Fig. 9):
(1) After initial alluvial fans, a playa to freshwater lake developed (Jankowski, 1981). Model calculations demonstrate that equilibration of groundwaters from the inner ring to atmospheric pCO2 together with calcite precipitation would lead to lake water with pH values, alkalinities, and ion concentrations similar to the alkaline freshwater lakes of the Bosumtwi crater.
(2) Further evaporation led to a stratified brackish soda lake [Na-HCO3-Cl-SO4-type] with high primary production (Jankowski, 1981), possibly comparable to, e.g., Lake Van (Reimer et al., 2009). High alkalinities and Na+-dominated lake water composition were due to silicate weathering in the catchment area. In addition, sulfate reduction may have contributed to increased alkalinities in the hypolimnion (see, e.g., Kempe, 1990). Model calculations show, to achieve alkalinities similar to those of Lake Van, further evaporation and carbonate precipitation of the Ries inner ring groundwater would be required. Then, at alkalinities of ∼150 meq L–1, Ca2+, Mg2+, Sr2+, and SO42– concentrations are also similar to those of Lake Van waters, while Na+ and Cl– concentrations are lower.
So far, this scenario is in accordance with previous interpretations. After erosion of a suevite blanket, the successive increase in water influx from the Bunte Breccia increased the Ca2+ concentrations, while surplus Si(IV) and Na+ were removed by authigenic silicate formation, and CO32– was lowered by CaCO3 precipitation at the lake-margin Cladophorites bioherms. By these processes and further ion accumulation, the Ries lake attained marine-like ion ratios and concentrations [Na-(Ca)-Cl-SO4-type] (units C and D; potential analogs: Lake Thetis and Lake Clifton; Grey et al., 1990; Moore, 1993; Moore and Burne, 1994; Reitner et al., 1996).
Model calculations demonstrate that the conversion to marine-like lake water similar to that of Lake Thetis is possible using a specific ratio of the soda lake water and waters from the Bunte Breccia, in association with carbonate precipitation. High nutrient concentrations successively decreased due to algal consumption and reduced influx, causing a shift from bituminous shales to stratified claystones and a long-term shift from the green algal (Cladophorites) bioherms to cyanobacterial stromatolites. Present-day examples show that cyanobacterial stromatolites do proliferate under nutrient-limited conditions, while nutrient influx stimulates overgrowth by eukaryotic algae (e.g., Lake Clifton; Moore, 1993; Moore and Burne, 1994) and thick exopolymer-rich biofilms that poorly calcify. Indeed, short-term climatic fluctuations with humid periods cyclically fostered green algal bioherm growth during lake-level highstands.
(3) Successive nutrient consumption, decrease in nutrient influx via surface runoff, and further increase in ion concentrations caused a change to hypersaline conditions with increased Sr2+ concentrations, high Mg/Ca ratios, and mesotrophic conditions. Then, aragonitic cyanobacterial bioherms developed at the lake margin (“aragonitic top horizon”—Wolff and Füchtbauer, 1976; “bioherms type Staudigberg”—Arp, 1995), followed by widespread radial-fibrous ooid formation (potential analog: Great Salt Lake; Eardley, 1938) and subsurface dolomitization characterized by high Sr2+ (Wolff and Füchtbauer, 1976; Arp and Wiesheu, 1997), high Na+ (Arp, 2006), and low 87Sr/86Sr values (Fig. 10). There are no preserved basinal sediments of this stage and the subsequent final lake stage. Model calculations demonstrate that waters similar to Great Salt Lake waters could have developed from Ries waters by ongoing evaporation and carbonate precipitation.
(4) The final stage involved the abrupt change from unfossiliferous intraclast and ooid grainstones to micrites and marls rich in charophytes, ostracodes, and freshwater gastropods (Gyraulus, Planorbarius, Radix; Bolten, 1977; Arp, 1995). This sudden change to oligotrophic freshwater points to the formation of an outlet very late in the lake history, i.e., much later than the deposition of marlstone and claystone members C and D (cf. Füchtbauer et al., 1977). To numerically achieve a freshwater lake water composition like the suggested analog Lake Ohrid, the whole water body of the hypersaline Ries lake needs to have been replaced by karstic water from the Upper Jurassic (“Malmkarst”) aquifer.
In summary, the model suggests that the development from an alkaline freshwater to soda to halite and finally karst water lake was largely controlled by intrinsic factors of the impact crater basin (i.e., mainly weathering of different ejecta deposits and mineral precipitation within the lake), with a superimposed short-term cyclicity due to climatic fluctuations. Further investigations, among them new drill cores, extensive strontium isotope and palynological analyses, as well as hydrochemical budget calculations, are required to test this new model of the Ries crater lake evolution.
(1) The four sedimentary (meter-scale) sequences encountered by the Enkingen (SUBO 18) drill core represent a fraction of the Ries crater lake history that can probably be correlated with the transition of the laminite member (unit B) to the marlstone member (unit C) of the Nördlingen 1973 research drill core (Fig. 9). These sequences represent flooding-evaporation cycles due to short-term climatic fluctuations in a hydrologically closed basin.
(2) The 87Sr/86Sr ratios of carbonate beds of the Enkingen drill core show a unidirectional trend reflecting a change from weathering of suevite to Bunte Breccia (mainly Jurassic marine sediments) in the catchment area as the major source of ions to the lake water.
(3) The general transition from bituminous shales to olive-gray marly claystones with carbonaceous clay intercalations is not associated with a decrease, but an increase in salinity, as indicated by palynomorphs and 13C-enriched biomarkers specific to halophilic Archaea.
(4) Extrapolating these results to the entire Ries crater lake succession, a new hypothetical model of chemical and ecological evolution is proposed for the Ries crater lake: Silicate weathering and sulfate reduction initiated a permanently stratified eutrophic soda lake, which turned into a hypersaline mesotrophic halite lake by successive influx of waters from the Bunte Breccia, carbonate precipitation, and proceeding ion accumulation. Sr2+ and Mg2+ accumulation finally led to phreatic dolomitization of porous algal bioherms and carbonate sands late in the lake history. The chemical lake water development inferred from mineralogical, paleontological, and biomarker data is consistent with model calculations.
(5) The abrupt change from unfossiliferous oolites composed of radial-fibrous ooids to micrites rich in charophytes and freshwater gastropods (oligotrophic conditions) in the youngest preserved lake sediments was due to the formation of an outlet.
(6) Changes in terrestrial palynomorph associations are interpreted as reflecting a decrease in fertile, moist soils on suevite, and an increase in dry karst soils and soils on Bunte Breccia, superimposed by short-term climatic fluctuations.
In summary, intrinsic factors such as successive weathering of different rocks dominated over extrinsic, climatic factors in controlling lake evolution in the Ries crater. These data suggest that the initial suevite blanket might have been much more widespread than previously assumed.
To test whether and how specific hydrochemical lake types could have developed in the Ries basin, hydrochemical model calculations (program PHREEQC; Parkhurst and Appelo, 1999) were performed using groundwater and surface waters from the Ries basin and vicinity (Table 7). The calculations were designed to obtain hydrochemical compositions similar to lakes considered as modern analogs of the Ries lake during various stages of its development. Arguments for the selection of the modern lake analogs are summarized in Table 6. From the early to the final conditions of the Ries crater lake, modern analogs are the alkaline freshwater Bosumtwi crater lake (Turner et al., 1996), the brackish soda Lake Van (Reimer et al., 2009), the marine-like saline Lake Thetis (Grey et al., 1990), the halite-rich hypersaline Great Salt Lake (Post, 1980), and the karstic Lake Ohrid (Stanković, 1960; Kunz, 2006) (Fig. 9). Near-present-day levels are assumed for Miocene pCO2 (Pearson and Palmer, 2000; Royer et al., 2001).
In the first step, numerical adjustment of groundwater of the inner ring (drilling Busse; Winkler, 1972) to present-day pCO2 resulted in alkaline freshwater conditions and high calcite supersaturation, similar to that of the Bosumtwi crater lake (Table 7).
In the second step, H2O was subtracted to simulate evaporation until the alkalinity of Lake Van (∼150 meq L–1) was achieved. Carbonate precipitation was simulated by adjusting calcite, dolomite, and strontianite saturation indices. As a result, a high pH of 9.78 and Mg2+, Sr2+, and SO42– concentrations similar to Lake Van were obtained. Na+ and Cl– were lower than in Lake Van. Ca2+ was very low, i.e., typical for soda lakes. Gypsum remained undersaturated.
In the third step, 0.88 parts of this soda lake water were mixed with 99.12 parts of spring waters of the Bunte Breccia (Table 7). This specific mixing ratio, determined by trial and error, is required to simulate lake water similar to that of Lake Thetis by subsequent evaporation to just below gypsum saturation and adjustment to atmospheric pCO2 and ten-fold calcite supersaturation. However, Mg2+, SO42–, and alkalinity were higher, and pH, Sr2+, Na+, and Cl– were slightly lower than in Lake Thetis. It has to be emphasized that mixing with waters from the Upper Jurassic limestone aquifer (“Malmkarst”) (instead of water from the Bunte Breccia) was unsuitable to reduce pH and alkalinities significantly, despite carbonate precipitation. Therefore, to achieve a “marine-like” stage, influx of water from the Bunte Breccia is crucial.
In the fourth step, further evaporation of the lake water obtained from the mixing was applied, with subsequent gypsum, celestite, and dolomite precipitation to simulate a water composition similar to that of the North Arm of the Great Salt Lake. Gypsum precipitation led to Ca2+ concentrations lower than in the Great Salt Lake because no sulfate reduction to reduce the high sulfate concentrations was included in the model calculations.
A fifth calculation was performed using spring water of the nearby Upper Jurassic limestone aquifer (“Malmkarst”). After equilibration to atmospheric pCO2 and calcite precipitation (SICc = 1.0), freshwater lake conditions similar to Lake Ohrid were achieved (Table 7).
Many thanks go to Alexander Satmari and Harald Tonn for preparing thin sections. Gerald Hartmann and Angelika Reitz are kindly acknowledged for their support in X-ray fluorescence analysis. We also wish to thank Klaus Wemmer for his support in X-ray diffraction analysis. We are grateful to Andreas Pack and Ingrid Reuber for stable isotope analysis, Walter Riegel for palynological sample preparation, and Volker Wilde for taking scanning electron microscope pictures. The manuscript benefited from discussions with Dieter Stöffler, Uwe Reimold, and Jean Pohl. We are indebted to James Head III for very helpful corrections and comments on a previous version of this manuscript. The study was supported by the Deutsche Forschungsgemeinschaft (grants AR 335/5, BL 971/1, 3, and Th 713/3). We are grateful to Elizabeth Gierlowski-Kordesch and three anonymous reviewers for constructive criticism, helpful suggestions, and linguistic corrections. In addition, Associate Editor Ken MacLeod gave valuable comments and suggestions.